— George W. Russell was convicted by a King County jury of the first degree murder of Mary Ann Pohl-reich and the aggravated first degree murders of Carol Beethe and Andrea Levine.
Russell appealed, and the Court of Appeals certified his appeal to this court pursuant to RCW 2.06.030. Before examining the issues presented, we will briefly outline the facts pertinent to each count.
Facts
Count 1 — Mary Ann Pohlreich
On Friday, June 22, 1990, Mary Ann Pohlreich went to Papagayo’s, a Bellevue nightclub, with two friends. The three drove in Pohlreich’s car. Her two friends left Pohlreich at Papagayo’s at approximately 9:30 p.m.
Early the next morning, Pohlreich’s body was found partially inside the dumpster corral area in the parking lot behind the Black Angus restaurant, about a mile from Pa-pagayo’s. Pohlreich’s body was unclothed, but she was wearing two pieces of jewelry. There was a Frito Lay dip container lid over her right eye and forehead, her arms were folded over her stomach, her legs were extended and crossed at the ankles, and she had a pinecone in one of her hands.
Though there were a number of significant injuries, the King County medical examiner determined that the most likely cause of death was manual strangulation. Pohlreich’s skull fracture and numerous facial injuries appeared to have been inflicted by a fist. Pohlreich’s liver had two widely separated lacerations, and she had a distinct anal tear that the medical examiner opined was caused by a solid, nonhuman object. Pohlreich had a blood alcohol level of .14 percent at the time of her death.
Pohlreich’s purse, sweater, and car were found at Papa-gayo’s. The police discovered that on the evening of Pohl-reich’s death George Russell went to Papagayo’s with his *31friend Smith McLain to have dinner. Once there, Russell and another friend talked to Pohlreich. After dinner, Russell borrowed the keys to McLain’s truck, explaining that he had to change into a shirt with a collar as required on Papagayo’s dance floor. Russell had a duffle bag in the truck.
A Bellevue police officer was working off duty that night as a doorman at Papagayo’s. He often spoke with Russell, a frequent patron of the nightclub. That night he saw Russell twice: once shortly after he began his shift at 10:30 p.m. and again approximately an hour later. On the second occasion, Russell told the officer that he was going to take "this girl” over to her place to get something. Verbatim Report on Appeal, at 4042, 4052. The officer did not see the woman well enough to identify her; however, he described her size as being similar to Pohlreich’s, and he noticed that the woman seemed very intoxicated.
Russell did not return to Papagayo’s that evening with the truck. McLain was upset and spent the rest of the evening waiting for him at the Overlake Denny’s where it was customary for Papagayo’s patrons to go after closing. At around 5:30 a.m., McLain found a ride home.
On the morning that Pohlreich’s body was discovered, Russell telephoned the McLain residence. Russell said that he had been out looking for McLain all night. At approximately 6 a.m., McClain’s sister, Shawn Calvo, saw Russell return in her brother’s truck. Russell told her that he had borrowed the truck to drive a friend home and then could not find McLain. During this conversation Calvo noticed a reddish-orange stain on the passenger seat of the truck. Russell explained that his friend had vomited clam chowder in the truck. Russell declined Calvo’s offer of a ride home and walked away with his duffle bag.
McLain woke up after Russell had left and went out to inspect his truck. He smelled a strong offensive odor that reminded him of vomit or the smell of a deer gutted after a hunting kill. Russell called McLain that morning and told him that he had thrown up in the truck after drinking too much. Russell informed McLain that he had driven a *32woman home in the truck because he did not want to be seen in the woman’s Porsche.
Russell had spoken previously about the woman with the Porsche. That woman was Tamara Francis. Francis testified that she knew Russell but had never left Papagayo’s with him.
The police also discovered that Russell had been a regular customer at the Black Angus, where Pohlreich’s body was discovered, from 1989 until March 1990, when he was banned from the restaurant. Russell was very angry about that decision.
On October 11,1990, almost 3x/2 months after Pohlreich’s murder, the police removed the interior of Smith McLain’s truck. While the interior had been cleaned and detailed during the summer, the floormats had not. McLain had removed the floormats, which were made from house carpet remnants, from his truck and had put them in the garage because they smelled so bad.
The upholstery in McLain’s truck reacted positively for blood and [¶] and A antigens which matched the [¶] and A antigens in the vaginal swab taken at the Pohlreich autopsy. Both antigens could have been contributed by Pohlreich herself who was a type A secretor. Russell was type O; although he could have been the source of the [¶] antigens, he could not have contributed the A antigens.
The State also sent the vaginal swab and upholstery samples for DNA testing. Because of the poor quality of the samples, the laboratory conducted a polymerase chain reaction (PCR) test. The PCR test results indicated that neither Russell nor McLain could have been the source of the blood in McLain’s truck, but that Pohlreich could have been. The testing also revealed that only Russell, of all the comparison samples, could have been the donor of the sperm.
In addition to the sperm and blood samples, one negroid hair, consistent with Russell’s, was found in the debris on the sheet in which Pohlreich’s body was wrapped. Five fibers found in the pubic combings were consistent with the truck carpet as was one fiber from the sheet debris. Another fiber *33in the sheet debris was consistent with the truck’s upholstery.
Count 2 — Carol Beethe
Carol Beethe was employed as a bartender at Cucina Cucina, a restaurant in Bellevue. She lived in a condominium with her two children. Beethe’s ex-husband, Paul, lived nearby. On August 8, 1990, Beethe spoke with Paul at around 9:30 p.m. At 10:30 p.m. she spoke with her boyfriend, Mike Suell, with whom she was planning to go on a vacation. At midnight she met another friend at the restaurant where he was the bartender. Beethe left at approximately 2:15 a.m.
At 4:30 a.m. Beethe’s daughter Kelly heard someone in the hall of the family’s condominium, and then saw the person shine a flashlight in the bathroom, her sister’s bedroom, and her own bedroom. Kelly assumed that the person was Mike Suell.
When Kelly awoke at 8:30 a.m., her mother was not up as she usually was, her bedroom door was locked, and Kelly could not wake her. When Kelly went outside to open the sliding glass door to her mother’s room, she saw her mother and became scared. She called her father who came over and entered the room through a sliding glass door.
Beethe was on her back on the bed. The bedspread was pulled down to the foot of the bed. Her body was unclothed except for a pair of red high-heeled shoes. Her feet were together with legs spread and knees bent. Blood had been smeared on her legs in a manner that resembled "finger painting”. Verbatim Report on Appeal, at 3290. A rifle had been placed resting symmetrically between Beethe’s legs, resting on her shoes. The firearm penetrated approximately five or six inches into her vagina. Her left arm was bent upward at the elbow, while her right arm was bent down at the elbow, nearly touching her hip. Beethe’s head was wrapped in a plastic bag and covered with a large pillow.
The medical examiner ascertained that Beethe’s death had been caused by head injuries. The head injuries were inflicted by an instrument swung with considerable force in *34rapid succession. The blows left distinct "Y” shaped marks and crushed the entire left side of Beethe’s skull. Beethe had also been struck many times with a knee or fist in the torso. Her ribs were broken and her liver was lacerated.
Testimony suggested that Beethe and Russell were acquaintances. One witness testified that both Russell and Beethe frequented the Overlake Denny’s. A waitress at the Black Angus testified that on two occasions she was talking to Beethe about a "situation . . . between George [Russell] and I” and saw Russell glaring at them. Verbatim Report on Appeal, at 4898-99. (This occurred before Russell was banned from the Black Angus.) After the murders, Russell told friends that he knew the victim of the second murder and that she was a bartender at the Cucina Cucina restaurant in Bellevue.
When Beethe’s body was found, she had rings on her right hand but not on her left hand. At the time of her death, Beethe owned two wedding ring sets, one from her mother and one from her previous marriage. The rings were kept in a jewelry box in Beethe’s bedroom, but they were never located after her death. During their investigation, the police published photographs of the rings in a Bellevue newspaper. At trial, one of the State’s witnesses testified that Russell had tried to sell him rings that resembled the missing set.
Beethe’s family also informed the police that she had a half dozen small Crown Royal bags in the top drawer of her dresser containing silver dollars and other change from tips. When police allowed Paul to reenter his ex-wife’s house he noticed the Crown Royal bags were missing.
About three weeks after Beethe’s murder, Russell and a friend drove to a wooded area on Mercer Island. Russell informed his friend that he had to pick up some money owed him. Russell stepped out of the car and returned with a paper bag full of silver dollars and change.
Forensic evidence did not reveal any of Russell’s fingerprints in Beethe’s residence. A fabric glove impression left on the sheet of her bed suggested that the murderer wore *35gloves. Negroid hairs were discovered on Beethe’s sheet, pillow and underwear, but the fragments were not suitable for comparison.
Count 3 — Andrea Levine
Andrea Levine rented a basement apartment in the home of Robert Hays and his wife. On Thursday, August 30,1990, Hays saw Levine after she returned from work. Later that evening, Levine met her boyfriend at a Kirkland restaurant, where the two discussed plans to go to the San Juan Islands. Levine declined a ride and drove herself home to pack at about 1:30 a.m.
Hays and his wife awoke on Friday morning at about 5 a.m. Hays opened the back door to let their dog out. The dog began barking wildly. Hays stepped out to investigate and saw a dark figure about 25 to 30 feet away. It was dark and Hays could see only an adult with a white form, approximately two-thirds of the width of the person, on or in front of the individual’s abdomen. Hays called out and the person fled. Hays chased the person a short distance but stopped because he was unarmed. He called the police who examined the yard but did not check Levine’s apartment.
The following Monday, Hays’ wife entered Levine’s apartment because one of Levine’s cats appeared hungry. As she walked down the hallway, she smelled something like old blood coming from the bedroom. She opened the door and discovered Levine’s body. Levine was on her back, on the bed. Her face was turned toward her left shoulder. Her legs were spread with knees straight. Her right arm extended above her shoulder while her left arm rested by her side. Under Levine’s left forearm was the book More Joy of Sex. A plastic dildo was partially inserted into Levine’s mouth.
The medical examiner determined that Levine had died from severe multiple head wounds inflicted with an object such as an iron bar. Levine’s body was covered with postmortem stab wounds. Forensic evidence revealed the presence of a single negroid pubic hair at the crime scene. This hair could not be matched to any samples taken from Rus*36sell, although Russell could not be excluded, either. No fingerprints could be found, suggesting to the police that someone had "wiped the scene down”. Verbatim Report on Appeal, at 5286, 5304-05.
At trial the State presented evidence showing that Russell knew Levine. On one occasion, Levine’s boyfriend drove her home in her truck and Russell followed in the boyfriend’s car. A few weeks later, Russell and some friends drove out to Renton to help Levine put a new battery in her truck. Russell rode back to Kirkland with Levine. On a third occasion, Levine was at a bar and Russell came over to talk. After Levine’s murder Russell made disparaging remarks about her, stating that she "slept around”, that she "used men”, and that she was a "whore”. Verbatim Report on Appeal, at 5464.
On the Labor Day weekend after Levine was killed, Russell and some friends went to Canada. On August 30, the night before they left, the group stayed at a motel. Russell left during the night dressed in dark pants, white tennis shoes, a dark blue sweatshirt, and a dark cap. He said something about going to work. He returned at about 6 a.m. wearing the same clothing. Russell did not have a car. Levine lived about a mile from the motel by car, but the walking distance was shorter.
A friend of Russell’s testified that she received a ring from Russell several days after Labor Day. She wore the ring several times and then gave it to a friend who pawned it. The police later retrieved the ring from the pawnshop. Levine’s sister-in-law identified the ring as one that she had given to Levine.
Police brought the ring to a jewelry store where an employee identified it as one he had worked on for Levine in February 1990. The owner of the jewelry store also testified that Levine had brought the ring in to be worked on.
Russell was arrested 8 days after Levine’s murder based on some outstanding misdemeanor warrants. After interviewing Russell, the police charged him with the murders of Mary Ann Pohlreich, Carol Beethe, and Andrea Levine. Rus*37sell raises several issues in this appeal. Other facts will be discussed where relevant to the issue presented.
I
Russell first challenges the trial court’s conclusion that PCR testing of DNA has gained sufficient scientific acceptance to admit the results of such testing in court. The PCR tests pertained to count 1, and were conducted on a vaginal swab taken from Mary Ann Pohlreich’s body and a piece of stained upholstery removed from Smith McLain’s truck. Before discussing the results of those tests and their admissibility, it is useful to briefly outline the PCR identification process. Our discussion is taken from Thomas M. Fleming, Annotation, Admissibility of DNA Identification Evidence, 84 A.L.R.4th 313 (1991) unless otherwise noted.
DNA identification techniques are based on a system of premises concerning the nature and function of the DNA molecule, a long, threadlike structure resembling a twisted ladder packed into the chromosomes in every nucleated cell. Each side of the ladder is composed of a chain of sugars and phosphates, while the rungs attached to them, approximately 3 billion in number, consist of pairs of molecules called bases — adenine, cytosine, guanine, and thymine — each of which bonds with only one of the others. The particular order of the four bases along the DNA molecule constitutes a genetic code which governs the production of proteins making up the organism. While most sections of this chain of bases are largely the same among individuals within a given species, certain sections are variable or "polymorphic”, meaning that they may take different forms in different individuals. A gene, which is the sequence of bases responsible for producing a particular protein, thus may be polymorphic, having two or more possible variations called "alleles”. Fleming, at 319.
Because of polymorphism in human genetic structure, no two individuals (except for identical twins) have identical base sequences throughout their DNA, although no one’s base sequence within a particular section is absolutely *38unique. Moreover, a given person’s DNA is the same in every nucleated cell in the body, whether from hair follicles, blood, semen, or other tissues, and remains the same throughout life, barring rare mutations. These are the qualities that make possible the identification of a specific person, to the practical exclusion of others, through analysis of his or her DNA. Fleming, at 319-20.
The two kinds of DNA tests most commonly used for identification are restriction fragment length polymorphism (RFLP) analysis, and polymerase chain reaction (PCR) amplification. PCR is the test at issue here. One microbiologist has described PCR as a genetic photocopy machine. Kamrin T. MacKnight, Comment, The Polymerase Chain Reaction on (PCR): The Second Generation of DNA Analysis Methods Takes the Stand, 9 Santa Clara Computer & High Tech. L. J. 287, 304 (1993).1 The PCR amplification system takes advantage of the natural DNA replication system and manipulates it to the advantage of the analyst to produce many millions of DNA copies. MacKnight, at 304.
In the PCR procedure, DNA is extracted from a sample, purified, and added to a buffer solution containing chemical primers and an enzyme called "Taq polymerase”. MacKnight, at 305. The solution is then placed in a heating device called a thermal cycler which cycles it through several successive temperature plateaus. After 30 or 40 of these cycles, the DNA has become denatured and the primers have annealed to it, identifying a "gene of interest” which will have been replicated or amplified fay the enzyme billions of times. Fleming, at 323. In a procedure called the reverse dot-blot process, the amplified DNA is flooded over a nylon membrane onto which have been dotted a number of "allele-specific” probes, each designed to recognize one variant of the "gene of interest”.2 *39Fleming, at 323. This may result in a color reaction and a visible dot on the membrane wherever a probe has identified one of the alleles. To determine whether two samples could have come from the same person, the analyst checks whether they have produced the same pattern of dots. If one sample produces a dot in response to a probe to which the other did hot react, the samples could not have a common source. Fleming, at 323.
The AmpliType DQ alpha test kit is one of the few PCR test systems sufficiently developed for forensic use, and was used to conduct the PCR tests in this case. See MacKnight, at 307. This kit was developed by the Cetus Corporation and will be referred to hereafter as the Cetus kit. The Cetus kit uses probes that identify the six different alleles present in the HLA DQ alpha genetic marker system. (HLA stands for the human leukocyte antigen system. Leukocytes are white blood cells; antigens are proteins.) The six alleles present in this system are denominated as 1.1, 1.2, 1.3, 2, 3, and 4. These alleles are combined in pairs in each person, because one is received from each parent. There are 21 possible pairs of these traits, and each pairing is called a "genotype”. Fleming, at 323.
If the DQ alpha genotype of a suspect is different from that of the evidence sample, the suspect presumably is excluded as the donor of the evidence. Unlike matches or inclusions, exclusions are independent of the frequencies of the genotype in the population. MacKnight, at 310. If the suspect and evidence have the same genotype, then the suspect is included as a possible source of the evidence sample. The probability that another, unrelated, individual would also match the evidence is equal to the frequency of that genotype in the relevant population. MacKnight, at 310.
*40PCR testing at the DQ alpha locus provides a power of discrimination of approximately 83 to 94 percent. Mac-Knight, at 312. This compares favorably to that of the ABO red cell typing system. By itself, however, PCR testing can neither provide individual identification nor the very high power of discrimination possible with RFLP methods of DNA typing. MacKnight, at 311. PCR analysis has proved useful, though, in including or excluding criminal suspects in circumstances where conventional typing has failed or insufficient DNA was available for RFLP testing. RFLP analysis requires relatively large samples of DNA, whereas PCR testing is capable of analyzing minute and degraded samples. MacKnight, at 297-98, 312. Other advantages of the PCR technique are that it takes much less time to achieve results than RFLP, is less expensive, and achieves results that are easier to interpret. State v. Williams, 252 N.J. Super. 369, 379-80, 599 A.2d 960 (1991).
The PCR test results in this case indicated that the sperm on the vaginal swab contained Russell’s DQ alpha genotype, and that the DQ alpha type recovered from the blood-stained upholstery was consistent with Pohlreich’s blood but inconsistent with Russell’s and McLain’s. Prosecution witnesses concluded that Russell’s DQ alpha genotype occurs in approximately 5 to 10 percent of the population. Therefore, 90 to 95 percent of the population (but not Russell) could be excluded as sources of the sperm in Pohlreich’s body. Since Pohlreich’s DQ alpha genotype is shared by 4 to 9 percent of the population, approximately 91 to 96 percent of the population (including Russell), could be excluded as sources of the bloodstain in McLain’s truck. This exclusion was significant since Russell had told McLain that his (Russell’s) vomit was the source of the upholstery stain.
The trial court admitted these test results after a lengthy pretrial Frye hearing. Under the Frye standard for novel evidence, scientific evidence will be admitted only if it is generally accepted in the relevant scientific community. See Frye v. United States, 293 F. 1013, 1014, 34 A.L.R. 145 (D.C. Cir. 1923). This court recently reaffirmed the Frye *41standard when it held RFLP evidence admissible in State v. Cauthron, 120 Wn.2d 879, 886, 846 P.2d 502 (1993).3
The core concern of Frye is whether the evidence being offered is based on an established scientific methodology. This involves both an accepted theory and general acceptance of the technique used to implement that theory. Cau-thron, at 889. The Frye inquiry does not require acceptance of the laboratory testing procedures used in the case before the court. Cauthron, at 889. If the methodology is sufficiently accepted in the scientific community at large, concerns about the possibility of error or mistakes made in the case at hand can be argued to the factfinder. Cauthron, at 889.
This reference to the scientific community at large is important — a court looks not only to the technique’s acceptance in the forensic setting but also to its acceptance by the wider scientific community familiar with the theory and underlying technique. See Cauthron, at 896-97; Williams, at 381. Under Frye, the court looks for "general acceptance in the appropriate scientific community”, that is, acceptance by the community of scientists familiar with the challenged theory. See Cauthron, at 887; People v. Young, 425 Mich. 470, 481, 391 N.W.2d 270 (1986) (ideal community would be scientists with direct empirical experience with the procedure in question). Furthermore, Frye requires only general acceptance, not full acceptance, of novel scientific methods. This only makes sense since full acceptance would obviate the necessity of a Frye hearing. The court’s review is de novo. See Cauthron, at 887.
During the 2-week Frye hearing in this case, the State introduced four experts to support its position that PCR *42analysis meets the scientific acceptance required by Frye. These experts included Dr. Rebecca Reynolds, a Cetus employee; Dr. Daniel Geraghty, a molecular geneticist employed by Fred Hutchinson Cancer Research Center; Dr. Cecelia Von Beroldingen, forensic DNA specialist at the Oregon state police crime laboratory; and Dr. Edward Blake of Forensic Science Associates, a forensic serologist who conducted the PCR tests at issue. The defense presented Dr. John Gerdes, DNA analysis director at Immunological Associates of Denver; Dr. Kristen Skogerboe, a clinical chemist at the Laboratory of Pathology; and Dr. Glenn Evans, associate molecular laboratory professor at the Salk Institute.
Dr. Reynolds acknowledged that PCR testing received attention first from biomedical researchers, but she added that over 30 laboratories were either in the process of implementing DQ alpha typing, using the Cetus kit, or already using it on casework. She regarded the PCR technique as a useful and reliable source of genetic markers.
Dr. Von Beroldingen also testified to the growing use of the Cetus kit by both private and governmental laboratories. She further stated that the forensic application of PCR testing in this case met with generally accepted scientific principles and practices.
Dr. Daniel Geraghty testified that PCR analysis was the major technical development of the last 10 years in molecular immunology and that it met with no pockets of resistance. He saw no serious dispute regarding the scientific principles involved in PCR testing. Fred Hutchinson uses the Cetus kit to help determine compatibility for bone marrow transplants. The same technology is used by Dr. Blake and Fred Hutchinson, with the only difference being the method of seeing the result. Dr. Blake uses an enzymatic color change which turns blue if the enzyme is present, while Fred Hutchinson uses a chemifluorescent method. Dr. Geraghty described them both as "perfectly reasonable and well established methods”. Verbatim Report on Appeal, at 216. Dr. Geraghty also uses the dot-blot method described *43earlier while the Cetus kit employs the reverse dot-blot method. He stated that Dr. Blake’s use of the reverse method, with its lower likelihood of error, made "perfect sense”. Verbatim Report on Appeal, at 220. Fred Hutchinson uses the same amplification method as Dr. Blake, with all reagents coming from Cetus, and also has approximately 50 thermal cyclers, which the doctor termed "very reliable”. Verbatim Report on Appeal, at 217.
Dr. Geraghty stated that PCR was now used by almost every molecular biologist in the world, and added that he had heard no criticisms of the Cetus kit. Dr. Geraghty said that Dr. Blake’s tests in this case were conducted properly and he expressed confidence in the results even though Dr. Blake conducted only one test on the swab. (The bloodstain sample was performed in triplicate.)4
Defense expert Dr. John Gerdes testified that while RFLP has been used by research labs for 15 to 20 years, PCR is a newer method. Dr. Gerdes felt it was too early for the Cetus kit to produce reasonable results in a forensic setting because of the danger of typing a contaminant, the low power of discrimination, and the lack of independent validation of laboratories other than Dr. Blake’s. On cross examination, Dr. Gerdes stated that he saw no evidence of contamination in the present case.
Dr. Skogerboe was concerned that insufficient validation studies had been performed on the use of the Cetus kit for crime-scene evidence. On cross examination, however, she acknowledged that her current laboratory had been using PCR on clinical samples for diagnosis for l1/2 years, and she regarded the benefits of using PCR as outweighing any problems associated with the test.
Dr. Evans regarded the Cetus kit as unreliable for use in a forensic setting. He added, however, that he believed PCR analysis would "revolutionize society, particularly forensics”, and saw no resistance to the techniques of enzymatic amplifi*44cation of DNA in general. Verbatim Report on Appeal, at 1546, 1549.
Following this testimony, the trial judge found the results of PCR testing at the HLA DQ alpha locus admissible under the Frye standard. The judge found that the underlying principle and techniques of PCR had been generally accepted by the scientific community, and added that DQ alpha testing and typing had gone sufficiently beyond the experimental stage to gain general acceptance in the scientific community. She observed further that the DQ alpha gene has been subjected to considerable scientific study, especially in the fields of immunology and medicine. The variations of the gene are well known, readily identified, and easily distinguished, making this gene an appropriate genetic marker for forensic use. She also noted that the population frequencies of the various genotypes occurring at the DQ alpha locus were not contested. The judge concluded that the fact that a scientific procedure might yield a false result if not performed properly did not render it inadmissible and that any problems associated with PCR testing at the DQ alpha locus went to the weight to be given the evidence.
On appeal, Russell challenges these conclusions, and argues that the State failed to demonstrate that PCR analysis has gone through the extended period of use and testing in the forensic community necessary to achieve the general scientific acceptance required under the Frye standard.
Russell derives most of his support for this position from a recent report prepared by the Committee on DNA Technology in Forensic Science, under the auspices of the National Academy of Sciences. The Committee was formed in 1990 to address the general applicability and appropriateness of the use of DNA technology in forensic science, the need to develop standards for data collection and analysis, aspects of the technology, management of DNA typing data, and legal, societal and ethical issues surrounding DNA typing. Nat’l Research Coun., DNA Technology in Forensic Science 1-2 (1992) (hereinafter DNA Technology). While most of the *45resulting report focuses on the RFLP method of DNA identification, it also discusses the PCR technique.
The Committee found no scientific dispute about the validity of the general principles underlying DNA typing. DNA Technology, at 51. The Committee apparently found less agreement, however, over whether a given DNA method might be scientifically appropriate for forensic use. "Before a method can be accepted as valid for forensic use, it must be rigorously characterized in both research and forensic settings to determine the circumstances under which it will and will not yield reliable results.” DNA Technology, at 51-52.
The basic thrust of the report was the need to standardize forensic DNA typing to the extent possible (realizing that the lack of control over crime-scene evidence makes standardization problematic). To achieve such standardization, the report recommended quality-assurance programs, individual certification, laboratory accreditation, and state or federal regulation. DNA Technology, at 16. With regard to PCR testing, the Committee observed that one commercial kit for forensic PCR analysis has been marketed. DNA Technology, at 69. While the report voiced no criticisms of the existing kit, it saw a potential for the introduction of unreliable kits and the misuse of kits.
The committee believes that nonexpert laboratories will run a significant chance of error in using kits. We therefore recommend that a standing committee . . . consider the issue of regulatory approval of kits for commercial use in forensic DNA analysis. Even though no precedent exists for regulation of tests in forensic applications, we believe that it might be necessary for a government agency to test and approve kits for DNA analysis before their actual forensic use.
DNA Technology, at 69. The report then summarized its findings regarding the forensic use of PCR analysis:
PCR analysis is extremely powerful in medical technology, but it has not yet achieved full acceptance in the forensic setting. The theory of PCR analysis ... is scientifically accepted and has been accepted by a number of courts. However, most forensic laboratories have invested their energy in development of RFLP technology and have left the development of forensic PCR technology to a few other laboratories. Thus, there is no *46broad base of experience in the use of the technique in identity testing.
Forensic PCR-based testing is now limited for the most part to analysis of genetic variation at the DQ locus in the HLA complex. . . .
. . . [F]urther experience should be gained with respect to PCR in identity testing. Information on the extent of the contamination problem in PCR analysis and the differential amplification of mixed samples needs to be further developed and published. A great deal of this information can be obtained when a number of polymorphic systems are available for PCR analysis. . . .
. . . Considerable advances in the use of PCR in forensic analysis can be expected soon; the method has enormous promise.
DNA Technology, at 70.
Russell points to this language as evidence of the lack of general scientific acceptance of PCR testing. In a subsequent chapter on the use of DNA in the legal system, however, the report acknowledges the admissibility of DNA evidence, without distinguishing between the PCR and RFLP methodology, so long as the precautions outlined are taken (as discussed later in this opinion):
It is not necessary, at this stage of development of DNA typing, to hold extensive admissibility hearings on the general validity of the scientific techniques, although cases will still arise in which the procedures used to report a match will be questioned.
As a general matter, so long as the safeguards we discuss in this report are followed, admissibility of DNA typing should be encouraged. There is no substantial dispute about the underlying scientific principles. However, the adequacy of laboratory procedures and of the competence of the experts who testify should remain open to inquiry.
DNA Technology, at 145-46.
The report’s message regarding the forensic use of DNA typing has been interpreted contrarily to Russell’s position both by experts in the scientific community and a number of court decisions. On April 14,1992, a New York Times article stated that courts should cease to admit DNA evidence until laboratory standards have been tightened and the technique is established on a stronger scientific basis. Gina Kolata, *47U.S. Panel Seeking Restriction on Use of DNA in Courts, N.Y. Times, Apr. 14, 1992, at Al. In response, the chairman of the DNA Committee stated that " '[w]e think that DNA can be used in court without interruption.’ ” Gina Kolata, Chief Says Panel Backs Courts’ Use of a Genetic Test, N.Y. Times, Apr. 15, 1992, at Al. A statement included in the Committee’s final report referred to the Times articles and provided this clarification:
We recommend that the use of DNA analysis for forensic purposes, including the resolution of both criminal and civil cases, be continued while improvements and changes suggested in this report are being made. There is no need for a general moratorium on the use of the results of DNA typing either in investigation or in the courts.
We regard the accreditation and proficiency testing of DNA typing laboratories as essential to the scientific accuracy, reliability, and acceptability of DNA typing evidence in the future. Laboratories involved in forensic DNA typing should move quickly to establish quality-assurance programs. After a sufficient time for implementation of quality-assurance programs has passed, courts should view quality control as necessary for general acceptance.
DNA Technology, at x.
A California Court of Appeal described the final National Research Council report as concluding that "there is indeed a need for standardization of laboratory procedures and proficiency testing (as well as appropriate accreditation of laboratories) to assure the quality of DNA laboratory analysis”. People v. Barney, 8 Cal. App. 4th 798, 812, 10 Cal. Rptr. 2d 731 (1992). The absence of such safeguards does not mean, however, that DNA analysis is not generally accepted. Rather, the question becomes whether a laboratory has complied with generally accepted standards in a given case. Barney, at 812-13.
The Colorado State Supreme Court also concluded that concerns about the forensic use of a given scientific theory bear on the weight accorded DNA typing evidence rather than its admissibility. Fishback v. People, 851 P.2d 884, 893 (Colo. 1993); see also FBI, The FBI’s Responses to Recommendations by the NRC’s Committee on DNA Technology in *48Forensic Science, 19 Crime Laboratory Dig. 49, 69 (1992) (issue of correct application of a valid scientific technique such as DNA analysis in a particular case is a question of fact and matter of weight which should be decided by jury). Similarly, a federal district court held that concerns about the forensic applications of RFLP did not bar its admissibility in United States v. Jakobetz, 747 F. Supp. 250, 256-58 (D. Vt. 1990), aff'd, 955 F.2d 786 (2d Cir.), cert. denied, 113 S. Ct. 104 (1992).
In addition to citing the DNA Technology report, Russell points more specifically to the absence of accreditation of forensic laboratories, the lack of published literature and professional testing of PCR, and the small number of forensic laboratories doing PCR work to demonstrate that PCR testing has not yet endured sufficient scientific analysis for general acceptance and thus admissibility under Frye.
As we evaluate these challenges it is important to reiterate that our inquiry is not confined to an examination of PCR testing in the forensic laboratory setting. Rather, we are concerned with the extent of peer review and acceptance as manifested in the general scientific community.
PCR analysis is in routine use in many settings. PCR testing has been used in HIV detection and diagnosis; in identifying microorganisms in the aquatic environment, as well as in food, dairy, soil, and clinical samples; in neonatal screening for cystic fibrosis and sickle cell anemia; in detecting chromosomal abnormalities and mutations; in gene replacement therapy; in human pedigree analysis; in studying the epidemiology of Lyme disease; and in the new fields of molecular anthropology and molecular paleontology. PCR analysis also is being used to monitor environmental contamination, to establish the new field of diagnostic molecular pathology, and to help identify those killed in the Persian Gulf War. MacKnight, at 302-04. In this case, Dr. Geraghty testified for the State that PCR is used by nearly every molecular biologist in the world, and defense expert Dr. Evans stated that he saw no resistance to the techniques of enzymatic amplification in general. Moreover, we do not *49agree with Russell’s contention that only a few forensic laboratories are using PCR analysis. Cetus reported that over 30 forensic labs were performing DQ alpha typing as of March 1991. MacKnight, at 319. (The same article listed five forensic labs as actually performing RFLP and stated that because of its difficulty, RFLP was best performed in research labs.) MacKnight, at 295. Cetus also reports that the FBI began using the Cetus kit in 1992. AmpliType HLA DQ-Alpha Forensic DNA Typing Customer Survey (1992). Dr. Reynolds testified that the British Home Office has adopted DQ alpha as its screening test. As of December 1989, the HLA DQ alpha typing system reportedly had been used in 106 forensic cases involving the analysis of over 1,000 evidence samples. Rhea Helmuth et al., HLA-DQ-Alpha Allele and Genotype Frequencies in Various Human Populations, Determined by Using Enzymatic Amplification and Oligonucleotide Probes, 47 Am. J. Hum. Genetics 515, 521 (1990). As of October 1991, PCR-based DQ alpha typing methods were used in biological evidence in over 250 cases. MacKnight, at 325.
Russell is correct that the National Research Council report cites accreditation and governmental regulation of forensic laboratories as potential external mechanisms needed to ensure quality science. DNA Technology, at 16. However, the report also observed that no precedent exists for regulation of forensic testing. DNA Technology, at 69. Furthermore, Dr. Geraghty stated that Fred Hutchinson, a medical research institute, had not yet received accreditation to perform DNA testing. While accreditation and regulation may be desirable in the medical as well as the forensic setting, it is not necessary to bar the use of DNA technology until such safeguards are in place. "Although the court is not the ideal forum for ensuring quality science, the adversary process is a means by which those who practice 'bad’ science may be discredited, while those who practice 'good’ science may enjoy the credibility they deserve.” MacKnight, at 341.
Russell also states that PCR testing lacks sufficient validation studies for admissibility. We disagree and conclude *50that extensive validation studies have been conducted on PCR testing. See MacKnight, at 344. Following repeated validation experiments, the FBI found that typing of the DQ gene by PCR and detection of specific alleles can be accomplished, when the typing is done using proper protocols, without producing false positive or false negative results. Catherine T. Comey & Bruce Budowle, Validation Studies on the Analysis of the HLA DQ Locus Using the Polymerase Chain Reaction, 36 J. Forensic Sci. 1633 (1991). An Oregon Court of Appeals found "no dispute” that scientific articles concerning the PCR method exist and that peer-reviewed journals publish many of the articles. State v. Lyons, 124 Or. App. 598, 608-09, 863 P.2d 1303 (1993). The court also noted that the Cetus Corporation has published a bibliography listing more than a thousand scientific articles on PCR analysis. Lyons, at 608-09; see also Williams, 252 N.J. Super.at 383 ("hundreds” of scientific articles supporting PCR testing).
Russell next argues that the problems with using PCR in the forensic setting are so serious that this method of DNA testing fails to meet the standard of general acceptance required under Frye. The areas of concern that he identifies include differential amplification, misincorporation, and contamination.
Again, we note that whether these problems occur in the forensic setting does not affect the general scientific acceptance of PCR methodology. See Cauthron, at 887; Young, at 481. Rather, these problems bear on the question of the reliability of the individual test at issue as we found in Cauthron. The defendant in Cauthron cited commentaries discussing the possibilities of contaminated or degraded samples, improper enzyme applications, and human error in forensic RFLP testing. We concluded that while these problems were of concern, they did not require exclusion of RFLP evidence under Frye. Cauthron, at 898.
Once the general underlying principles are accepted, as they are here, then both the proponents and opponents of a particular test should be able to garner the necessary information to *51present both sides of the issue to the factfinder. Any remaining questions about the reliability of the particular tests in this case should be examined under the standards for admissibility of expert testimony, which is within the trial court’s discretion.
Cauthron, at 898. These standards are set forth in ER 702. Under this rule, the trial court has discretion to admit expert testimony if the witness qualifies as an expert and if the expert testimony would be helpful to the trier of fact. State v. Kalakosky, 121 Wn.2d 525, 541, 852 P.2d 1064 (1993); Cauthron, at 890.
It is important at this point to explain the relationship between Frye and ER 702 and how it affects this case. As stated earlier, the concern of Frye is whether the evidence being offered is based on generally accepted scientific theory and methodology. Cauthron, at 889. Frye is not concerned with the acceptance of the results of a particular study or of the particular testing procedures followed in the case before the court. Cauthron, at 889. These concerns are addressed under the ER 702 inquiry of whether the expert testimony would be helpful to the trier of fact. If the testing before the trial court shows that the testing procedure as performed was so flawed as to be unreliable, the results may be inadmissible because they are not helpful to the trier of fact. Kalakosky, at 541. If the evidence survives an ER 702 challenge, however, these questions then are considered by the trier of fact in assessing the weight to be given the evidence. See Kalakosky, at 543 (alleged infirmities in performance of test usually go to weight of evidence, not admissibility); Cauthron, at 899.
The problems that Russell now raises do not affect the general acceptance of the underlying PCR methodology. The possibility that differential amplification, contamination, or misincorporation affected the test results in a given case can be assessed by a trial court pursuant to ER 702, as the following discussion illustrates.5
*52A. Differential Amplification.
One authority sees no scientific basis for the belief that differential amplification or "allelic dropout” occurs with PCR testing. MacKnight, at 314. He explains that the phenomenon is possible, but improbable, as long as the equipment is properly calibrated and maintained. MacKnight, at 315. Another article concurs, concluding that the problem of allelic dropout with DQ alpha testing can be alleviated by performing the typing under appropriate conditions. Comey & Budowle, at 1633,1647. These conditions include avoiding the front two rows of the thermal cycler and using stringent denaturation conditions. Dr. Geraghty saw no problems with allelic dropout at the DQ alpha locus, and there is no claim that it occurred during the tests at issue.
B. Misincorporation.
This occurs when the Taq polymerase enzyme makes mistakes in pairing the bases that form the "rungs” of the DNA ladder. Misincorporation occurs very rarely (once or twice every 20,000 to 1 million bases) and usually at random. Office of Technology Assessment, U.S. Congress, Genetic Witness: Forensic Uses of DNA Tests 70 (1990); DNA Technology, at 64. One authority states that "misincorporation does not create problems in DNA sequencing analysis or probe typing of PCR-amplified DNA . . ., because the entire population of molecules is being examined, not a single molecule”. Genetic Witness, at 70. The article then adds, however, that misincorporation of nucleotides could be an issue if the initial amount of DNA is minute. Genetic Witness, at 70. Because there is no way to predict which PCR assays will be subject to misincorporation, each assay must be thoroughly characterized. DNA Technology, at 64. Dr. Geraghty testified that he saw no likelihood of a false result occurring because of misincorporation since it is easily detected, and added that he had never seen misincorporation occur "in the thousands of tests that have been done”. Verbatim Report on Appeal, at 229-31.
*53C. Contamination.
This risk can be minimized by strict adherence to sterile technique; the use of separate work areas for sample processing, solution preparation, amplification, and type testing; the use of separate pipettes in each area; and maintenance of a one-way flow of materials from the evidence-storage area to the sample-preparation area to the type-testing area. DNA Technology, at 66-67.6 (Drs. Blake and Geraghty testified to following these procedures in using the Cetus kit.) The study cited earlier conducted tests on several contaminated samples and found that the DQ amplification and typing system was "relatively unaffected by various environmental insults to bloodstains”. Comey & Budowle, at 1646. Moreover, the scientists found that the routine handling of evidence did not introduce contaminating DNA from the handler to the sample. Comey & Budowle, at 1646.
Thus, the potential testing problems that Russell cites are either detectable or preventable. Expert witnesses may assist the trial court in determining whether the testing procedures are so flawed that exclusion of the results are warranted under ER 702. If the results are admitted, the same experts can assist the trier of fact in determining what weight to give the test results in light of the perceived problems.
Adherence to proper laboratory procedure is essential in assessing the reliability of PCR test results and thus their admissibility under ER 702. As one commentator notes, it will be advisable for attorneys to be extremely familiar with the laboratory and with the person who conducts the PCR tests.
Thus, if the opponent to the test procedure raises issues regarding contamination, the well-prepared proponent of the evidence should be able to counter the arguments with specific descriptions, photographs or other documentation of the care and diligence with which samples are handled and tested in *54the laboratory. On the other side of the fence, if the opponent of the evidence is aware of sloppy technique, the lack of controls and/or unsuitable laboratory design which could foreseeably lead to contamination, this would be an important argument against the evidence.
MacKnight, at 322.
We see no question that the principles and methodology underlying PCR analysis at the DQ alpha locus have been generally accepted by the scientific community.7 We hold that the Cetus kit test results in this case were the product of generally accepted scientific theory and technique and were properly admitted under Frye. In so holding, however, we caution that this conclusion by no means assures the automatic admission of PCR DQ alpha test results. Serious flaws in a given test may render PCR evidence unreliable and thus inadmissible pursuant to ER 702. In seeking to admit PCR evidence, counsel must be prepared to establish adherence to proper laboratory procedures and protocols. It is also important to note that while both parties have assumed here that Frye requires the general acceptance of the Cetus kit as well as DQ alpha testing, that is not the case. The trial court properly determined that the issue here was the admissibility of PCR testing at the DQ alpha locus. The Cetus kit is one means by which such testing is performed; there undoubtedly *55will be others. The Cetus kit is simply one tool for carrying out generally accepted PCR methodology, and any concerns about its implementation in a given case are matters to be addressed to the trial court under ER 702. See MacKnight, at 306; DNA Technology, at 68-69.
Russell’s final issue on the admissibility of the test results relates to the Cetus ties of the forensic experts in this case and the degree to which they were "interested” witnesses. It is true that most of the State’s witnesses had some tie to Cetus. Dr. Reynolds was a Cetus employee, Dr. Von Beroldingen a former employer, and Dr. Blake a collaborator with Cetus in that he is paid by it to conduct workshops on PCR analysis.
A similar issue was raised in Barney, where a California Court of Appeal held that the self-interest underlying the testimony of FBI experts went to the weight to be attributed to their testimony rather than its admissibility. Barney, at 812; see also Williams, at 382 (evidence of financial rewards witness will gain from use of new scientific technique goes to jury in assessing weight to give testimony). We agree that the issue of self-interest is for the jury’s consideration, and we also observe that in the case at bar, the trial court appeared most impressed by the testimony of Dr. Geraghty, who had no financial ties to Cetus.
II
The second issue presented is whether article 1, section 9 of the Washington State Constitution requires suppression of evidence derived from a statement that Russell gave to the police without the benefit of Miranda warnings. Bellevue police arrested Russell on an outstanding warrant in order to talk to him about the murders. Russell explained to the police how he knew Pohlreich and Levine and what he was doing at the time of their murders. He also told the police he used Smith McLain’s pickup truck on the night of the Pohlreich murder. His statements led the police to two key witnesses in the Pohlreich case, McLain and his sister Shawn Calvo, and to the discovery of the bloodstains in the truck’s cab. Russell’s statement also included the names of a *56number of other witnesses and gave police enough information to eventually secure warrants authorizing the taking of samples of Russell’s blood, saliva, and hair.
Russell argued in pretrial hearings that the police failed to read him his Miranda rights prior to this questioning. He sought to have the trial court suppress not only the statement, but also the "fruit” of that statement.
The trial court concluded that the questioning occurred in a custodial setting, requiring police to read Miranda warnings. Due to conflicting testimony, the court found the State had failed to prove by a preponderance that Russell had been advised of his rights. The State was not permitted to use his statement in its case in chief.
The court found, however, that Russell’s statement was voluntarily given and that his "free will was not overborne”, noting that "the atmosphere in the interview room was relaxed and friendly”. Clerk’s Papers, at 423, 431A. Evidence also indicated that Russell had been "Mirandized” on previous occasions, including one occasion only 4 months before the questioning in this case. That Russell knew of his right to legal counsel was also clear since he eventually terminated the questioning by asking to speak to an attorney.
Because Russell’s statement was voluntarily given in a noncoercive atmosphere, the trial court ruled that evidence derived from his un-Mirandized statement would be admissible. Russell now challenges the admission of the "fruit” of his statement to the police under the Washington State Constitution.
We begin our analysis with a cursory discussion of the federal constitution in order to put the evaluation of our own state constitution in context.
The Fifth Amendment does not require the suppression of physical evidence derived from an un-Mirandized confession unless the statement was actually coerced. State v. Wethered, 110 Wn.2d 466, 473-75, 755 P.2d 797 (1988); see also State v. Putman, 61 Wn. App. 450, 456, 810 P.2d 977 (1991), adhered to on reconsideration, 65 Wn. App. 606, 829 P.2d 787 (1992), review denied, 122 Wn.2d 1015 (1993). *57Wethered makes clear that in the absence of coercion, the federal constitution requires suppression only of the unMirandized statement itself. See generally Wethered.8 Given the findings here, the trial court’s refusal to suppress the evidence derived from Russell’s statement was proper under the federal constitution.9
In light of the federal analysis, Russell seeks an independent interpretation of Const, art. 1, § 9 as it applies to this case. We therefore look to the factors outlined in State v. Gunwall, 106 Wn.2d 54, 61-62, 720 P.2d 808, 76 A.L.R.4th 517 (1986) in deciding whether the state constitutional provision extends broader rights than the federal constitution. Russell has provided an analysis of these factors.
The State contends that a resort to Gunwall is misplaced under State v. Earls, 116 Wn.2d 364, 805 P.2d 211 (1991). In Earls, this court addressed a similar argument regarding independent interpretation of Const, art. 1, § 9, and concluded as follows:
[R]esort to the Gunwall analysis is unnecessary because this court has already held that the protection of article 1, section 9 is coextensive with, not broader than, the protection of the Fifth Amendment. State v. Moore, 79 Wn.2d 51, 483 P.2d 630 (1971).
Earls, at 374-75. Thus, according to the State, no Gunwall analysis is necessary since this court has previously held that the two constitutional provisions are coextensive.
*58The State, however, takes this language out of context,10 thereby giving Earls an overly expansive interpretation and running afoul of an important principle of constitutional construction. A determination that a given state constitutional provision affords enhanced protection in a particular context does not necessarily mandate such a result in a different context. State v. Boland, 115 Wn.2d 571, 576, 800 P.2d 1112 (1990). Similarly, when the court rejects an expansion of rights under a particular state constitutional provision in one context, it does not necessarily foreclose such an interpretation in another context.
In Boland, this court set out the analysis to apply where the state provision in question has been analyzed in a different context. The previous case will have already analyzed the first, second, third and fifth Gunwall factors* as these factors arise whenever the two constitutional provisions are compared. The fourth and sixth factors, however, are generally unique to the context in which the interpretation question arises. The cqurt thus examines the fourth and sixth factors in light of the new context presented, and then takes these two factors into account along with those previously analyzed. See Boland, at 576.
Because Earls and Moore did not pertain to the admissibility of the fruits of an un-Mirandized confession, these cases do not preclude the possibility of an independent interpretation of Const, art. 1, § 9 in this context.* 11 While a full Gunwall analysis would not otherwise be required, we must fully analyze all six factors, for unlike the situation in Boland, we are not presented with any previous cases in *59which a majority opinion has evaluated all of the Gunwall factors. We will, however, take into account, where relevant, the analysis contained in the Earls and Moore decisions.
1
Factors 1 and 2 — Constitutional Texts
Under the state constitution "[n]o person shall be compelled in any criminal case to give evidence against himself . . .”. (Italics ours.) Const, art. 1, § 9. The parallel federal provision states "nor shall [any person] be compelled in any criminal case to he a witness against himself . . .”. U.S. Const, amend. 5. Thus, the difference is between "giving evidence” (state constitution) and "being a witness” (federal constitution).
This court has already held that this difference in language is without meaning. See Moore, at 55-57. In Moore, and again in Earls, this court concluded that the purpose of each provision was the same: to prohibit the compelling of self-incriminating testimony from a party or witness. Moore, at 56; Earls, at 376.
The dissents in both Moore and Earls pointed out that the framers of the state constitution had originally drafted a provision using language similar to that found in the federal constitution, but instead adopted the different "give evidence” language. The dissent in each case concluded this change in language signified that the state framers intended an independent interpretation of the state constitution. Moore, at 65 (Rosellini, J., dissenting); Earls, at 390-91 (Utter, J., dissenting). This argument was rejected by the majority in each case — in Earls, by a vote of 8 to 1.
2
Factor 3 — State Constitutional and Common Law History
The Earls majority opinion did not discuss this point. The Earls dissent pointed out that article 1 of the state constitution was based primarily on language from other states’ constitutions rather than from the federal constitution. The dissent thus concluded that the framers intended an independent interpretation. Earls, at 391 (Utter, J., dissenting).
*60This recognition certainly has some force to it, but it would have much greater significance if the framers had used language that differed in any great degree from that used in the federal constitution. The court has not been presented with any evidence suggesting that the framers of these "model” state constitutions intended any different result than that reached under the federal constitution.
3
Factor 4 — Preexisting State Law
Russell contends that Washington courts have historically held that in the absence of Miranda warnings a statement is necessarily coerced. As support, he cites State v. Lavaris, 99 Wn.2d 851, 664 P.2d 1234 (1983), where this court stated that
[a]ny form of custodial interrogation is inherently coercive. Therefore, any confession obtained in the absence of proper Miranda warnings is by definition "coerced” — regardless of how "friendly” the actual interrogation.
(Citations omitted.) Lavaris, at 857. The Defendant also cites two Court of Appeals cases holding that the fruits of un-Mirandized statements must be suppressed. State v. Marko-vich, 17 Wn. App. 809, 814-15, 565 P.2d 440 (1977) (holding that a gun found as a result of the defendant’s un-Mirandized statement should have been suppressed), review denied, 89 Wn.2d 1015 (1978); State v. Galloway, 14 Wn. App. 200, 202, 540 P.2d 444 (holding that amphetamines found as a result of the defendant’s un-Mirandized statement should have been suppressed), review denied, 86 Wn.2d 1006 (1975). We also note an earlier decision of this court holding that a finding of voluntariness is precluded if Miranda warnings are not given. State v. Creach, 77 Wn.2d 194, 199, 461 P.2d 329 (1969).
We acknowledge this preexisting law, but we are persuaded by some important countervailing considerations. The cases Russell cites have all involved interpretation of Miranda, a federal judicial decision, and we have never held that Miranda warnings are independently required under *61the state constitution.12 Thus, the "state law” cited by Russell is to a large degree based on federal law. An additional consideration is that the holdings upon which Russell relies have been around only since 1969 and were supplanted by Wethered in 1988; they do not represent long-held principles of law.
4
Factor 5 — Structural Differences of the Constitutions
The state constitution limits powers of state government, while the federal constitution grants power to the federal government. Gunwall, at 66. This difference favors an independent state interpretation in every Gunwall analysis.
5
Factor 6 — National Versus State or Local Concerns
Russell contends that Washington has historically applied the exclusionary rule broadly. As support he cites State v. Bonds, 98 Wn.2d 1, 10, 653 P.2d 1024 (1982), cert. denied, 464 U.S. 831 (1983), where this court summarized the ways in which Washington law has generally "extended the exclusionary rule beyond the original Fourth Amendment context”. For example, this court adopted an exclusionary rule decades before the requirement was extended to the states in Mapp v. Ohio, 367 U.S. 643, 6 L. Ed. 2d 1081, 81 S. Ct. 1684, 84 A.L.R.2d 933 (1961), and Washington has extended the rule (sometimes judicially and sometimes legislatively) to statutory violations as well. Bonds, at 9-10. Moreover, as the Earls dissent argues, criminal law in general involves local, not national, concerns. See Earls, at 396-97 (Utter, J., dissenting).
As with the fourth Gunwall factor, Russell’s position does not tell the entire story. First, Russell gives no indication that the exclusionary rule at issue, suppression of the fruits of an un-Mirandized confession, has historically been applied more broadly under state law than under federal law. As discussed *62above, the specific exclusionary rule here at issue is peculiarly federal in nature. It is based on a federal case (Miranda v. Arizona, 384 U.S. 436, 16 L. Ed. 2d 694, 86 S. Ct. 1602, 10 A.L.R.3d 974 (1966)) interpreting the federal constitution. Moreover, this court has not held that Miranda (or similar) warnings are required independently under the state constitution. Thus, this case involves a national issue to a greater extent than do many other issues of criminal law.
On balance, we conclude that the Gunwall factors do not support extending greater protection through Const, art. 1, § 9 than that provided by the federal constitution in the present context.
What Russell argues is essentially policy. He maintains that the State should not be allowed to use evidence derived from un-Mirandized statements, regardless of the surrounding circumstances. Policy considerations alone are insufficient, however, to trigger'an expansive reading of Const, art. 1, §9.
Ill
The third issue raised is whether the trial court erred in denying Russell’s motion to sever count 1 from counts 2 and 3. The defense made a pretrial motion to sever these counts and renewed the motion at the close of the State’s case. The court denied both motions and all three counts were joined. Russell now contends that these denials were error because the prejudice of joinder far outweighed any considerations of judicial economy.
CrR 4.3(a) authorizes joinder of counts where the offenses are of the same or similar character. Joinder of counts should never be used in such a way as to unduly embarrass or prejudice a defendant or deny him or her a substantial right. State v. Smith, 74 Wn.2d 744, 754-55, 446 P.2d 571 (1968), vacated in part, 408 U.S. 934 (1972), overruled on other grounds by State v. Gosby, 85 Wn.2d 758, 539 P.2d 680 (1975). Prejudice may result from joinder if the defendant is embarrassed in the presentation of separate defenses, or if use of a single trial invites the jury to cumu-*63late evidence to find guilt or infer a criminal disposition. State v. Watkins, 53 Wn. App. 264, 268, 766 P.2d 484 (1989) (citing Smith, 74 Wn.2d at 754-55). In determining whether the potential for prejudice requires severance, a trial court must consider (1) the strength of the State’s evidence on each count; (2) the clarity of defenses as to each count; (3) court instructions to the jury to consider each count separately; and (4) the admissibility of evidence of the other charges even if not joined for trial. Smith, 74 Wn.2d at 755-56; State v. York, 50 Wn. App. 446, 451, 749 P.2d 683 (1987), review denied, 110 Wn.2d 1009 (1988). In addition, any residual prejudice must be weighed against the need for judicial economy. State v. Kalakosky, 121 Wn.2d 525, 539, 852 P.2d 1064 (1993); State v. Bythrow, 114 Wn.2d 713, 723, 790 P.2d 154 (1990); State v. Markle, 118 Wn.2d 424, 439, 823 P.2d 1101 (1992). On review, a trial court’s refusal to sever charges is reversible only where it constitutes a manifest abuse of discretion. Markle, at 439; York, at 450. The defendant bears the burden of demonstrating such abuse. Bythrow, at 720; York, at 450.
In considering the strength of the State’s evidence on each count, the trial court found that the strength of counts 1 and 3 was similar and that count 2 was a weaker case. The defense disputes this conclusion. In assessing this contention, we need only address the evidence presented on counts 1 and 3 since the defense never moved to sever count 2 from count 3.
With regard to count 3, the State’s evidence indicated that Russell knew Levine and knew where she lived. On the morning of Levine’s murder, Russell was staying in a motel that was less than a 15-minute walk from Levine’s residence. Early in the morning, Russell left the motel. At approximately 5 a.m., an intruder was seen leaving Levine’s residence. The intruder, like Russell, was wearing dark clothing. An eyewitness saw something white, about two-thirds as wide as the person, superimposed against the chest or abdomen of the intruder. Russell was wearing a dark blue sweatshirt with a white logo on the front.
*64After Levine’s murder, Russell told some of his friends that she was a "whore” , who had been "sleeping with” a friend of his and who "used men”. Verbatim Report on Appeal, at 5463-64, 5360-61. A hair found on Levine’s body after her murder was similar to Russell’s. Finally, and perhaps most importantly, it was discovered that a distinctive ring which Levine wore regularly was missing from her apartment. Less than 1 week after Levine’s murder, Russell gave a similar ring to an acquaintance. Witnesses identified the ring as belonging to Levine.
In count 1, the evidence showed that Russell borrowed a truck the evening before the Pohlreich murder and returned it the next day with the explanation that he had vomited in it. The truck’s upholstery and carpeting later revealed blood and semen that PCR testing showed could be connected to Pohlreich and to Russell. Fibers found on Pohlreich’s body matched fibers from the carpet and upholstery in Smith McLain’s truck. Blood testing revealed that the blood in the upholstery could not have been Russell’s. The evidence also showed that Russell knew Pohlreich and left Papagayo’s with someone who matched her general description.
The trial court concluded that the relative strengths of counts 1 and 3 were not sufficiently dissimilar to merit severance, especially given the relatively low power of discrimination inherent in PCR testing. The trial court specifically indicated that had the DNA testing been with RFLP, severance would have been warranted because of its high power of discrimination.
We see no abuse of discretion in the court’s conclusion. We agree the strengths of counts 1 and 3 were not dissimilar, with evidence of the ring in count 3 perhaps equal in strength to the PCR evidence in count 1. Neither piece of evidence pointed with certainty to the Defendant, but each was significant.
The second factor to consider is whether the clarity of defenses to each count was prejudiced by joinder. The likelihood that joinder will cause a jury to be confused as to the accused’s defenses is very small where the defense is identical on each charge. State v. Hernandez, 58 Wn. App. 793, *65799, 794 P.2d 1327 (1990), review denied, 117 Wn.2d 1011 (1991). Russell’s defense on each count in this case was denial. The trial court recognized that the defense would attempt to portray count 2 as a domestic homicide, with Beethe’s boyfriend Mike Suell as the perpetrator, but did not see this as significant. "It isn’t as though there will be a self-defense argument on one and a different type of defense on another one, or that there will be an admission of one or denial of another.” Verbatim Report on Appeal, at 2067.
The defense argues that joinder made it impossible to defend in count 2 by portraying that murder as a domestic homicide because the argument did not apply to the Pohl-reich count. We note, however, that this defense would not have worked well even if count 1 were severed from counts 2 and 3 because Suell could not have been the source of the Negroid hairs found on Beethe’s bed in count 3.
Russell also claims joinder prejudiced his ability to present separate defenses because, had the counts been severed, he probably would have testified on counts 2 and 3 but not on count 1. Joinder, Russell claims, dissuaded him from testifying.
A defendant’s desire to testify only on one count requires severance only if a defendant makes a "convincing showing that she has important testimony to give concerning one count and a strong need to refrain from testifying about another”. Watkins, at 270; State v. Weddel, 29 Wn. App. 461, 467, 629 P.2d 912, review denied, 96 Wn.2d 1009 (1981). In the instant case, Russell made no offer of proof as to the content of the anticipated testimony. In denying the motion to sever, the trial court made the following finding:
Although there has been some recent suggestion that the defendant might elect to testify on one count and not on the others, there has been no offer of proof as to which count the defendant might elect to testify about, no offer of proof as to what he might say, and no showing that he would be prejudiced by any decision he might make regarding his decision to testify on any count or counts and not on another.
Clerk’s Papers, at 407. Absent an offer of proof, it is difficult to conclude that joinder affected Russell’s decision not to *66testify. We do not find the clarity of the Defendant’s defenses prejudiced by joinder under the facts presented.
The third factor to consider is whether the court properly instructed the jury to consider each count separately. The defense now claims that the trial court should have instructed the jury to "decide separately what the evidence in the case shows about the crime”. See United States v. Johnson, 820 F.2d 1065, 1071 (9th Cir. 1987). Since the defense never proposed such an instruction, however, and since the instruction it did propose is both the one that the trial court gave and a correct statement of the law, we find no error. See Harris v. Groth, 99 Wn.2d 438, 447, 663 P.2d 113 (1983); Hoglund v. Raymark Indus. Inc., 50 Wn. App. 360, 368, 749 P.2d 164 (1987), review denied, 110 Wn.2d 1008 (1988).
The final factor is whether evidence of each count would be cross admissible under ER 404(b) if severance were granted. ER 404(b) permits evidence of other crimes to show identity, motive, intent, preparation, plan, knowledge, absence of mistake or accident, opportunity, or an alternative means by which a crime could have been committed. State v. Lord, 117 Wn.2d 829, 872 n.11, 822 P.2d 177 (1991), cert. denied, 113 S. Ct. 164 (1992). Such evidence is not admissible "to prove the character of a person in order to show action in conformity therewith”. ER 404(b); State v. Smith, 106 Wn.2d 772, 775, 725 P.2d 951 (1986). Of relevance here is identity. In determining the admissibility of other crimes to prove identity, a trial court must determine that the evidence is relevant to identity and that any prejudicial effect is outweighed by the probative value: It must then properly limit the purpose for which the jury may consider the evidence. Smith, 106 Wn.2d at 772; Watkins, at 270.13
Evidence of other crimes is relevant on the issue of identity only if the method employed in the commission of *67both crimes is "so unique” that proof that an accused committed one of the crimes creates a high probability that he also committed the other crimes with which he is charged. Hernandez, at 799 (citing Smith, 106 Wn.2d at 777). In other words, the device used must be so unusual and distinctive as to be like a signature. State v. Coe, 101 Wn.2d 772, 777, 684 P.2d 668 (1984) (citing McCormick’s Evidence § 190, at 449 (Edward W. Cleary gen. ed., 2d ed. 1972)); see also State v. Lynch, 58 Wn. App. 83, 88, 792 P.2d 167, review denied, 115 Wn.2d 1020 (1990).
The trial court found cross admissibility on the basis of signature and entered the following written findings:
6. Each crime bears the perpetrator’s unique signature, comprised of the manner in which these women were killed, the elaborate manner in which they were posed after their deaths, and the proximity in time and place of the three murders.
7. Evidence of each crime is highly probative of the identity of the murderer in each of the other crimes; the probative value of this evidence greatly outweighs its prejudice to the defendant.
Clerk’s Papers, at 407. The trial court also cited the opinions of two criminal investigators who testified during the severance hearing that the three homicides bore the same signature and were committed by the same person.
Russell now contends that the factors identified in finding of fact 6 do not meet the test for signature crimes, and that no unique signature was identified by the trial court. The State counters by citing other "signature” cases in which the criminal methods were less distinctive than those employed here. See State v. Laureano, 101 Wn.2d 745, 682 P.2d 889 (1984) (evidence of prior robbery admissible under ER 404(b) where crimes committed 3 weeks apart where both involved forcible entry into family residences by three persons dressed in army fatigues (though not the same three) and where both involved firearms and similar use of a shotgun); see also Lynch (two prior robberies admissible where all crimes involved wearing a brown wig, similar time of day, a red 10-speed bicycle, display of a gun tucked in a waistband, and theft of car keys from victims); but see Hernandez, at *68799 (no showing of unusual or unique manner sufficient to show identity where robber entered the store, pulled a knife, asked for money and fled upon receiving it).
We find that the factors cited by the trial court support the finding that certain evidence in this case was quite unique. Each count involved a victim killed by violent means who was then sexually assaulted and posed, naked, with the aid of props. The murders occurred within a few weeks of one another in a small geographic area. We agree with the trial court and with the expert witnesses that these similarities were not due simply to coincidence. Accordingly, we do not regard the trial court’s conclusion as to cross admissibility as an abuse of discretion.
Finally, the court must weigh any prejudice to the defendant resulting from joinder against the need for judicial economy. The trial court found that apart from the evidence of signature, "a great deal of evidence, particularly from those witnesses who were acquainted with the defendant during the period of time in which these crimes were committed, would be repeated in each trial if the defendant’s motion to sever Count I from the others is granted”. Clerk’s Papers, at 407-08; The court thus concluded that judicial economy was served by a single trial on all counts.
Excluding pretrial motions, this case took 33 days to try. Considering our evaluation of the other severance factors we cannot find that the prejudice resulting from joinder in this case outweighed considerations of judicial economy. Accordingly, we conclude that the trial court did not abuse its discretion in denying Russell’s motion for severance.
IV
The fourth issue the Defendant raises is whether the trial court erred in admitting expert and lay testimony regarding the rarity of posed murder victims. Russell here raises several arguments regarding this testimony. We will first address Russell’s contention that the expert testimony was inadmissible because the State’s experts improperly relied on unproved scientific methodologies in determining that the same person committed all three murders.
*69At issue here are references made by John Douglas and Robert Keppel to the HITS and VICAP computer programs during their testimony regarding the rarity of posing.14 These programs use forms, filled out by local law enforcement officers, listing the various characteristics of homicides in Washington and the nation, respectively. The trial court found that the expert testimony referring to HITS and VI-CAP did not involve novel scientific evidence and was, therefore, subject only to the requirements of ER 702.
As stated earlier, expert testimony is admissible under ER 702 if the witness qualifies as an expert and if the expert testimony would be helpful to the trier of fact. Kalakosky, at 541; State v. Cauthron, 120 Wn.2d 879, 890, 846 P.2d 502 (1993). Testimony which does not involve new methods of proof or new scientific principles from which conclusions are drawn need not be subjected to the Frye test. State v. Ortiz, 119 Wn.2d 294, 311, 831 P.2d 1060 (1992); State v. Young, 62 Wn. App. 895, 906, 802 P.2d 829, 817 P.2d 412 (1991). Decisions based on ER 702 are reviewed under the abuse of discretion standard. Kalakosky, at 541.
In the case at bar, the trial court ruled that both Keppel and Douglas were widely recognized as authorities in crime scene analysis. Both men have extensive experience in serial crime analysis and investigation. The court then found that their testimony would not involve the application of a new scientific technique and that a Frye hearing was unnecessary. Finally, the court ruled that the testimony concerning the rarity of posing would be helpful to the jury under ER 702:
The jury does not have the specialized knowledge of how common the problem is or how often there is sexual penetration, open display of bodies, or the posing of the body after death. So I think it is within the scope of an opinion of somebody’s experience to indicate whether these are common or not common or unique. I would find that the relevance of the testimony, as it goes to the identity of the perpetrator, and the inference to be drawn, is that the same person committed all three homicides.
Verbatim Report on Appeal, at 2331.
*70We agree with the trial court that the Frye test clearly was inapplicable to the expert testimony regarding the HITS and VICAP programs. These programs are nothing more than sophisticated record-keeping systems. The court correctly analyzed the admissibility of this testimony under ER 702 and we find no abuse of discretion in the admission of the experts’ testimony.
Russell also objects to this testimony on the ground that it was statistical. Neither expert expressed his opinion about the rarity of posing in precisely quantified terms, though Douglas testified ás to the number of cases on VICAP and Keppel testified as to the number of cases on HITS. Russell maintains, however, that by specifying the extent of these databases, Keppel and Douglas implicitly testified that Russell was guilty as a matter of mathematical probability.
We first note that there is no prohibition against using well-founded statistics to establish some fact that will be useful to the trier of fact. State v. Briggs, 55 Wn. App. 44, 62-63, 776 P.2d 1347 (1989) (citing People v. Collins, 68 Cal. 2d 319, 332, 438 P.2d 33, 66 Cal. Rptr. 497 (1968)). Second, both experts relied on the databases primarily as support for the conclusion that posing is a rare occurrence and not for the conclusion that there was a statistical probability that Russell committed the murders. Finally, both experts relied more on case materials and personal expertise than on the databases in forming their opinions and both expressed their opinions in nonquantifiable terms.
Russell next contends that the trial court erred in allowing three lay witnesses to testify about the rarity of posing and thus to reinforce the expert testimony. In addition to Keppel and Douglas, three detectives testified that each of the bodies seemed posed and only one said he had ever seen another murder scene involving posing.
This court recently explained the appropriate conditions for admissibility of lay testimony as follows:
Under Rule 701 and Rule 602, the witness must have personal knowledge of matter that forms the basis of testimony of opinion; the testimony must be based rationally upon the *71perception of the witness; and of course, the opinion must be helpful to the jury (the principal test).
Ortiz, at 308-09 (citing McCormick’s Evidence 29 (Edward W. Cleary gen. ed., 3d ed. 1984)).
We have already concluded that the court did not abuse its discretion in finding testimony regarding the rarity of posing helpful to the trier of fact. The detectives testified regarding their personal knowledge of crime scenes, and their perceptions of the three murders at issue. While the testimony may have been cumulative, we do not see that its admission rises to the level of an abuse of discretion.
The defense argues further that the State’s expert testimony amounted to expert opinion on the ultimate question of guilt. The State points out, however, that if the testimony was improper, the defense opened the issue when it presented testimony that the murders were not related.
These arguments bring us once again to Russell’s challenge that this expert testimony was improperly admitted under ER 404(b) to show identity and thus to prove that the same person committed all three murders. In her pretrial ruling the judge limited the experts by allowing them to testify only that the criminal methods employed in each case were unique and rose to the level of signature evidence. The experts were precluded from testifying that they thought that the same person committed all three crimes.
During the course of trial, the court modified this ruling when defense counsel, during cross examination of State’s expert Keppel, asked whether he knew of any other cases related to these three murders. The State objected because the line of questioning went beyond the scope permitted by the court. The court told the defense:
You may do that. Then they have the right to come back on redirect and ask if these three cases are related, which they have not done. They have not gone into the relationship among the three cases. If you want to relate . . . the Pohlreich case to other cases, then I think in all fairness they can come back and relate the three cases to each other.
*72Verbatim Report on Appeal, at 5770. When the defense pursued the issue of the similarities and differences among these three murders and others, the court allowed Keppel to testify on redirect that in his opinion all of the murders were committed by the same person. Keppel based his opinion on the posing and on the facts that all the victims were nude, all were female, and all were killed within a short period of time of their contact with the offender. He also observed that each crime involved the sexual insertion of a foreign object and that the offender needed to display these victims and ensure their discovery. Russell did not raise an ER 404(b) objection to this line of testimony.
Later, the State’s other expert, John Douglas, also testified that all of the victims were posed and that all of the murder scenes exhibited the same signature. Douglas based his opinion regarding signature on the facts that all of the victims had been posed in degrading and humiliating positions and on the fact that the murders occurred within a 67-day period within a small geographical area. Again, Russell raised no ER 404(b) objection but instead, in cross examination, sought to emphasize the differences between Pohl-reich’s murder and the other two murders. Douglas agreed that there were differences among the crimes, but explained on redirect that the differences were insignificant compared to the similarities. "The significant part is the posing of the victims, the posing in this degrading type of position, that is critical.” Verbatim Report on Appeal, at 6042.
The defense then called Robert Gebo as an expert witness who testified that, at one point, he had believed that the Pohlreich and Beethe murders were not connected. On cross examination, the prosecution elicited Gebo’s current opinion that all of the murders were committed by the same person. The court overruled defense counsel’s objection and motion to strike, and again ruled that Gebo’s testimony on direct brought the issue of whether the homicides were related onto the field of play.
In considering Russell’s contention that these experts improperly gave opinions on the ultimate question of guilt, *73we observe that the purpose of showing identity under ER 404(b) is to demonstrate the probability that the same person committed the crime. Coe, at 777-78; Smith, 106 Wn.2d at 778. Having found the expert testimony admissible to show identity, we will not rule inadmissible the inference to be drawn from such evidence. Moreover, the express assertions that the same person committed the three murders were invited by defense counsel. Once the defense brought up the issue of whether these crimes were related to other crimes, the court properly ruled that the State could, in turn, ask the experts whether the crimes were related to one another and had been committed by one person. See State v. Gefeller, 76 Wn.2d 449, 455, 458 P.2d 17 (1969); State v. Crenshaw, 27 Wn. App. 326, 333, 617 P.2d 1941 (1980), aff'd, 98 Wn.2d 789 (1983). We find no error in the court’s ruling.
In a related argument, the defense complains that Gebo improperly based his change of opinion on the totality of the evidence against Russell, and that his testimony thus doubly invaded the province of the jury. The trial court cured any error in this regard by allowing the State to call Gebo as a rebuttal witness so that he could explain that his change of opinion was based not on information relating to Russell but on the views of his colleagues, Robert Keppel and John Douglas. The defense now claims that Gebo was lying. We have no way to assess this claim and will not consider it further.
Russell next contends that the trial court’s denial of discovery of police reports from which the HITS data is drawn violated his Sixth Amendment right of confrontation. The Sixth Amendment and Const, art. 1, § 22 (amend. 10) grant criminal defendants the right to confront and cross-examine adverse witnesses. State v. Hudlow, 99 Wn.2d 1, 15, 659 P.2d 514 (1983); State v. Boast, 87 Wn.2d 447, 453, 553 P.2d 1322 (1976).
ER 703 governs the bases of opinion testimony given by experts:
The facts or data in the particular case upon which an expert bases an opinion or inference may be those perceived by or made known to the expert at or before the hearing. If of a *74type reasonably relied upon by experts in the particular field in forming opinions or inferences upon the subject, the facts or data need not be admissible in evidence.
ER 705, in turn, governs the disclosure of the facts underlying an expert’s opinion and provides as follows:
The expert may testify in terms of opinion or inference and give reasons therefor without prior disclosure of the underlying facts or data, unless the judge requires otherwise. The expert may in any event be required to disclose the underlying facts or data on cross examination.
ER 703 thus permits expert opinion testimony based on hearsay data that would be otherwise inadmissible in evidence, while ER 705, which is identical to Federal Rule of Evidence 705, authorizes the admission of expert opinion testimony without the prior disclosure of the facts or data which underlie the opinion. See 11 James W. Moore & Helen I. Bendix, Federal Practice § 705.10, at VII-70 (2d ed. 1976); see also Robert H. Aronson, Evidence in Washington 705-3 (2d ed. 1993) (under ER 705, trial court has discretion to require an expert to disclose the basis for opinion).
Initially, the defense requested access to the HITS system and specifically disavowed any request for police reports of cases included in that system. The State offered to allow a reputable defense expert to review the data with Keppel, and Keppel reviewed the database with defense counsel for several hours. The trial court granted the defense’s discovery request for forms in the relevant categories, e.g., posing/ unusual position, sexual insertion, ritual. After access to the HITS system was provided, the defense requested a variety of police reports referenced in HITS. The defense argues that the trial court’s denial violated the confrontation clause.
Keppel himself did not have access to the actual police reports in the HITS system. The trial court reasoned that meaningful cross examination was possible without access to the police reports by pointing out that Keppel was basing his conclusions on data interpreted by another person.
Other Washington cases have allowed the admission of expert opinion based on data interpreted by another. *75See State v. Ecklund, 30 Wn. App. 313, 318, 633 P.2d 933 (1981); Tennant v. Roys, 44 Wn. App. 305, 311, 722 P.2d 848 (1986). We see no abuse of discretion or violation of the confrontation clause resulting from a similar admission here. The State allowed the defense to review the same data on which Keppel relied. Keppel himself did not have access to the actual police reports that make up the databases; moreover, he did not base his opinions exclusively on the databases. We hold that Russell was not denied due process and find no abuse of discretion in the trial court’s denial of discovery of the police reports underlying the HITS system.
V
Russell next argues that the trial court erred in excluding evidence of other suspects and other crimes which he contends would connect another person with the three murders at issue.
Washington law on this point is clear:
While evidence tending to show that another party may have committed the crime may be admissible, before such testimony can be received there must be such proof of connection ... or circumstances as tend clearly to point out someone besides the one charged as the guilty party.
State v. Kwan, 174 Wash. 528, 532-33, 25 P.2d 104 (1933) (citing State v. Downs, 168 Wash. 664, 667, 13 P.2d 1 (1932)); see also State v. Mak, 105 Wn.2d 692, 716, 718 P.2d 407, cert. denied, 479 U.S. 995 (1986). The trial court’s decision on admissibility of this type of evidence is reviewed for abuse of discretion. Mak, at 717.
Russell asserts initially that the trial court erred in excluding evidence of two other assaults that occurred in September 1990. The defense states that the incidents were similar to the Beethe and Levine counts because the female victims lived on the ground floor of Bellevue apartment buildings and both received head injuries inflicted by a heavy object. A review of the facts, however, indicates substantial differences among the incidents.
The first assault occurred early in the morning on September 9, 1990. The victim had locked herself out of her *76apartment building and was banging on her door when a man approached, offering his help. The two walked around to the back of the building to see if she could climb in a window. The man struck the victim on the head with what appeared to be a rock; she screamed and he fled.
The trial court found no obvious similarities between injuries inflicted upon this victim and upon Beethe. Furthermore, the court noted that Russell had no alibi for the time of the assault, and could have been the assailant.
The second assault occurred on September 15,1990, when the victim surprised a burglar in her home. The burglar struck her once in the eye and then fled with money. Although Russell had an alibi in this case because he was in custody, the court found no similarity between either this assault and the one described above or between this assault and the murders at issue.
Russell also assigns error to the trial court’s refusal to admit evidence that two men, George Grumbs and Brent Carlson, may have murdered Pohlreich and Levine, respectively. George Grumbs was considered a suspect in a rape/ murder case where the body was left in a parking lot. The victim in that case was an African-American with a history of prostitution and a heroin habit. She had been seen driving with Grumbs, and although her body was left in a parking lot, it was fully clothed and not posed. The trial court found that the Grumbs case bore little resemblance to the Pohlreich case.
Brent Carlson had had a previous romantic relationship with Levine, and the two were friends up to her death. Carlson initially offered little in the way of an alibi for the night of Levine’s death. Later he admitted that he had been with another woman the night of the murder, but he could not produce the woman. The trial court concluded that neither the relationship between Carlson and Levine nor the lack of an alibi was sufficient to bring this evidence into the case. "The case law seems to focus on having more than motive and opportunity.” Verbatim Report on Appeal, at 2261.
We agree. As this court stated shortly after Downs was decided,
*77[m]ere evidence of motive in another party, or motive coupled with threats of such other person, is inadmissible, unless coupled with other evidence tending to connect such other person with the actual commission of the crime charged.
Kwan, 174 Wash. at 533. We see no evidence connecting either George Grumbs or Brent Carlson to the present case. Nor do we see any connection between the three murders and the two assaults described earlier. Accordingly, we conclude that the trial court did not abuse its discretion in refusing to admit evidence of other crimes and suspects.
VI
Russell also challenges the trial court’s rulings admitting into evidence unused condoms, a police scanner, and a crime scene handbook. Russell argues that the admission of each item violated Evidence Rules 401, 403, and 404.
A. Condoms.
Prior to trial Russell moved to exclude unused condoms found in his belongings. The State asserted that the condoms were relevant to rebut an argument that Carol Beethe and Andrea Levine had been killed by someone other than Mary Pohlreich’s killer, because semen was found at the scene of Pohlreich’s murder but not at Beethe’s and Levine’s. (The State intended to argue that the absence of semen in the bodies of Levine and Beethe could be attributed to the use of condoms.)
The trial court concluded the condoms had little probative value, as they are easily obtained, and determined this value was outweighed by the condoms’ prejudicial effect. The court indicated, however, that it might reconsider its decision if the defense presented this theory at trial. The State objected that jurors might make the inference regarding different killers even if the defense did not directly present it to them.
The next day, the trial court reversed its decision and admitted the condoms. The trial court noted that during the voir dire of potential jurors, none had found the possession of condoms unusual or offensive, and none believed it showed any intent to commit a sexual offense. The trial court concluded the condoms would not prejudice these *78jurors against the Defendant and found the condoms relevant in rebutting the potential inference of separate killers.
Russell contends that the condoms were not relevant to any issue in the case. He points out that the State did not have to prove intercourse, that the condoms obviously had not been used in the murders, and that condoms are widely available. He argues further the condoms raised two impermissible inferences: that Russell was not celibate, and that Russell kept condoms in order to rape and murder women without leaving evidence.
A trial court’s evaluation of relevance under ER 401 and its balancing of probative value against prejudicial effect under ER 403 will be overturned only for manifest abuse of discretion. See State v. Rice, 110 Wn.2d 577, 598-600, 757 P.2d 889 (1988), cert. denied, 491 U.S. 910 (1989); State v. Harris, 106 Wn.2d 784, 791, 725 P.2d 975 (1986), cert. denied, 480 U.S. 940 (1987). Discretion is abused only when no reasonable person would have decided the issue as the trial court did. Rice, at 600.
We find no basis for overturning the court’s ruling here. Although the balancing of the probative and prejudicial aspects of the condoms is close, the trial court’s decision was within the bounds of appropriate discretion. While the evidence was not highly relevant, neither was it highly prejudicial in light of the jurors’ responses to voir dire. Since a reasonable person could conclude under these circumstances that the prejudicial nature of this evidence did not outweigh its probative value, we find no abuse of discretion under ER 401 and ER 403. Moreover, because the condoms related to the identity of the killer or killers, we find no violation of ER 404(b).
B. Scanner.
Russell possessed a police scanner until the police took it away from him a few months prior to the killings. The scanner issue arose during presentation of the defense case. One of the defense theories was that Russell was working undercover for the police. The defense sought to introduce the following testimony from one of Russell’s roommates: The *79roommate had called the police after noticing a man lurking outside; shortly thereafter, Russell returned to the residence saying he was angry she had called the police; and Russell said he had been in a police car when the call came in. Defense counsel said she wanted to introduce this evidence to show Russell was working with the police.
The State countered that if this evidence were admitted, it would seek to introduce Russell’s previous ownership of the scanner to show that Russell could have learned of the police call even if he were not employed by the police.
The court stated, "Either I am going to admit [Defendant’s proposed testimony] and let the state go ahead with access to the police scanner, or I’m going to exclude it.” Verbatim Report on Appeal, at 5385. Defense counsel responded by indicating that "our position would be to let the evidence in and to let the state adduce evidence that he was in possession of a scanner in May”. Verbatim Report on Appeal, at 5385. The trial court then allowed each party to present its evidence.
Russell now contends admission of the scanner was improper. He argues that the evidence was not related to the elements of any of the charged crimes, that the scanner does not make any more probable his guilt in these charges, and that the evidence prejudicially implied that he had a criminal disposition.
The scanner was relevant in rebutting the testimony proposed by defense counsel. Moreover, while this evidence has the potential for creating unfair prejudice, the prejudicial eifect was markedly reduced in this case since the Defendant claimed to be working for the police. Also, the prosecutor elicited testimony informing the jury that possession of a scanner is not illegal. The trial court’s decision was well within the range of discretion and we conclude that the evidence was properly admitted under ER 401 and 403.
As to ER 404(b), the scanner was admitted for the purpose of explaining Russell’s opportunity to learn of the police call other than through employment with the police. Because the evidence was not admitted to prove character, ER 404(b) was not violated.
*80C. Crime Scene Handbook.
After Russell was questioned by the police, he telephoned one of his roommates and asked her to give the police his copy of Crime Scene Search and Physical Evidence Handbook. The handbook outlines police procedures in gathering evidence from the scene of a crime and contains chapters on fingerprints and body fluids. The book does not contain a chapter on DNA testing.
In a pretrial hearing, the State argued the handbook was relevant because it showed knowledge of techniques that were apparently used by the killer of Beethe and Levine. The trial court initially ruled the handbook inadmissible, but noted it would reconsider its decision if circumstances changed. The issue was raised again during trial. The State pointed out that the Beethe murderer had worn gloves; that Levine’s bedroom had been wiped down to remove clues; that no semen was left at the Beethe and Levine crime scenes; and that the jury had already heard testimony that Russell had said he worked for the police. The trial court reversed its prior decision, noting that the handbook was relevant given the steps taken to minimize evidence left at the crime scene, and noting reduced prejudice due to Russell’s contention he was working for the police. The trial court allowed the State to have a witness summarize and read from the handbook, but did not allow it to go to the jury.
Russell contends admission of the handbook evidence was unfairly prejudicial. He cites Coe, 101 Wn.2d at 780, where the court reversed convictions for rape because, among other reasons, the State had cross-examined the defendant about his sexually oriented fiction. He relies additionally on State v. Hanson, 46 Wn. App. 656, 663-64, 731 P.2d 1140, review denied, 108 Wn.2d 1003 (1987), where the court held the defendant was improperly cross-examined about his own fictional writing and cited with approval the statement that "[n]o inference of any kind can be drawn about a person’s character from the kinds of books that he reads”. Hanson, at 663 (citing United States v. Giese, 597 F.2d 1170, 1207 (9th *81Cir.) (Hufstedler, J., dissenting), cert. denied, 444 U.S. 979 (1979)); see also United States v. McCrea, 583 F.2d 1083, 1086 (9th Cir. 1978) (harmless error when defendant charged with possession of unregistered firearm was cross-examined about possession of books entitled Improvised Munitions Handbook and OSS Sabotage and Demolition Manual).
Although the issue is close, this case concerns evidence of greater relevance and less prejudice than was involved in the cases cited by Russell. Unlike the evidence at issue in Coe and Hanson, the handbook related directly to the facts of the crimes themselves. The crime scenes revealed that steps had been taken to ensure certain evidence was not left behind, and relevant topics on these points were covered in the handbook. Moreover, the prejudicial effect of the evidence was moderated in this case because possession of the handbook would have been consistent with Russell’s claim to be working for the police.
For these reasons, we uphold the trial court’s ruling on ER 401 and 403 as being within the range of discretion. As to ER 404(b), the handbook evidence was admitted not to prove Russell’s character, but for the purpose of demonstrating Russell’s knowledge on a point relevant to the case.
In conclusion, we affirm the trial court’s rulings admitting the condoms, the scanner, and the handbook. We pause to emphasize, however, that we are not broadly authorizing trial courts to admit evidence of this nature in all criminal trials. Here the trial judge admitted the evidence after carefully evaluating each item’s relationship to the issues in the case and after fully considering the restrictions of ER 401, 403 and 404. We commend the trial judge for her detailed examination.
VII
The next issue is whether the trial court erred in denying a mistrial based on the State’s cross examination of Matt McCaulley. Matt McCaulley was a 16-year-old neighbor of Carol Beethe. Soon after the murder McCaulley told police *82investigators that he saw a blue Corvette with a black top driving near Beethe’s house around the time of her murder. He also said he heard a cat scream around that same time.
Prosecutors interviewed McCaulley on March 26, 1991. McCaulley described using alcohol and drugs and having related hallucinations during the summer of 1990. He admitted drinking some bourbon the night of Beethe’s murder.
At trial, McCaulley was called as a defense witness to support the theory that Beethe was killed by a boyfriend, Mike Suell. McCaulley identified a photograph of Mike Suell’s white-topped Corvette as possibly being the car he saw.
On cross examination, McCaulley denied drinking any bourbon on the night of the murder. The State reminded him of his March 26 statement in which he admitted consuming bourbon that night. McCaulley then admitted drinking some bourbon that night, but denied it was enough to "affect [his] system”. Verbatim Report on Appeal, at 6602.
The State also asked McCaulley on cross examination if he had any problems on the night in question with drug-related hallucinations. McCaulley denied having any such problems that night. The State repeatedly questioned Mc-Caulley about the discrepancies between his earlier statement that the top of the Corvette was black and his in-court testimony regarding Suell’s white-topped Corvette.
Russell first argues that a mistrial should have been granted on the basis of undue surprise. He contends the State used interview notes for impeachment that had not been disclosed in violation of the court’s discovery order.15 That order required each party to divulge in advance certain impeaching material gathered from witness interviews.16
At trial the judge ruled that the discovery order did not require the State to divulge the interview notes. Rather, the *83order had been limited to a witness’ prior inconsistent statements and did not include general impeachment evidence. Because McCaulley’s statements to the prosecutors did not involve any inconsistent statements, the court allowed the cross examination based on the interview.
We accord discretion to trial court decisions on the scope of discovery in criminal cases where not inconsistent with CrR 4.7. See State v. Blackwell, 120 Wn.2d 822, 826, 845 P.2d 1017 (1993); State v. Hoffman, 116 Wn.2d 51, 80, 804 P.2d 577 (1991). The trial judge did not abuse her discretion in ruling that the State had complied with her discovery order.
Russell argues that due process is violated when the defendant is required to provide discovery that the State is not likewise required to provide, citing Wardius v. Oregon, 412 U.S. 470, 37 L. Ed. 2d 82, 93 S. Ct. 2208 (1973). In this case, however, both parties were required to comply with the same standard. Russell also argues due process is violated when the State seeks tactical advantage through surprise, citing Coe, at 783. This argument too misses the mark, for the State acquired no tactical advantage when both parties were operating under the same rule.
Russell next argues that the State improperly inquired into McCaulley’s drug and alcohol use. He contends that the prosecution’s questions should have been limited to McCaul-ley’s use of alcohol or drugs on the night of Beethe’s murder.
It is well settled in Washington that evidence of drug use is admissible to impeach the credibility of a witness if there is a showing that the witness was using or was influenced by the drugs at the time of the occurrence which is the subject of the testimony. State v. Dault, 19 Wn. App. 709, 719, 578 P.2d 43 (1978); State v. Hall, 46 Wn. App. 689, 692, 732 P.2d 524, review denied, 108 Wn.2d 1004 (1987); State v. Smith, 103 Wash. 267, 269, 174 P. 9 (1918); see also David J. Oliveiri, Annotation, Use of Drugs as Affecting Competency or Credibility of Witness, 65 A.L.R.3d 705 § 6[b] (1975) (generally accepted in other jurisdictions that drug use may be admissible to show impairment of a witness’ faculties).
*84Here, the State’s interview of McCaulley indicated that he drank bourbon on the night of Carol Beethe’s murder and that he had drug- and alcohol-related hallucinations in the summer of 1990. Considering this evidence, the State properly questioned McCaulley about drugs or alcohol that may have influenced him on the night of Beethe’s murder and whether he had problems with drug-induced hallucinations on the same night. We see no error in this cross examination.
Russell also sought a mistrial based on a prosecution question relating to cats. After reminding McCaulley he had told police that he heard a cat scream at the time of Beethe’s murder, the prosecutor asked how McCaulley knew the sound was made by a cat. McCaulley replied he had previously heard mating cats scream. The prosecutor then asked:
Q: Do you recall you also could identify a cat scream sound from an occasion on which you killed a cat and was dangling it over a bridge?
[defense counsel]: Objection, move to strike.
the court: Sustained. That will be stricken, the jury is directed to disregard it.
Verbatim Report on Appeal, at 6608.
At no point since, despite adequate opportunity, has the prosecution presented any valid basis for asking this question. The question insinuates that McCaulley had killed a cat and dangled it over a bridge, yet the prosecution has not demonstrated that McCaulley ever said such a thing, that he ever did such a thing, or that the prosecution had good reason to think this had happened. All the State can point to is McCaulley’s statement that he heard a cat scream on the night of the murder. The question was undeniably improper, inflammatory, and unprofessional.
The State argues that the trial court’s actions sufficiently cured the error, and that any remaining prejudice did not require a mistrial. While we are unhappy with the State’s tactics we do not find that a mistrial is warranted for the reasons below.
Courts generally presume jurors follow instructions to disregard improper evidence. See State v. Swan, 114 *85Wn.2d 613, 661-62, 790 P.2d 610 (1990), cert. denied, 498 U.S. 1046 (1991). Trial courts are accorded discretion in denying a motion for mistrial; such denials will be overturned only when there is a "substantial likelihood” the prejudice affected the jury’s verdict. State v. Crane, 116 Wn.2d 315, 332-33, 804 P.2d 10, cert. denied, 111 S. Ct. 2867 (1991). Trial courts "should grant a mistrial only when the defendant has been so prejudiced that nothing short of a new trial can insure that the defendant will be tried fairly”. Mak, 105 Wn.2d at 701, quoted in State v. Hopson, 113 Wn.2d 273, 284, 778 P.2d 1014 (1989).
The question’s prejudicial effect was mitigated under the circumstances of this case. The single, isolated question occurred in the context of a lengthy trial. The trial court immediately struck the question from the record, even before McCaulley had a chance to answer it, and instructed the jury to disregard it. Under these circumstances, the trial judge did not abuse her discretion in denying Russell’s request for a mistrial.
VIII
Russell next contends that prosecutorial misconduct during cross examination and closing argument was so egregious that Russell was denied a fair trial. Before discussing the instances of alleged misconduct, we will briefly review the law applicable to prosecutorial misconduct.
Where improper argument is charged, the defense bears the burden of establishing the impropriety of the prosecuting attorney’s comments as well as their prejudicial effect. Hoffman, 116 Wn.2d at 93; State v. Hughes, 106 Wn.2d 176, 195, 721 P.2d 902 (1986). Reversal is not required if the error could have been obviated by a curative instruction which the defense did not request. Hoffman, at 93; State v. York, 50 Wn. App. 446, 458, 749 P.2d 683 (1987), review denied, 110 Wn.2d 1009 (1988).
Allegedly improper arguments should be reviewed in the context of the total argument, the issues in the case, the evidence addressed in the argument, and the instructions *86given. State v. Graham, 59 Wn. App. 418, 428, 798 P.2d 314 (1990); State v. Green, 46 Wn. App. 92, 96, 730 P.2d 1350 (1986). Remarks of the prosecutor, even if they are improper, are not grounds for reversal if they were invited or provoked by defense counsel and are in reply to his or her acts and statements, unless the remarks are not a pertinent reply or are so prejudicial that a curative instruction would be ineffective. State v. Dennison, 72 Wn.2d 842, 849, 435 P.2d 526 (1967); Graham, at 428-29.
Lastly, failure to object to an improper remark constitutes a waiver of error unless the remark is so flagrant and ill intentioned that it causes an enduring and resulting prejudice that could not have been neutralized by an admonition to the jury. Hoffman, at 93; York, at 458-59. In other words, a conviction must be reversed only if there is a substantial likelihood that the alleged prosecutorial misconduct affected the yerdict. State v. Lord, 117 Wn.2d 829, 887, 822 P.2d 177 (1991), cert. denied, 113 S. Ct. 164 (1992); State v. Wood, 44 Wn. App. 139, 145, 721 P.2d 541, review denied, 107 Wn.2d 1011 (1986).
Most of the statements Russell complains of must be analyzed under the "enduring and resulting prejudice” standard since only one was objected to at trial. Furthermore, in moving for a mistrial based on prosecutorial misconduct, Russell complained of only 3 of the 12 instances of alleged misconduct now at issue. The trial court denied the motion. The decision to deny a request for mistrial based upon alleged prosecutorial misconduct lies within the sound discretion of the trial court, and it will not be disturbed absent an abuse of discretion. State v. Ray, 116 Wn.2d 531, 549, 806 P.2d 1220 (1991); Hopson, 113 Wn.2d at 284.
Russell first complains of three statements made during closing argument that he claims were based on facts not in evidence: the statement that McCaulley said his brother was a hit man; the statement that Russell owned a Seattle Police cap with a patch on it; and the statement that additional incriminating evidence could have been developed. The defense objected to none of these statements when made.
*87The State responds accurately that the hit man and police cap statements were drawn directly from the testimony of two witnesses. The statement regarding additional incriminating evidence was made in response to Russell’s theory that the police did an inadequate job of investigating the murders and that they did not test every conceivable item of evidence. The State framed this response as follows:
But you want to listen very, very carefully to arguments made by the defense in this connection, when you know that six months ago in April they were told that if there was any work they wanted done on any item of evidence, any test, it would be done. It was said to them, folks, we really don’t want to hear about this at trial. You may have reason to guess that there is incriminating evidence that has not been developed. You really think that there is evidence of innocence there? The police are only human. They made mistakes, they did the best they could. They developed a lot of incriminating evidence. There may be some that remained undeveloped.
Verbatim Report on Appeal, at 7063.
A prosecutor may not suggest that evidence not presented at trial provides additional grounds for finding a defendant guilty. United States v. Garza, 608 F.2d 659, 663 (5th Cir. 1979). It is not misconduct, however, for a prosecutor to argue that the evidence does not support the defense theory. Graham, at 429; State v. Contreras, 57 Wn. App. 471, 476, 788 P.2d 1114, review denied, 115 Wn.2d 1014 (1990). Moreover, the prosecutor, as an advocate, is entitled to make a fair response to the arguments of defense counsel. United States v. Hiett, 581 F.2d 1199, 1204 (5th Cir. 1978).
It appears that the cited statement was aimed more at responding to defense criticisms than at finding additional reasons to convict Russell. The inadequacy of the police investigation was a constant defense theme, and the prosecutor’s statement constituted a fair response to that theory.
Moreover, the trial court ameliorated the effect of the statement when it gave a curative instruction in response to defense counsel’s subsequent statement that "there is a lot of evidence which we haven’t seen which has been held back from us. . . .” Verbatim Report on Appeal, at 7133. The court instructed the jury
*88to base your decision solely on the evidence presented in court. You are to disregard any remarks that there is additional evidence being withheld from you, the jury. You are to decide the case on the basis of the evidence presented in court and not on speculation.
Verbatim Report on Appeal, at 7148.
Despite this admonition, the deputy prosecutor could not resist commenting further on the possibility of withheld evidence. During rebuttal, she referred to defense counsel’s "notions that evidence is being held back”, and to the large amount of discovery information made available to the defense. The deputy prosecutor then added, "You know they have had access to their own experts to look at. . . this evidence, very few of whom you heard from”. Verbatim Report on Appeal, at 7269. The defense objected, and the court reminded the jury that it was to consider the evidence before it, the exhibits and the instructions. While the prosecutor improperly referred to facts not in evidence, we find that defense counsel’s objection and the court’s prompt response arguably cured any resulting prejudice. In light of this we do not find the remark so prejudicial as to warrant a new trial.
Russell next objects to the deputy prosecutor’s statement in closing argument that "[t]he killing stopped with these three women and it should go no further”. Verbatim Report on Appeal, at 7053. The defense made no objection to this statement when it was made.
Russell now claims that this remark was improper because it was based on facts not in evidence. The State asserts, however, that it is a reasonable inference from the evidence that no additional posed victims were found after Russell’s arrest. See Hoffman, at 94-95; State v. Ranicke, 3 Wn. App. 892, 897, 479 P.2d 135 (1970) (in closing argument, prosecuting attorney permitted reasonable latitude in drawing inferences from the evidence). The State also points out that defense counsel stated later that there weren’t any other posed bodies. Even if the remark was error, the prejudicial effect of this isolated statement could have been cured had the defense objected.
*89We find more serious a later comment made by the prosecutor:
Mr. Russell was going to go to California, San Diego, I think he said. If you have a reasonable doubt that he killed these women, let him go. He’ll find new friends. There is no shortage of naieve [sic], trusting, foolish young people in the cities of this country. He will settle in. He will begin looking for work. You could say he will be hunting for a job and he will find it. If you have a reasonable doubt that he’s the killer, let him go.
Verbatim Report on Appeal, at 7117-18.
Russell made no objection to these comments, but he did refer to them in his motion for a mistrial. He argues that the comments were a deliberate appeal to the jury’s fears and thus inappropriate. See State v. Belgarde, 110 Wn.2d 504, 507, 755 P.2d 174 (1988); State v. Claflin, 38 Wn. App. 847, 851, 690 P.2d 1186 (1984), review denied, 103 Wn.2d 1014 (1985). In Belgarde, the court found that the prosecutor’s inflammatory comments were a deliberate appeal to the jury’s passion and prejudice, encouraging it to render a verdict based on the defendant’s associations rather than on the evidence. Since an objection and instruction to disregard could not have erased the fear and revulsion jurors would have felt, a new trial was the mandatory remedy for such misconduct. Belgarde, at 508.
While egregious, it is doubtful that the prosecutor’s statements herein created a sense of revulsion. Moreover, defense counsel repeated the California remark in her closing argument: "The state suggested Mr. Russell killed three people in Bellevue, he was on the way to California and he would kill again.” Verbatim Report on Appeal, at 7163. This statement was made to illustrate how the State was trying to identify Russell as a serial killer. The incorporation of this statement into the defense argument weakens the contention that it denied Russell a fair trial. While we do not approve of the prosecutor’s statement, we do not find it sufficiently flagrant under the facts presented to warrant a new trial. See Darden v. Wainwright, 477 U.S. 168, 179-82, 91 L. Ed. 2d 144, 106 S. Ct. 2464 (1986) (remarks about a defendant’s future dangerousness were criticized but not regarded as reversible error).
*90Russell next contends that the State committed misconduct in commenting on his right to remain silent and in shifting the burden of proof by stating that the defense failed to bring in (1) anyone to testify he bought a ring in Canada; (2) any evidence to establish who he was working with on the morning Levine was killed; (3) some of his acquaintances to show where he went in Smith McLain’s truck; and (4) someone to establish that he worked for the police. Defense counsel made no objections to any of these statements during closing argument.
The defense now claims error, arguing the statements were not justified by the "missing witness” doctrine outlined in State v. Blair, 117 Wn.2d 479, 485-86, 816 P.2d 718 (1991). Under this doctrine, where a party fails to produce otherwise proper evidence which is within his or her control, the jury may draw an inference unfavorable to that party. Blair, at 485-86 (citing State v. Davis, 73 Wn.2d 271, 276, 438 P.2d 185 (1968)). Most jurisdictions permit the missing witness inference in criminal cases where the defense fails to call logical witnesses. Blair, at 486. The inference may be drawn only where there is an unexplained failure to call a witness whom it would be natural for a party to call if that party knew that the testimony would be favorable. Blair, at 488 (citing Davis, at 279-80). The inference may not be drawn when to do so would infringe on the defendant’s constitutional rights, including the right to remain silent. Blair, at 491. The Blair court did not agree, however, that any comment on a defendant’s failure to produce witnesses is an impermissible shifting of the burden of proof. "Here, nothing in the prosecutor’s comments said that the defendant had to present any proof on the question of his innocence. The prosecutor was entitled to argue the reasonable inference from the evidence presented.” Blair, at 491.
The first comment of which Russell complains referred to an amethyst ring allegedly belonging to Andrea Levine. The State presented evidence that shortly after Levine was murdered, Russell gave the ring to Dacia Jubinville. On cross examination, Jubinville said that Russell told her he got the *91ring in Canada. During closing argument, the prosecutor posed some questions for the defense.
If Mr. Russell bought that ring in Canada, who did he buy it from? . . . why didn’t they bring somebody down from Vancouver? They could have saved themselves a lot of trouble, just to find the man or woman that sold Mr. Russell the ring. Where is he? Doesn’t that make you wonder a little bit?
Verbatim Report on Appeal, at 7110. The comments at issue thus were based on testimony elicited by the defense, and pointed to logical support of the defense theory that was missing. These comments are justifiable under the missing witness doctrine.
Russell next complains of the following remarks:
Not one of Mr. Russell’s acquaintances called by the state or by the defense has any idea where he went in Smitty’s truck. Not one of them has any idea where he was on the mornings that Carol Beethe and Randy Levine were killed.
Verbatim Report on Appeal, at 7111. The defense contends that these comments were the same as an outright statement that Russell did not take the stand and provide an alibi. While the remarks point to absence of alibi, they do not refer to Russell’s failure to testify. These comments did not violate his right to remain silent or shift the burden of proof.
Russell also complains of the prosecutor’s reference to the absence of witnesses to corroborate that Russell worked as an undercover police informant. The prosecutor first acknowledged that Russell had no burden of proof and then said:
But when they talk about people who are peculiarly within their knowledge, for instance . . . these mysterious police that they want to claim Mr. Russell really was working for, those are people peculiarly within their knowledge. When those people are not brought forward you are entitled to ask why not? And the logical explanation is because they don’t exist.
Verbatim Report on Appeal, at 7273. These comments were made after the defense stated in closing argument that Russell could not introduce witnesses on this point because of the danger to which it would subject Russell in jail. Thus, the defense implied that it could have produced such wit*92nesses, and the State’s remarks seem well within the bounds of the missing witness doctrine.
Russell also argues that the State disparaged his right to present a defense by its cross examination of defense investigator Debra Malcom. The law allows cross examination of a witness into matters that will affect credibility by showing bias, ill will, interest, or corruption. State v. Jones, 67 Wn.2d 506, 512, 408 P.2d 247 (1965); State v. Roberts, 25 Wn. App. 830, 834, 611 P.2d 1297 (1980). The scope of such cross examination is within the discretion of the trial court. State v. Robbins, 35 Wn.2d 389, 396, 213 P.2d 310 (1950); Roberts, at 834. While the State subjected Debra Malcom to rigorous cross examination, the questions were designed to emphasize her position as part of the defense team and to show that her investigative techniques were often suggestive and incomplete. The inquiry was within the proper bounds of cross examination.
Finally, Russell contends that two personal attacks on his attorneys deprived him of his right to counsel. The first comments occurred during rebuttal, and followed statements by the defense that the State pursued Russell because of his race and that it changed and fabricated evidence to fit its theory of the case. The State responded as follows:
I think defense counsel in this case has kind of stooped to new lows in the argument you just heard, essentially going from claiming people are lying and fabricating evidence, whenever there is a piece of evidence they can’t explain away, to claiming that the police — going so far as to claim that the police planted the defendant’s Negroid hairs in the ME exhibits, that they got from his bag.
You’ve heard all the witnesses, you will judge their credibility. Thank God you, and not Ms. Schwartz, is the judge of the credibility of the witnesses ....
Verbatim Report on Appeal, at 7267-68. The State further complained that the defense had "attacked and vilified” Dr. Blake. "[N]o matter what we do, if it suits someone’s interest to claim that that evidence is not reliable, or is fabricated, they will stoop to any level to do so. You have seen that here.” Verbatim Report on Appeal, at 7289.
*93The defense made no objection to any of the comments cited above. These remarks appear to have been provoked by defense counsel and arguably constitute a fair response to attacks made by the defense on the deputy prosecutor, her witnesses, and the work of government agents. See United States v. Nanez, 694 F.2d 405, 410 (5th Cir. 1982), cert. denied, 461 U.S. 909 (1983); Graham, at 428-29. The Ninth Circuit has stated that "absent specific evidence in the record, no particular defense counsel can be maligned”. Bruno v. Rushen, 721 F.2d 1193, 1195 (9th Cir. 1983), cert. denied, 469 U.S. 920 (1984). It would appear that specific evidence supported the prosecutor’s statements here. While inflammatory, the remarks were not so prejudicial that a curative instruction would have been ineffective.
In summary, only one of the many instances of alleged prosecutorial misconduct was objected to at trial. Thus, the admonition stated by this court in Jones v. Hogan, 56 Wn.2d 23, 27, 351 P.2d 153 (1960), may be applicable: "Counsel may not remain silent, speculating upon a favorable verdict, and then, when it is adverse, use the claimed misconduct as a life preserver on a motion for new trial or on appeal.” Moreover, only the statement regarding the likelihood of Russell continuing to kill in California appears to have been capable of lasting prejudicial effect. While it is not, by itself, sufficient to warrant a new trial, we reiterate here our unhappiness with this comment.
IX
Finally, Russell maintains that he was denied a fair trial due to the cumulative effect of errors that include improper joinder, the improper admission of expert testimony and other physical "character” evidence, and the denial of the right to confront and present evidence in defense.
It is well accepted that reversal may be required due to the cumulative effects of trial court errors, even if each error examined on its own would otherwise be considered harmless. See State v. Coe, 101 Wn.2d 772, 789, 684 P.2d 668 (1984); State v. Badda, 63 Wn.2d 176, 183, 385 P.2d 859 *94(1963); State v. Alexander, 64 Wn. App. 147, 154, 822 P.2d 1250 (1992). Analysis of this issue depends on the nature of the error. Constitutional error is harmless when the conviction is supported by overwhelming evidence. State v. Whelchel, 115 Wn.2d 708, 728, 801 P.2d 948 (1990); State v. Guloy, 104 Wn.2d 412, 425, 705 P.2d 1182 (1985), cert. denied, 475 U.S. 1020 (1986). Under this test, constitutional error requires reversal unless the reviewing court is convinced beyond a reasonable doubt that any reasonable jury would have reached the same result in absence of the error. Whelchel, at 728; Guloy, at 425. Nonconstitutional error requires reversal only if, within reasonable probabilities, it materially affected the outcome of the trial. State v. Halstien, 122 Wn.2d 109, 127, 857 P.2d 270 (1993); State v. Tharp, 96 Wn.2d 591, 599, 637 P.2d 961 (1981).
We have identified no error, harmless or prejudicial, resulting from the trial court’s rulings regarding joinder, the admission of expert testimony, or the admission of evidence of other assaults. More troublesome are the deputy prosecutor’s references to a "cat killing” during Matt McCaulley’s cross examination and to Russell’s future dangerousness during closing argument. Given the scope of this trial, however, we do not find that the comments had a material effect on its outcome, nor do we believe that a different result would have been reached in their absence.
Accordingly, we affirm the judgment of the trial court and uphold the Defendant’s convictions in this case.
Brachtenbach, Dolliver, Durham, and Guy, JJ., concur.
MacKnight, editor in chief of the Santa Clara Computer and High Technology Law Jourhal, has an M.S. in microbiology from San Jose State University, a Ph.D. in microbiology from the University of California, and experience as a legal consultant in biotechnology.
Some scientists use a dot-blot process where the sample DNA is dotted onto nylon membrane strips held in individual wells, and different solutions, each *39containing a different probe, are added to each strip. The test kit used in this case employs the reverse dot-blot method, which has less chance of human error because it requires fewer steps and less manipulation. Kamrin T. MacKnight, Comment, The Polymerase Chain Reaction (PCR): The Second Generation of DNA Analysis Methods Takes the Stand, 9 Santa Clara Computer & High Tech. L. J. 287, 306 n.84 (1993).
The United States Supreme Court has since rejected the Frye test for use by the federal courts. See Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 125 L. Ed. 2d 469, 113 S. Ct. 2786 (1993). The Court has replaced the "general acceptance” inquiry of Frye with a 2-part test that examines "whether the reasoning or methodology underlying the testimony is scientifically valid and . . . whether that reasoning or methodology properly can be applied to the facts in issue”. Daubert, 113 S. Ct. at 2796. Since the trial court in this case applied the Frye test and since neither party requests this court to replace the Frye test with Daubert, we will analyze the challenge in this case under the Frye standard.
Following Dr. Geraghty’s testimony, the State moved to have him retest the vaginal swab at state expense, with a defense witness observing. The defense opposed the motion and the court denied it as untimely.
Russell does not contend that any of these problems occurred here. Furthermore, none of the experts who testified either for the defense or the prosecution raised any of the concerns that Russell now mentions in connection with the tests performed in this case.
Even with carefully controlled tests, some argue that results in forensic casework should be reconfirmed by an independent repetition from the original sample. (This is practical because of the minimal sample requirements of the PCR test and the ease of the procedure.) Office of Technology Assessment, U.S. Congress, Genetic Witness: Forensic Uses of DNA Tests 70 (1990). As stated earlier, the State sought to provide such confirmation but it was opposed by Russell and rejected by the trial court.
While no other appellate decision considering PCR evidence has evaluated admissibility using the Frye standard it is nonetheless significant that PCR evidence has been admitted under tests that evaluate the reliability and relevancy of a new scientific technique. See State v. Lyons, 124 Or. App. 598, 863 P.2d 1303 (1993); Spencer v. Commonwealth, 240 Va. 78, 393 S.E.2d 609, cert. denied, 498 U.S. 908 (1990); Clarke v. State, 813 S.W.2d 654 (Tex. Ct. App. 1991), aff'd, 839 S.W.2d 92 (Tex. Crim. App. 1992), cert. denied, 113 S. Ct. 1611 (1993); Trimboli v. State, 817 S.W.2d 785 (Tex. a. App. 1991), aff'd, 826 S.W.2d 953 (Tex. Crim. App. 1992).
A New Jersey trial court did admit PCR evidence pursuant to Frye in State v. Williams, 252 N.J. Super. 369, 383, 599 A.2d 960 (1991). Experts in Williams included the director of the Human Genetics Department at Cetus, the director of the Center for Medical Genetics at Johns Hopkins University, a research scientist with the New Jersey Department of Health, the scientific director of the Center for Blood Research Laboratories, and the director of Connecticut’s Police Forensic Science Laboratory. "These highly qualified scientists testified to the overwhelming acceptance within the scientific community of PCR-amplified DNA testing.” Williams, at 381. The test in that case was conducted by Dr. Blake using the Cetus kit. Williams, at 381-82.
We note that a slightly different analysis is necessary in cases where the "fruit” of an un-Mirandized statement is a second confession. In these "successive confession” cases, courts must also consider whether the taint from the first statement carries over to negate the voluntariness of the second statement. See Oregon v. Elstad, 470 U.S. 298, 309, 84 L. Ed. 2d 222, 105 S. Ct. 1285 (1985); State v. Wethered, 110 Wn.2d 466, 473, 755 P.2d 797 (1988). The present case does not involve this issue, for the derivative evidence was not a second statement from Russell.
A portion of the fruit from Russell’s statement was the testimony of other witnesses, not physical evidence. We note that even in the face of a coerced statement, courts are more reluctant to exclude the testimony of other witnesses than they are physical evidence. See United States v. Ceccolini, 435 U.S. 268, 274-79, 55 L. Ed. 2d 268, 98 S. Ct. 1054 (1978); see generally Wayne R. LaFave & Jerold H. Israel, Criminal Procedure §§ 9.4(e), 9.5(d) (2d ed. 1992).
A close reading reveals that Earls discussed extensively the analyses of Const, art. 1, § 9, offered by both the majority and the dissent in State v. Moore, 79 Wn.2d 51, 483 P.2d 630 (1971). Earls relied on the Moore majority’s analysis, and rejected the competing arguments raised in the Moore dissent, in declining to extend the scope of that provision beyond its federal counterpart.
At issue in Earls was the validity of an express waiver of constitutional rights, while Moore involved a challenge to the implied consent law that requires motorists suspected of driving while intoxicated to submit to a blood-alcohol-content test. See State v. Earls, 116 Wn.2d 364, 373, 805 P.2d 211 (1991); State v. Moore, 79 Wn.2d 51, 54-55, 483 P.2d 630 (1971).
In one case, this court specifically declined to resolve the issue for failure of the parties to analyze the Gunwall factors. Wethered, 110 Wn.2d at 472.
The State here proposed and the trial court agreed to give an ER 404(b) limiting instruction. Since the Defendant expressly rejected the proposed instruction, failed to prepare an alternative, and raises no issue in this regard here, we will not consider the limiting instruction requirement.
VICAP stands for Violent Incident Criminal Apprehension Program; HITS is the acronym for the Homicide Information and Tracking System. HITS is Washington’s version of the nationwide VICAP computer database program.
The defense is not arguing that the State’s failure to disclose McCaulley’s statements violated the court rule on criminal discovery, CrR 4.7. Indeed, the defense essentially conceded to the trial court that CrR 4.7 did not require disclosure.
The discovery order itself is not a part of the record on appeal. All we have is a transcribed colloquy containing the attorneys’ and the trial judge’s recollections as to the terms of that order.