dissenting.
The majority erroneously parses the appellate record and fashions a result by reducing the gatekeeper’s Robinson hearing to a nullity. My colleagues have decided to treat reliability as an evidentiary jury issue; moreover, they misapply appellate rules and case precedent. I dissent.
The Issue is Reliability and the GATEKEEPER IS ESSENTIAL
The issue on appeal is reliability: Exxon contends that plaintiffs’ experts’ opinions1 *192“are not evidence because they are unreliable and founded on unproven assumptions.” The majority reviews the reporter’s record from trial and scientific studies that are not included in this appellate record and concludes the plaintiffs presented no evidence of reliability before the jury and thus no evidence of medical causation. There are two glaring errors in the majority’s approach. First, they simply ignore the evidentiary Robinson hearing held by the trial court, disregarding it as though it is a nullity. Second, the majority confuses the jury’s role, mistaking reliability as an element to be proved before the jury.
1. Trial Court Determines Reliability, Preferably Pretrial
Because the issue is reliability, it is imperative that we review both the trial record and the pretrial Robinson hearing. Tex.R.App. P. 34.1. The trial court has the heightened responsibility to act as a gatekeeper in a Robinson hearing to screen out unreliable expert evidence. Gen. Motors Corp. v. Sanchez, 997 S.W.2d 584, 590 (Tex.1999); Maritime Overseas Corp. v. Ellis, 971 S.W.2d 402, 409 (Tex.1998); E.I. du Pont de Nemours & Co., Inc. v. Robinson, 923 S.W.2d 549, 556 (Tex.1995). This is true because reliability of an expert’s opinions is a question of admissibility. Helena Chem. Co. v. Wilkins, 47 S.W.3d 486, 499 (Tex.2001); Gammill v. Jack Williams Chevrolet, 972 S.W.2d 713, 726 (Tex.1998) (“All expert testimony should be shown to be reliable before it is admitted.”). “Preliminary questions ... concerning admissibility of evidence shall be determined by the court_” Tex.R. Evid. 104(a); see also Gammill, 972 S.W.2d at 718; Robinson, 923 S.W.2d at 556; Tex.R. Evid. 705(c) (“If the court determines that the underlying facts or data do not provide a sufficient basis for the expert’s opinion under Rule 702 or 703, the opinion is inadmissible.”).
The majority, without citation to any authority, brushes aside the evidentiary Robinson hearing in this case as “a different decision-maker in a different proceeding.” They could not be more wrong. First, the roles of the trial court and jury as decision-makers are not coterminous; the trial court is not a fact finder: “The trial court’s role is not to determine the truth or falsity of the expert’s opinion.” Robinson, 923 S.W.2d at 558; Weingarten Realty Investors v. Harris County Appraisal Dist., 93 S.W.3d 280, 285 (Tex.App.-Houston [14th Dist.] 2002, no pet.). To the contrary, “the trial court must be careful not to step into the role of the fact finder in weighing credibility....” Green v. Tex. Workers’ Comp. Ins. Facility, 993 S.W.2d 839, 843 (Tex.App.-Austin 1999, pet. denied). Instead, determination of admissibility is a question of law, not a jury decision. See N. Dallas Diagnostic Ctr. v. Dewberry, 900 S.W.2d 90, 93-94 (Tex.App.-Dallas 1995, writ denied); Weingarten, 93 S.W.3d at 285 (“Reliability is an admissibility issue for the trial court, not a weight-of-the-evidence issue for the fact finder.”).
Second, a Robinson hearing, whether held pretrial or at trial, is the same proceeding as trial. An opponent of expert testimony may object to reliability when the evidence is offered at trial. Maritime, 971 S.W.2d at 409. At trial, a challenge to an expert is typically made through a voir dire examination outside the presence of the jury:
*193Prior to the expert giving the expert’s opinion or disclosing the underlying facts or data, a party against whom the opinion is offered ... in a civil case may [ ] be permitted to conduct a voir dire examination directed to the underlying facts or data upon which the opinion is based. This examination shall be conducted out of the hearing of the jury.
Tex.R. Evtd. 705(b); see, e.g., Guadalupe-Blanco River Auth. v. Kraft, 39 S.W.3d 264, 266 (Tex.App.-Austin 2001), rev’d, 77 S.W.3d 805 (Tex.2002) (both opinions identifying voir dire challenge to reliability); N. Dallas Diagnostic Ctr., 900 S.W.2d at 93, 96. However, through case law, Texas courts have been urged to determine Robinson issues pretrial: “it is imperative to ventilate any Robinson issues as early as possible, preferably as a pretrial matter.” Maritime, 971 S.W.2d at 412 (Gonzalez, J., concurring); see Regan v. Schlumberger Tech. Corp., No. 01-00-00026-CV, 2001 WL 1344077, at *3 (Tex.App.-Houston [1st Dist.] Nov. 1, 2001, pet. denied) (not designated for publication) (“We, too, urge early Robinson/Havner hearings.... ”). A pretrial Robinson hearing is no different than one held during trial. We appropriately consider an evidentiary Robinson hearing to be the same as a voir dire held during trial. Naturally, we would consider a voir dire directed at an expert’s reliability upon appellate review of the issue. See Guadalupe-Bianco River Auth., 77 S.W.3d at 807. Accordingly, it is imperative that we consider the evidentiary hearing held pretrial. We should not punish those litigants who ventilate reliability issues early by refusing to consider the gatekeeper’s hearing. Under the majority’s approach, the trial court is converted from a gatekeeper to “idle spectator.” See Maritime, 971 S.W.2d at 412 (quoting Robinson, 923 S.W.2d at 554).
2. If Held Reliable, No Repetition Necessary
The majority also states that when Exxon renewed its objections at trial, the plaintiffs had the opportunity to prove scientific reliability and develop a full record. The impact of this holding is, frankly, a waste of judicial resources and an eviden-tiary absurdity. First, having passed muster in the gatekeeper’s hearing, a proponent of expert testimony does not need to parrot the evidence of reliability at trial. See Judge Harvey Brown, Procedural Issues Under Daubert, 36 Hous. L.Rev. 1133, 1148 (1999). In fact, the rules of evidence permit an expert to testify at trial without disclosure of underlying facts or data: “The expert may testify in terms of opinion or inference and give the expert’s reasons therefor without prior disclosure of the underlying facts or data, unless the court requires otherwise.” Tex.R. Evid. 705(a) (emphasis added).
Second, addressing reliability pretrial, rather than before the jury, allows a full presentation of the issue. As the Texas Supreme Court has noted, judges may be better-equipped to determine reliability than juries. See Robinson, 923 S.W.2d at 558. This is because a judge has the benefit of documents and briefs, whereas similarly detailed information could easily overwhelm the issues before the jury. Id. The judge “can freely ask questions in a preliminary hearing and thus glean more information” without the risks associated with explaining reliability to a jury. Id. In so doing, the trial court independently evaluates the underlying data. See Havner, 953 S.W.2d at 713. Additionally, a hearing before the trial court, as opposed to trial before the jury, solves rule-based problems with developing the record. For instance, in reviewing the trial record, the majority states it is hard-pressed to conduct a Havner review without the epidemi*194ological studies on which plaintiffs’ experts relied. However, such studies are not admissible at trial even under the learned treatise hearsay exception. See Tex.R. Evid. 803(18) (statements from learned treatises may be read into evidence, but not received as exhibits). In contrast, in determining reliability in hearing, a trial court “is not bound by the rules of evidence except those with respect to privileges.” Tex.R. Evid. 104(a). In the case before us, this explains why the trial court examined epidemiological studies, but such studies were not admitted at trial before the jury.
Third, a proponent of expert testimony has the burden to prove reliability the first time the opponent objects. See Guadalupe-Blanco River Auth., 77 S.W.3d at 807 (“Once the [opponent] made its objection, the burden was on [the proponent] to establish that [the expert’s] opinion was reliable.”); Robinson, 923 S.W.2d at 557 (at point of motion to exclude, proponents bore burden to establish admissibility); see also Frias v. ARCO, 104 S.W.3d 925, 927 (Tex.App.-Houston [14th Dist.] 2003, no pet.) (summary judgment opinion). In this case, Exxon objected pretrial and, because the court denied Exxon’s motions pretrial, the plaintiffs met their burden to the trial court’s satisfaction. No authority then requires repeated proof of reliability just because objections are renewed. Nor is there any rule requiring plaintiffs to incorporate by reference at trial their proof from the pretrial hearing.2 Here, Exxon’s repeated objections to reliability accomplished two things only. They assured preservation of error, see United Parcel Serv., Inc. v. Tasdemiroglu, 25 S.W.3d 914, 916 (Tex.App.-Houston [14th Dist.] 2000, pet. denied), and gave the trial court the opportunity to reconsider its previous rulings.3 Cf. Clark v. Trailways, Inc., 774 S.W.2d 644, 647 (Tex.1989) (renewed objections promote efficient administration of justice by providing a trial court with a final opportunity to prevent the erroneous admission of evidence).
3. Jury Determines Weight and Credibility
The majority concludes that a no-evidence challenge to reliability “tests the jury’s decision after trial, looking at the whole trial record to see if any evidence supports the verdict.” Although the majority examines the trial record for more than a scintilla of reliability,4 reliability need not be proved before the jury like an element of plaintiffs’ cause of action. A negligence cause of action has just four elements: (1) a legal duty; (2) breach of that duty; (3) proximate cause; and (4) damages. Mission Petroleum Carriers, Inc. v. Solomon, 106 S.W.3d 705, 710 (Tex.2003). Causation evidence only becomes “no evidence” if it is unreliable. See Weingarten, 93 S.W.3d at 285. Further, as explained above, reliability is a question of law for the trial court. If an opponent never objects, experts can testify at trial to causation without ever proving their opinions are based on reliable foundations. See Maritime, 971 S.W.2d at 411 (failure to timely object waives issue).
*195The trial court’s gatekeeping function, however, does not supplant the opportunity to expose weaknesses in expert testimony before the jury. Weingarten, 93 S.W.3d at 285. Under Robinson standards, the jury continues “to assess the weight and credibility of the proffered testimony.” Robinson, 923 S.W.2d at 558. This is consistent with the rules of evidence: “This rule [that the trial court determines admissibility] does not limit the right of a party to introduce before the jury evidence relevant to weight and credibility.” Tex.R. Evid. 104(e). Cross-examination remains “the traditional and appropriate means of attacking shaky but admissible evidence.” Gammill, 972 S.W.2d at 728; cf. Keo v. Vu, 76 S.W.3d 725, 734 (Tex.App.-Houston [1st Dist.] 2002, pet. denied) (“Further, factual weaknesses underlying an expert’s causation opinion generally go to the testimony’s weight, rather than its admissibility....”).
4. Exxon’s Cross-Examination
Given the generalities of direct testimony proffered by plaintiffs, the majority should hesitate to conclude that the experts’ testimony was unreliable based on cross-examination and studies absent from the appellate record. As I pointed out above, an expert may testify on direct without revealing underlying facts and data and without repeating the evidence of reliability from the pretrial hearing. See Tex.R. Evid. 705(a). Thus, plaintiffs’ experts testified on direct at trial in generalities, without offering studies, literature, or data on which they relied. For instance, Dr. Goldstein testified that since his deposition in this case, he found literature supporting a relationship between benzene and ALL.5 Dr. Stuart Lloyd Shalat testi-fled, “Most of what I’m relying on already exists in the peer-reviewed literature.” Dr. Marvin Legator testified that “science” helped him form the opinion that ALL could result from benzene exposure because “[a]s we did more and more studies, we added more and more leukemias and lymphomas in terms of what benzene can do.”
The majority concludes, via Exxon’s cross-examination, that no epidemiological study meeting Havner’s standards supports the experts’ opinions. Aside from epidemiological studies, perhaps the experts relied upon in vitro human cellular studies. Or in vivo studies. Or animal studies. The cross-examination does not foreclose these possibilities. While cross-examination of Dr. Goldstein regarding epidemiological studies was very effective, perhaps Dr. Shalat or Dr. Legator relied upon an epidemiological study meeting Havner standards. Again, this was not foreclosed by cross-examination. In my opinion, it’s extremely difficult to determine reliability solely based on cross-examination at trial if a plaintiff meets its burden of proof in a pretrial hearing. And in this case, while the cross-examination was quite damaging, without reviewing the Robinson hearing record, cross-examination goes only to weight and credibility. Cf. Maritime, 971 S.W.2d at 414 (Gonzalez, J., concurring) (stating cases in which expert testimony is untenable on its face, and thus excusing a Robinson hearing, “will be comparatively rare”); Gammill, 972 S.W.2d at 728 (availability of cross-examination does not relieve the trial court of its threshold responsibility as gatekeeper).
Further, the Robinson and Havner factors are flexible and non-exhaustive. See Havner, 953 S.W.2d at 718-19; Robinson, *196923 S.W.2d at 557; see also Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 593-94, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993) (promoting flexible inquiry rather than a definitive checklist); Coastal Tankships, U.S.A., Inc. v. Anderson, 87 S.W.3d 591, 599, 611 (Tex.App.-Houston [1st Dist.] 2002, pet. denied) (en banc) (“pertinent, suggested inquiries in assessing the reliability of expert testimony are applied flexibly and are not exclusive or required.”). Trial courts are also granted deference in their selection of the factors to use in determining reliability. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152-53, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999). Without the pretrial hearing record, we do not know if the trial court used additional factors or which Robinson and Havner factors it assessed.
Because there was an evidentiary Robinson hearing, the full bases of the experts’ opinions should have been revealed there. It is unsound and poor precedent to assume that cross-examination was conclusive without consideration of the very hearing in which reliability was tested. Using a partial record contradicts the Texas Supreme Court’s admonition “that courts should make a determination of reliability from all the evidence.” Havner, 953 S.W.2d at 720 (emphasis added).
5. What We Review and How
Throughout this dissent, I have emphasized review of the Robinson hearing, even when it is held pretrial. The majority contends I have misconstrued admissibility of expert testimony with its legal sufficiency. I am not the first to be accused of eroding the distinction. See Lucinda M. Finley, Guarding the Gate to the Courthouse: How Trial Judges are Using Their Evidentiary Screening Role to Remake Tort Causation Rules, 49 DePaul L.Rev. 335, 376 (1999). Litigants have also urged that admissibility and sufficiency reviews are distinct, although both involve the reliability of expert testimony. See Austin v. Kerr-McGee Ref. Corp., 25 S.W.3d 280, 284 (Tex.App.-Texarkana 2000, no pet.).
However, others have recognized that review of reliability, even in the context of a no-evidence challenge, requires review of the Robinson hearing. The pretrial Robinson hearing “should be recorded by the court reporter. Without a record from the hearing, the appellant will not be able to show harm.” Michol O’ConnoR, O’Con-NOR’s Texas Rules: Civil TRIALS 289, § 3.5 (2002). Further, an appeal “from a final judgment rendered in a trial in which the trial court ruled on a motion to exclude an expert ... must include the record from the hearing in the appellate record.” Id. 290, § 5.1. This is because a parly may offer evidence at the hearing and, having prevailed, not present it before the jury. See Judge Harvey Brown, Procedural Issues Under Daubert, 36 Hous. L.Rev. 1133, 1139 (1999). “[T]he evidence from the hearing will be considered in determining both admissibility and, indirectly, the sufficiency of the evidence, even if the evidence is not fully presented to the jury.” Id. Indeed, the Texas Supreme Court has stated that reliability issues go to “the admissibility of expert evidence rather than the legal sufficiency of the evidence.” Gen. Motors Corp., 997 S.W.2d at 590.
Finally, as to standard of review, the Texas Supreme Court has recognized that a no-evidence complaint may be sustained when the record shows one of the following: (1) a complete absence of a vital fact; (2) the court is barred by rules of law or evidence from giving weight to the only evidence offered to prove a vital fact; (3) the evidence offered to prove a vital fact is no more that a mere scintilla; or (4) the evidence establishes conclusively the opposite of the vital fact. Maritime, 971 S.W.2d at 409. By challenging reliability *197of the experts’ testimony, and thus admissibility, Exxon contends the testimony amounts to no evidence because a “rule of law or evidence,” specifically Rule of Evidence 702, precludes giving the evidence any weight. See id.; Exxon Pipeline Co. v. Zwahr, 35 S.W.3d 705, 718 (Tex.App.Houston [1st Dist.] 2000) (Taft, J., dissenting), rev’d, 88 S.W.3d 623 (2002). Accordingly, I would begin our analysis by examining the reliability, and therefore admissibility of the testimony, Zwahr, 35 S.W.3d at 718 (Taft, J., dissenting), recognizing the trial court’s discretion in determining the matter. See id. at 719.
I recognize that some courts conduct de novo or something similar to a de novo review in this context. See Mo. Pac. R.R. Co. v. Navarro, 90 S.W.3d 747, 750 (Tex.App.-San Antonio 2002, no pet.); Austin, 25 S.W.3d at 285. However, I would review for an abuse of discretion because the Texas Supreme Court has done so in the three post-Havner cases in which it addressed reliability in the context of a no-evidence challenge. See Exxon Pipeline Co. v. Zwahr, 88 S.W.3d at 628-29 (reversing Zwahr, 35 S.W.3d at 705); Guadalupe-Bianco River Auth., 77 S.W.3d at 807 (reversing Guadalupe-Bianco River Auth., 39 S.W.3d at 264); Helena Chem. Co., 47 5.W.3d 486, 499 (Tex.2001) (affirming Helena Chem. Co. v. Wilkins, 18 S.W.3d 744 (Tex.App.-San Antonio 2000)). In dissenting, however, I find that the degree of deference to give on appellate review is a less pressing question than whether we can ignore evidence from the gatekeeper’s hearing.
Appellate Law Spurned
1. The Record Filed
The primary reason the majority discounts the pretrial Robinson hearing is that the transcript and evidence from it are not in the appellate record.6 Under established appellate law, we cannot consider whether error exists unless the record relating to the alleged error is before us. Foust v. Estate of Walters, 21 S.W.3d 495, 504 (Tex.App.-San Antonio 2000, pet. denied). Further, as appellant, Exxon had the burden to bring forth a sufficient record showing error. See Melendez v. Exxon Corp., 998 S.W.2d 266, 274 (Tex.App.-Houston [14th Dist.] 1999, no pet.). “Absent such a record, the reviewing court must presume that the evidence before the trial judge was adequate to support the decision.” Simon v. York Crane & Rigging Co., 739 S.W.2d 793, 795 (Tex.1987); see also Foust, 21 S.W.3d at 504.
The effect of these precepts on this case ends in an affirmance: two plaintiffs keep their verdict and four plaintiffs collect nothing. For Exxon’s appeal, because we are missing the pretrial Robinson record, we must assume that evidence of reliability presented at the Robinson hearing supports the trial court’s decision to admit the experts’ testimony. Because we must assume the testimony was properly admitted, the experts’ causation evidence is legally sufficient for plaintiffs James Makof-ski, Jr. and John Russell. The lack of a Robinson record cuts both ways in this case, however, and similarly hinders the cross-appeal filed by plaintiffs Andrea Russell, Janie Devora, Phillip Devora, and Codi Stennett. They contend that one study, the ATSDR, supports causation testimony regarding benzene exposure and their health problems. However, the ATSDR is not in the trial record. By the same standards that stymie Exxon’s appeal, we must assume the Robinson hear-*198mg record supports the trial court’s ultimate decision that expert testimony in their cases was unreliable.
2. Legal Contortions
To avoid the effect of the missing record, the majority engages in legal contortions. First, the majority excuses Exxon from bringing forward a complete appellate record because neither party raised or briefed the argument. Certainly, inadequate briefing can result in waiver of an appellant’s issue. See Campbell v. State, 85 S.W.3d 176, 184 (Tex.2002) (cited by majority). However, waiver of error due to inadequate briefing is distinct from the burden to bring forth a record on appeal. Additionally, because it is an appellant’s burden to bring forth a record showing error, I do not assign blame to the appel-lees, as the majority does, for failing to brief the omissions in the appellate record. I also do not excuse the missing record, as the majority does, because “the very capable attorneys and experienced trial judge all certainly knew when a record is required and when it is not.” There is no deference on appellate review based on attorneys’ or judges’ reputations.
Second, the majority implies that 270 pages of Exxon’s motions to exclude and supporting exhibits suffice in lieu of the reporter’s record from the Robinson hearing.7 In this approach, the majority pretends that the Robinson hearing was held by submission. In fact, when the plaintiffs originally sought a hearing by submission, Exxon protested and sought an evidentiary hearing, which was held 11 days before the start of trial. The majority also fails to acknowledge that plaintiffs filed written objections to some of the exhibits attached to Exxon’s motions. We do not know whether these objections were ruled upon. And we also do not know whether the exhibits attached to Exxon’s motions encompass those offered and admitted in hearing. I do not imply that a trial court must hold an evidentiary hearing instead of ruling by submission.8 However, if an evidentiary hearing is held, one party’s filings in the clerk’s record are not a substitute on appeal for the reporter’s record of the hearing.
Third, the majority surprisingly and wrongly concludes that there was no evidence presented at the pretrial hearing.9 To this end, the majority assumes that a trial court would hold an evidentiary hearing only to receive live testimony from the experts. I disagree. I do not suggest that the pretrial hearing in this case included live testimony. Nor do I imply that all evidentiary hearings on reliability must include live testimony. Instead, evidence in a pretrial evidentiary hearing could include depositions, studies, and documents.10 In this case, the trial record and clerk’s record reflect that the trial court held an evidentiary hearing, but do not delineate what evidence was offered or admitted in that hearing.
*199Fourth, the majority wrongly concludes that under my analysis, “every time a pretrial motion ... is denied without a reporter’s record, we must presume any jury verdict thereafter is valid (no matter how scientifically unsound).” We will only need a reporter’s record of evidentiary Robinson hearings, not hearings by submission. Further, I leave open the possibility that in rare cases, expert testimony will be untenable on its face. See Maritime, 971 S.W.2d at 414 (Gonzalez, J., concurring).
Fifth, the majority presumes that the parties “never requested that the hearing be recorded.” The record is silent as to whether the parties requested or excused the court reporter from the pretrial hearing. However, failure to request that a hearing be recorded does not excuse an appellant’s burden to bring forth a record showing error. Further, Exxon specifically designated the pretrial Robinson record as part of the appellate record:
The pre-trial “Robinson ” hearing held on September 30,1999.... This request includes all exhibits offered or admitted into evidence either for the Court or for the jury at the hearing of September 30, 1999 or at trial on the merits beginning October 11, 1999 and ending November 19,1999.
Thus, we cannot presume the record we have is the entire record pertinent to reliability. See Andrews v. Sullivan, 76 S.W.3d 702, 704 (Tex.App.-Corpus Christi 2002, no pet.) (sufficiency challenge waived where party did not request partial record and failed to file exhibits in appellate record); see also Sandoval v. Comm’n for Lawyer Discipline, 25 S.W.3d 720, 722 (Tex.App.-Houston [14th Dist.] pet. denied).
Sixth, in discounting the missing Robinson record, the majority states it was plaintiffs’ “burden to present scientifically reliable evidence of causation at trial” In citing Havner and Robinson for this proposition, the majority ignores Maritime, decided after those cases. In Maritime, the court encouraged pretrial resolution of reliability issues. 971 S.W.2d at 409. Thus, the majority’s citations to Havner and Robinson are taken out of context. Neither case supports the contention that we review only the evidence presented at trial before the jury.
Seventh, the majority circularly concludes that the trial cross-examination both proves unreliability and rebuts the presumption we give to the missing record. However, the majority fails to cite a single case in which the presumption given the missing record is rebutted by that which is available.11 I have already discussed the limitations of cross-examination in this case given the generalities of direct testimony. The cross-examination simply did not foreclose the bases of the experts’ opinions. If the bases of their opinions are *200not foreclosed, we presume that the missing record provides the evidence of reliable bases. I think a stronger case for rebuttal would be presented, for example, where the existing record proves that no evidence was offered, admitted, or submitted in the missing pretrial hearing. Or where the existing record clearly delineates what the trial court reviewed in pretrial hearing. In truth, it is the majority who pretends in this case, allowing its reactions to the trial testimony to blur the requisites for appellate review.
Finally, if all the above approaches fail the majority’s reasoning, the majority creates a safety net by simply making its own appellate record. Before oral argument, we became aware of some of the deficiencies in the record. During argument, the parties acknowledged that the trial court examined scientific studies not included in the appellate record. At Chief Justice Blister’s request, the parties filed fourteen studies as exhibits to their briefing. His request was an implicit acknowledgment that we must review the evidence which the trial court reviewed. Exxon properly objected, asserting that such studies were outside the trial record. We may not create a new record on appeal. Richards v. Comm’n for Lawyer Discipline, 35 S.W.3d 243, 251 (Tex.App.-Houston [14th Dist.] 2000, no pet.); Intermarque Auto. Prod., Inc. v. Feldman, 21 S.W.3d 544, 547 n. 3 (Tex.App.-Texarkana 2000, no pet.). Materials attached to briefing, but not in the appellate record, cannot be considered by an appellate court. Till v. Thomas, 10 S.W.3d 730, 733-34 (Tex.App.-Houston [1st Dist.] 1999, no pet.); Tex.R.App. P. 34.1.
Judicial PeRspective
Judges are human and are affected by the facts of the cases before them. In this case, the trial record imparts a gut reaction that plaintiffs’ experts are, at best, on the minority (though perhaps leading) edge of science and, at worst, absolute quacks. As judges, we are also aware of cases discussing the limitations of science, benzene and toxic exposure, and cancers. See Frias, 104 S.W.3d at 925; Daniels v. Lyondell-Citgo Ref. Co., Ltd., 99 S.W.3d 722, 726 (Tex.App.-Houston [1st Dist.] 2003, no pet.); Navarro, 90 S.W.3d at 747; Austin, 25 S.W.3d at 280. However, justice should not give into visceral reactions to a case.
We must follow procedural rules. We must enforce requisites for appellate review. If procedural and appellate law preclude a review on the merits, we must write accordingly. And in this case, the incomplete record precludes a true review on the merits. Whether an appellate court can disregard the gatekeeper’s hearing in a no-evidence context is the material issue in this case, upon which a take nothing judgment or an affirmance rests. The answer to the issue will also clarify for all courts and litigants the correct procedure for challenging an expert and proving reliability. I thus urge review of this case by the Texas Supreme Court. Although we may have decided the case differently were we jurors, appellate court judges must apply rules and law consistently and evenhandedly. Otherwise, we diminish rule of law, predictability, and regularity necessary to guide future litigants.
Under the standards I have discussed in this dissent, I would overrule Exxon’s issues, overrule plaintiffs’ cross-issues, and affirm the trial court’s judgment. Because the majority has not followed established rules and standards for appellate review, I dissent.
. Exxon challenges the contamination causation testimony presented by Professor Patrick Agostino (a geologist), Dr. Joseph Hughes (a professor of environmental science and engineering), Dr. Phillip Gschwend (a chemical oceanographer), Dr. Dennis McLaughlin (a hydrologist), and Dr. Jack Matson (an envi*192ronmental engineer). Further, Exxon challenges the medical causation testimony provided by Dr. Marvin Legator (a toxicologist), Dr. Bernard Goldstein (a toxicologist and hematologist), Dr. Stuart Lloyd Shalat (an epidemiologist), and Dr. Michael Gray (a medical doctor).
.Nor did Exxon need to incoiporate by reference evidence from the pretrial hearing. However, by suggesting that either party could incorporate such proof by reference at trial, the majority again implicitly acknowledges our need to review the evidence which the trial court reviewed as gatekeeper.
. And the trial court did reconsider its rulings as to four plaintiffs, finding no reliability and thus no evidence of causation.
. For instance, the majority states "it cannot grant even a scintilla of weight” to general references to studies that are not in the trial record.
. The majority emphasizes cross-examination of Dr. Goldstein in which he admits he provided no animal or epidemiological studies to Exxon since his deposition. However, this goes to supplementation of discovery, not the existence or non-existence of studies.
. Nor can it appear in the record. On file with this court at our request is a letter from the official court reporter in which she advises that there was no reporter’s record made by her on September 30, 1999, the day of the Robinson hearing.
. If the plaintiffs Bled responses to Exxon’s motions to exclude, Exxon did not bring them forward on appeal. The clerk’s record is entirely one-sided.
. Cf. Tanner v. Westbrook, 174 F.3d 542, 546 (5th Cir.1999) (where trial court refused to hold pretrial hearing, appellate court reviewed documents that were submitted pretrial).
. The majority reaches this conclusion because Exxon’s counsel introduced himself to the experts at trial and cross-examined them without referring to the pretrial hearing.
. For instance, with regard to contamination causation, the plaintiffs contend they used "Exxon’s numbers” to calculate the amount of benzene lost in the blowout. Exxon’s documents evidencing such numbers could be presented without live testimony.
. The majority miscites Candelier v. Ringstaff, 786 S.W.2d 41, 44 (Tex.App.-Beaumont 1990, writ denied). Unlike this case, in Candelier, the appellant requested a partial reporter’s record from two post-trial hearings only. Under the rules of appellate procedure, the "appellate court must presume that the partial reporter’s record designated by the parties constitutes the entire record for purposes of reviewing the stated points or issues.” Tex.R.App. P. 34.6(c)(4). In one of the post-trial hearings, the Candelier trial court clearly stated that it was considering the evidence from trial in determining the post-trial issue. Thus, the Beaumont appellate court realized it did not have the entire record pertinent to the issue.
Here, we know the record pertinent to reliability is incomplete because (1) during trial, the trial court and attorneys referred to the pretrial objections, rulings, and evidence; (2) the clerk's record reflects that Exxon demanded and received an evidentiary Robinson hearing; and (3) in oral argument, the attorneys acknowledged the trial court examined studies not a part of the trial record.