Drumm v. Freedom of Information Commission

*********************************************** The “officially released” date that appears near the be- ginning of each opinion is the date the opinion will be pub- lished in the Connecticut Law Journal or the date it was released as a slip opinion. The operative date for the be- ginning of all time periods for filing postopinion motions and petitions for certification is the “officially released” date appearing in the opinion. All opinions are subject to modification and technical correction prior to official publication in the Connecticut Reports and Connecticut Appellate Reports. In the event of discrepancies between the advance release version of an opinion and the latest version appearing in the Connecticut Law Journal and subsequently in the Connecticut Reports or Connecticut Appellate Reports, the latest version is to be considered authoritative. The syllabus and procedural history accompanying the opinion as it appears in the Connecticut Law Journal and bound volumes of official reports are copyrighted by the Secretary of the State, State of Connecticut, and may not be reproduced and distributed without the express written permission of the Commission on Official Legal Publica- tions, Judicial Branch, State of Connecticut. *********************************************** JOHN DRUMM, CHIEF OF POLICE, ET AL. v. FREEDOM OF INFORMATION COMMISSION (SC 20656) Robinson, C. J., and McDonald, D’Auria, Mullins, Ecker, Dannehy and Elgo, Js.* Syllabus Pursuant to a provision of the Freedom of Information Act (§ 1-210 (b) (3) (D)), records of law enforcement agencies not otherwise available to the public that were compiled in connection with the detection or investi- gation of a crime are exempt from disclosure if the disclosure of such records would result in the disclosure of ‘‘information to be used in a prospective law enforcement action if prejudicial to such action . . . .’’ The plaintiffs, the town of Madison, its police department, and its chief of police, D, appealed to the trial court from the decision of the defendant, the Freedom of Information Commission, ordering the disclosure, pursu- ant to the Freedom of Information Act (FOIA), of certain documents contained in the police department’s homicide investigation files to the intervening defendant, N, a filmmaker who was the complainant before the commission. The requested documents related to the 2010 homicide of B. Immediately after the homicide, DNA evidence was recovered, but the case went unsolved. In 2013, N and B’s son, H, began working on a documentary about B’s unsolved homicide. Between 2013 and 2019, H met with members of the police department, including S, three times in an attempt to obtain information about the investigation. During those meetings, H was left with the impression that the police department had classified B’s homicide as a cold case. S indicated to H that, although the DNA evidence was sufficient to identify potential suspects, it was inadequate to prosecute any particular individual, and that the police department had had the same prime suspect since one week after the homicide. Nevertheless, D, citing the ongoing investigation, would not permit the police department to cooperate with the documentary project. As a result, N filed an FOIA request with the police department, which denied the request and declined to produce any of the requested docu- ments on the ground that they were not subject to public disclosure pursuant to § 1-210 (b) (3) (D). N then filed a complaint with the commis- sion. The only evidence the plaintiffs offered at the hearing before the commission was S’s testimony. S acknowledged that the case had gone cold by 2016 and that, with the technology available at that time, the DNA evidence could not lead to an arrest. S also testified, however, that the investigation remained open and active, that he was working on the case periodically throughout the year, and that the police periodically received new leads, which he then would investigate. S further testified that he monitored changes in forensic technology and suggested that new DNA technologies might help the police make an identification in the future. The commission ruled in favor of N with respect to most of the requested documents, ordering that the plaintiffs provide N with copies of those documents. The commission found that, although B’s death continued to be investigated, the requested documents did not fall within the exception from disclosure of law enforcement records because the plaintiffs had failed to establish either prong of § 1-210 (b) (3) (D), namely, that the requested records would ‘‘be used in a prospec- tive law enforcement action’’ and that their release would be prejudicial. The trial court upheld the commission’s decision and dismissed the plaintiffs’ appeal. In so doing, the trial court concluded that the law enforcement exception to the FOIA is governed by a reasonable possibil- ity standard, articulated a list of seven, nonexclusive factors that the commission should use to determine whether a prospective law enforce- ment action is a reasonable possibility, and determined that the plaintiffs were unable to satisfy that standard. On the plaintiffs’ appeal from the trial court’s judgment, held: 1. The plaintiffs could not prevail on their claim that the commission arbi- trarily and capriciously had failed to articulate and apply the correct legal standard that governs claims of exemption under § 1-210 (b) (3) (D): a. The trial court correctly concluded that, under the first prong of § 1-210 (b) (3) (D), a ‘‘prospective law enforcement action’’ is a law enforcement action that is at least a reasonable possibility: Insofar as the FOIA does not define the term ‘‘prospective,’’ this court consulted dictionary definitions of the term and concluded that the statute was facially ambiguous, as plausible arguments could be made that the legislature, in enacting § 1-210 (b) (3) (D), may have intended ‘‘prospective’’ to have a probabilistic meaning, such as potential, antici- pated, expected, likely or possible, or to mean in prospect or pertaining to a prospect, as in prospective buyers. In resolving that ambiguity, this court adopted the probabilistic definition of ‘‘prospective’’ and agreed with the trial court’s conclusion that a ‘‘prospective law enforcement action’’ refers to a future law enforcement action that has at least a reasonable possibility of occurring, meaning that the occurrence is more than theoretically possible but not necessar- ily likely or probable. Moreover, although the law enforcement exception plausibly could be read to impose either a more exacting standard, such as by requiring that the police show that an arrest or prosecution is pending or likely, or a less demanding standard, such as by requiring that the police demon- strate only a speculative or theoretical possibility of some future law enforcement action, the reasonable possibility standard was the most reasonable reading of the law enforcement exception insofar as it struck a proper balance between the competing public interests underlying the FOIA, namely, fostering openness and transparency while protecting important governmental functions that demand a degree of confiden- tiality. Furthermore, application of the rules of statutory interpretation to a related FOIA provision (§ 1-215 (b) (3)), which exempts from disclosure information relating to records of arrest that may ‘‘prejudice a pending prosecution or a prospective law enforcement action,’’ and the legislative history of § 1-210 (b) (3) (D) provided further support for the reasonable possibility standard and ruled out the most extreme readings of the term ‘‘prospective.’’ b. This court clarified that, under the first prong of § 1-210 (b) (3) (D), a respondent before the commission must establish only that it is at least reasonably possible that the information contained in a requested document will ‘‘be used in’’ support of an arrest or prosecution: In Dept. of Public Safety v. Freedom of Information Commission (51 Conn. App. 100), the Appellate Court stated in dictum that the law enforcement exception is satisfied when there is an evidentiary showing that the requested records are going to be used in a law enforcement action, and Superior Court judges and the commission have relied on that language to require a showing that it is at least probable, if not certain, that the subject records would be used for an arrest or prose- cution. Construing the statute to create a coherent and cohesive scheme, this court presumed that the same standard must govern both the ‘‘to be used’’ and the ‘‘prospective law enforcement action’’ elements of the first prong of the statutory law enforcement exception, and, accordingly, it overruled the Appellate Court’s dictum in Dept. of Public Safety to the extent that it imposed a standard in connection with the ‘‘to be used’’ element that was different from the reasonable possibility standard that applied to the ‘‘prospective law enforcement action’’ element. c. This court rejected the policy arguments advanced by the plaintiffs and the amicus curiae, the Division of Criminal Justice: The division’s claim that the reasonable possibility standard should be deemed satisfied when a law enforcement investigation is open, a suspect has been identified, and no insurmountable obstacles exist to a future arrest or prosecution was not supported by the language or legislative history of the statute, and that proposed rule followed a categorical approach that failed to account for the specific facts and context of each individual case, placed too much weight on the law enforcement side of the scale, precluded the public from exercising any effective oversight in all cases in which any applicable statute of limitations has not run, and did not account for the fact that, with the passage of time, it becomes increasingly likely that openness, rather than secrecy, is what will unearth the elusive lead that will help the police solve the case. Moreover, the claim that, as a matter of public policy, courts and the commission should defer to the expertise of law enforcement agencies or officials when construing or applying the law enforcement exception was not supported by the statutory scheme or the legislative history, especially in view of the fact that the legislature, which has been clear in the context of exceptions to the disclosure of records when it intends to give the agency possessing the subject records increased deference, did not do so when it drafted § 1-210 (b) (3). d. This court articulated various factors for determining, in the context of the crime of murder and other crimes that involve lengthy or no statutes of limitations, whether a future law enforcement action is reason- ably possible: This court agreed that the following seven, nonexclusive factors identi- fied by the trial court were relevant to a determination of whether a future law enforcement action is reasonably possible: the length of time that has passed since the crime; the length of time that has elapsed since the law enforcement agency last obtained significant new evidence or leads; whether the investigation, even if officially open, is classified as a cold case; the number of investigators presently assigned to the investigation; the amount of time investigators are committing to the case; whether the agency has a suspect and, if so, whether the agency’s suspicion is supported by more than speculation; and whether advances in science or technology, such as improvements in DNA analysis, may help solve the crime. This court also noted that those individual factors may vary in importance from case to case and are not intended to serve as a complete or mecha- nistic checklist, and sight should not be lost of the two fundamental issues that underlie the factors, namely, whether the law enforcement agency continues to investigate the crime actively and earnestly, and, if the investigation has gone cold, whether there remains a reasonable possibility that the investigation ultimately will culminate in some law enforcement action. 2. The existing administrative record was not sufficient to permit this court to apply the newly adopted reasonable possibility standard as a matter of law, and, accordingly, this court remanded the case for further pro- ceedings before the commission: The commission’s finding that the police had not identified a suspect was without support in the record and was contradicted by S’s and H’s testimony that the police department had enough DNA to develop suspects and that it had a prime suspect, the erroneous finding was apparently important to the commission’s law enforcement exception analysis insofar as the commission had relied on that finding to distin- guish the present case from previous cases in which it had found that release of requested records would be prejudicial to a prospective law enforcement action, and, accordingly, this court concluded that the com- mission must be afforded an opportunity on remand to consider what weight and importance, if any, to give to the testimony that the police department identified a prime suspect early in the investigation. Moreover, the commission’s final decision could be understood to require a probability, even a certainty, that the requested records will be used for an arrest or prosecution, a stringent standard that is not consistent with the plain language of the statute, and, because the commission’s factual findings, such as that the plaintiffs’ claims were wholly specula- tive, were conclusory and may be tied up with the legal standard that it applied, this court could not apply the new reasonable possibility standard to the existing record without running the risk of substituting its judgment for that of the commission as to the weight of the evidence on questions of fact, and, accordingly, the commission must be provided with the opportunity on remand to assess whether some law enforcement action still remains a reasonable possibility. Furthermore, in concluding that a law enforcement action was not rea- sonably possible, the trial court relied on certain statistical evidence regarding the probability of a prosecution in cold murder investigations and the extent to which public disclosure of information in such cases would improve the likelihood of an arrest and prosecution rather than prejudice that outcome, but the commission was the fact finder, and it should have the opportunity on remand to review any available statistical data in the first instance, with the input from the parties and any expert testimony they wish to offer, before making relevant findings. 3. Although the trial court did not address the commission’s determination that the plaintiffs had failed to establish that disclosure of the requested records would be prejudicial to a prospective law enforcement action, this court declined to give the trial court the opportunity on remand to consider the prejudice prong of § 1-210 (b) (3) (D) because, under the circumstances of the present case, further review by the trial court was unnecessary: The two prongs of the law enforcement exception, although distinct, are not wholly unrelated, several of the factors that the trial court identified as relevant to the prospective law enforcement action prong of § 1-210 (b) (3) (D), such as the existence of a suspect and the availability of DNA evidence for future testing, also are relevant to the issue of prejudice, and the commission’s reconsideration of these matters will necessarily bear on prejudice, as well. Moreover, it seemed likely that the commission also applied the wrong legal standard to the prejudice prong insofar as it improperly conflated the two prongs of the statutory exception and improperly relied on S’s offhand statement that he could ‘‘go on with speculating’’ as to how the requested information might be used, and, on remand, any documents that the plaintiffs submit for in camera review as containing potentially prejudicial information should be reviewed by the commission under its established standards. Argued October 19, 2022—officially released February 27, 2024 Procedural History Appeal from the decision of the defendant ordering the disclosure of certain records, brought to the Supe- rior Court in the judicial district of New Britain, where the court, Klau, J., granted the motion to intervene as a defendant filed by Anike Niemeyer; thereafter, the case was tried to the court, Klau, J.; judgment for the defendant, from which the plaintiffs appealed. Reversed; further proceedings. Floyd J. Dugas, for the appellants (plaintiffs). Valicia Dee Harmon, commission counsel, with whom, on the brief, was Colleen M. Murphy, general counsel, for the appellee (defendant). Stephanie Rice, law student intern, with whom were David A. Schulz and, on the brief, Emile Shehada, law student intern, for the appellee (intervenor Anike Niemeyer). Matthew A. Weiner, former assistant state’s attorney, and Sarah Hanna, former senior assistant state’s attor- ney, filed a brief for the Division of Criminal Justice as amicus curiae. Dan Barrett and Elana Spungen Bildner filed a brief for the American Civil Liberties Union Foundation of Connecticut as amicus curiae. Opinion MULLINS, J. This appeal arises from the 2010 homi- cide of Barbara Beach Hamburg and the ensuing efforts of her son, Madison Hamburg, and the intervener, Anike Niemeyer, a filmmaker who was the complainant before the defendant, the Freedom of Information Commission (commission), to produce a documentary film exploring the victim’s death and the failure of the plaintiffs and respondents before the commission—the town of Madi- son, its police department, and its chief of police, John Drumm (respondents)—to solve the crime.1 In 2019, Niemeyer filed a request under the Freedom of Informa- tion Act (FOIA), General Statutes § 1-200 et seq., with the Madison Police Department to review the victim’s case file. The police department denied the request, which led Niemeyer to file a complaint with the com- mission. After a contested hearing, the commission ruled in favor of Niemeyer and ordered the respondents to pro- vide copies of the police department’s investigation files. Specifically, the commission rejected the respon- dents’ contention that the request fell within the excep- tion from disclosure of law enforcement records pursuant to General Statutes § 1-210 (b) (3), which exempts from disclosure ‘‘records . . . compiled in connection with the detection or investigation of crime, if the disclosure of such records would not be in the public interest because it would result in the disclosure of . . . (D) information to be used in a prospective law enforcement action if prejudicial to such action . . . .’’ (Emphasis added.) The trial court affirmed the decision and disclosure order of the commission and dismissed the respon- dents’ appeal. In doing so, the trial court clarified that the statutory phrase ‘‘ ‘prospective law enforcement action’ . . . refers to a future law enforcement action, i.e., an arrest [or] prosecution, the occurrence of which is at least a reasonable, not a mere theoretical, possibil- ity.’’ Then, after applying this standard, the trial court rejected the respondents’ contentions that the decision of the commission (1) was arbitrary and capricious, insofar as the legal standard applied by the commission was novel, incorrect, or insufficiently articulated, and (2) rested on factual findings that were clearly errone- ous and not supported by substantial evidence. We agree with the trial court that, in order to satisfy their burden of showing that the law enforcement exception applies, the respondents were required to demonstrate that a future law enforcement action was a reasonable possibility. We also conclude that the case must be remanded to the commission for further proceedings to allow the commission to apply that standard. Accord- ingly, we reverse the judgment of the trial court. I The record reveals the following relevant facts and procedural history. On March 3, 2010, the victim’s sister, Conway Beach, and her daughter, Barbara Alexandra Hamburg, discovered the victim’s dead body outside the victim’s home on Middle Beach Road West in Madison. Barbara Alexandra Hamburg called 911 to report the death. The medical examiner determined that the victim died from multiple blunt and sharp force injuries, lead- ing the examiner to deem her death a homicide. DNA matching the male Hamburg lineage was found under the victim’s fingernails. After the crime, the police department issued press reports stating that it was look- ing for a Hamburg relative for questioning. That relative submitted DNA for testing, but the police did not con- sider the DNA match strong enough to justify an arrest warrant. The victim’s homicide has yet to be solved, and the case remains open. The department has not held any press conferences, published any press releases, publicly named or cleared any suspects, or otherwise updated the public on the case since March, 2010. In 2013, Hamburg and Niemeyer began working on a documentary film about the victim. In connection with that documentary, Hamburg met with representa- tives of the police department in 2013, 2016, and 2019 in an ongoing attempt to obtain information about the homicide investigation. At the initial, February, 2013 meeting, two Madison police detectives, Detective Ser- geant Neal Mulhern and Detective Christopher Sudock, indicated that they were still pursuing leads and actively working on the case. At that time, they indicated that, if the case ever went cold, they might be open to doing an interview or releasing a statement regarding the investigation in cooperation with the documentary. In October, 2016, Hamburg again met with Mulhern and Sudock, as well as with Drumm and two cold case detectives. At that time, Hamburg was under the impres- sion that the police department had run out of resources and leads and, therefore, had classified the victim’s homicide as a ‘‘cold case.’’ The officers indicated that the cold case unit of the Office of the Chief State’s Attorney was reviewing the case.2 The detectives again inquired about Hamburg’s male relative. In June, 2019, Hamburg requested and was granted a third meeting, this time with only Sudock. Sudock indicated that the police department had no unidenti- fied DNA related to the case, that the DNA test kits that had been used were faulty or expired, and that the department lacked the resources to run additional tests. He also indicated that, although the existing DNA evi- dence was sufficient to identify potential suspects, it was inadequate to initiate a prosecution against any particular individual. Sudock further indicated that the cold case review had not led to any new interviews and that the department had had the same ‘‘number one’’ suspect since the week after the homicide. That sus- pect’s telephone—presumably a cell phone—had been turned off during a twenty-four hour period around the time of the crime. Sudock provided the following summary of the meeting: ‘‘[A]t that time, we were stale; we hadn’t gotten any new leads, we had looked at all the evidence, technology hadn’t changed. We haven’t seen anything nation[wide] or worldwide that would assist us in possibly identifying the perpetrator.’’ Accord- ingly, Sudock expressed a willingness to cooperate with the documentary project, pending Drumm’s approval. Drumm did not give his approval, citing the ongoing investigation as the reason. Consequently, the police department declined to provide further information or assistance. As a result, in October, 2019, Niemeyer filed an FOIA request with the department. Niemeyer’s request sought copies of five categories of documents: all investi- gatory records, witness statements, interrogation records, and crime scene recordings related to the homicide, as well as a transcript or recording of the 911 call. The request encompasses hundreds, if not thousands, of individual documents, including photographs. The police department denied the request and initially declined to produce any documents. The department contended that the requested documents were not sub- ject to public disclosure pursuant to § 1-210 (b) (3).3 In response, Niemeyer filed a complaint with the com- mission. The commission heard the matter as a contested case in February, 2020. By that time, nearly one decade had passed since the homicide. The only evidence the respon- dents offered at the hearing was Sudock’s testimony. With respect to the first prong of § 1-210 (b) (3) (D), whether the information requested was ‘‘to be used in a prospective law enforcement action,’’ Sudock testified that, when he joined the police department in 2012, approximately two years after the victim’s death, the department was still focusing on the crime, and it was a high priority for Drumm at that time. By approximately 2016, however, the case had gone cold. Sudock further acknowledged that several ‘‘cold case looks’’ at the file since 2016 had failed to turn up any potential investigative steps or additional individuals to interview. He conceded having told Hamburg that ‘‘[n]obody has made a phone call . . . or written an anonymous letter, [or] anything; we’ve got nothing. We’ve done hundreds and hundreds and hundreds of interviews. . . . [The state forensic science labora- tory] . . . [has] identified suspects, but [the] DNA is not enough to go forward with the arrest warrants, or anything like that.’’ He also conceded that, at the June, 2019 meeting, he might have told Hamburg that ‘‘the case is not moving forward and that it’s held up . . . .’’ Sudock also confirmed having previously told Hamburg something to the effect of, ‘‘I need something to investi- gate, and . . . right now, I’m stuck . . . .’’ Sudock also testified, however, that the investigation remained open and active. He acknowledged that the case was ‘‘stale’’ and that long periods of time—as much as eight weeks—pass without any work being done on it. But he stated that he does work on the case periodi- cally throughout the year when ‘‘something pops up, and we’re back to [having] something to investigate again.’’ He was unable to put a firm number on how frequently he works on the case or how much time he dedicates to it but ultimately suggested that he typically works on it at least once per month. As recently as the week prior to the hearing, Sudock testified, there had been new activity in the case. He specifically stated that, ‘‘[l]ast week, we had informa- tion on this investigation,’’ but he declined to answer any questions about that development. Looking to the future, Sudock testified that his ongo- ing work on the homicide investigation includes analyz- ing evidence, monitoring changes in forensic technology, and pursuing possible leads. He suggested that unspeci- fied new DNA technologies might help the police make an identification. He expressed no opinion, however, as to the likelihood of, or potential timetable for, that development in this particular case. With respect to the second, prejudice prong of § 1-210 (b) (3) (D), Sudock stated that many of the requested documents, including photographs and video recordings, contained ‘‘information . . . that only the perpetrator of this crime would know . . . .’’ He opined that releas- ing such information to the public—especially for expected use in a documentary film—would make it more diffi- cult for the police to assess the credibility of any infor- mants who might come forward and might help suspects to create fake alibis. When asked to elaborate about the possible prejudice, Sudock explained: ‘‘There’s a myriad of things that could happen if all the information in this case was public and public knowledge, and it would certainly prevent a successful prosecution. It would make it more difficult. I shouldn’t say prevent; it would make it extremely more difficult to make an arrest and also to prosecute the case.’’ He further con- curred with the statement that ‘‘[i]t is of paramount importance that any new information not be colored by information that is made publicly available. That is to say that the credibility of any new information will be amplified if it is information that would only be known by the killer, as opposed to information that could be gathered from watching a film.’’ (Internal quo- tation marks omitted.) Finally, when asked to provide additional information as to the types of law enforcement actions that might be prejudiced by release of the victim’s homicide files, and the nature of the potential prejudice, Sudock responded: ‘‘I can’t say that there’s something in a report that’s viewed [as] insignificant now that may be important tomorrow, [one] year from now, five years from now. . . . I don’t know what it could be. . . . [T]here could be somebody out there [who] has infor- mation [who is] holding on to that information because of a fear . . . and now a documentary comes out, and now they get more fear . . . . I just don’t know. There’s so much information there. It could be a search warrant; it could be an arrest warrant. [It] could be another scene. I mean, there’s so many things that it could be. I can go on . . . speculating, but I don’t know at this point in time. . . . The information . . . in my opinion should . . . remain with the police department and not be public because it’s only going to hamper any further investigative tool or technique that we would have in the future.’’ Following the hearing, the respondents reconsidered their position, in part, and released the 911 telephone call recording to Niemeyer. In its decision, the commis- sion ruled in favor of Niemeyer with respect to most of the remaining documents, ordering that copies be provided.4 Although the commission found that the vic- tim’s death continues to be investigated as the police receive new information, it concluded that the respon- dents had failed to establish either that the records ‘‘would be used in a prospective law enforcement action’’ or that their release would be prejudicial. In reaching this conclusion and distinguishing the case from prior ones in which the commission had determined that the law enforcement exception was satisfied, the commission relied on, among other things, its observations that (1) the respondents had not called any additional witnesses to corroborate Sudock’s testimony, (2) Sudock’s testi- mony as to both prongs of § 1-210 (b) (3) (D) was purely speculative, (3) the police department has not identified any suspects in the crime, and (4) there was no evidence that an arrest was expected. The respondents appealed from the commission’s decision to the trial court pursuant to the Uniform Administrative Procedure Act (UAPA). See General Statutes § 4-183. Although the trial court acknowledged that the commission’s decision was not a model of clarity, the court found no reversible error. Looking to federal law for guidance, the trial court concluded that the law enforcement exception to the FOIA is governed by a reasonable possibility standard, and the court artic- ulated a list of seven nonexclusive factors that the com- mission should use to determine whether a prospective law enforcement action is a reasonable possibility. Hav- ing clarified the standard that governs the prospective law enforcement action exception (law enforcement exception) and determined that the respondents were unable to satisfy that standard, the court affirmed the commission’s decision and dismissed the appeal. The respondents appealed from the trial court’s judgment to the Appellate Court, and we transferred the appeal to this court pursuant to General Statutes § 51-199 (c) and Practice Book § 65-1. Additional facts will be set forth as necessary. II On appeal, the respondents’ primary claim is that the commission arbitrarily and capriciously failed to articulate and apply the correct legal standard that gov- erns claims of exemption under § 1-210 (b) (3) (D). They further contend that the trial court adopted and retroactively applied a new version of the standard with- out affording them a fair chance to satisfy that standard. We address the legal standard in this part of the opinion, and, in part III, we consider whether the trial court’s application of the new standard was correct as a matter of law rather than a remand of the case to the com- mission. A The following well established principles guide our interpretation of the law enforcement exception to the FOIA. ‘‘Under the UAPA, it is [not] the function . . . of this court to retry the case or to substitute its judg- ment for that of the administrative agency. . . . Even for conclusions of law, [t]he court’s ultimate duty is only to decide whether, in light of the evidence, the [agency] has acted unreasonably, arbitrarily, illegally, or in abuse of its discretion. . . . [Thus] [c]onclusions of law reached by the administrative agency must stand if the court determines that they resulted from a correct application of the law to the facts found and could reasonably and logically follow from such facts. . . . [Similarly], this court affords deference to the construc- tion of a statute applied by the administrative agency empowered by law to carry out the statute’s purposes.’’ (Internal quotation marks omitted.) Commissioner of Emergency Services & Public Protection v. Freedom of Information Commission, 330 Conn. 372, 379, 194 A.3d 759 (2018). ‘‘Cases that present pure questions of law, however, invoke a broader standard of review . . . . [T]he tradi- tional deference accorded to an agency’s interpretation of a statutory term is unwarranted when the construc- tion of a statute . . . has not previously been subjected to judicial scrutiny [or to] a governmental agency’s time- tested interpretation . . . . Even if time-tested . . . an agency’s interpretation of a statute [is subject to deference] only if it is reasonable . . . [as] determined by [application of] our established rules of statutory construction.’’ (Citation omitted; internal quotation marks omitted.) Id., 379–80. It is well established that § 1-210 (b) (3) (D) has two prongs—that the records are ‘‘to be used in a prospec- tive law enforcement action’’ and that disclosure would be prejudicial. See, e.g., Commissioner of Public Safety v. Freedom of Information Commission, 312 Conn. 513, 545 n.31, 93 A.3d 1142 (2014). Although the first prong has been interpreted in a few instances by the lower courts, those interpretations have not been out- come determinative. Nor has the commission articu- lated a clear, time-tested interpretation of the legal standard required to establish the first prong of the law enforcement exception, particularly with respect to cold case murder investigation records. Our review of that pure question of law is, therefore, de novo. It also is settled law that the custodians of the records —in this case, the respondents—bear the burden of establishing that an FOIA exception applies. ‘‘[T]he overarching legislative policy of [the FOIA] is one that favors the open conduct of government and free public access to government records. . . . [I]t is well estab- lished that the general rule under the [FOIA] is disclo- sure, and any exception to that rule will be narrowly construed in light of the general policy of openness expressed in the [FOIA]. . . . [Thus] [t]he burden of proving the applicability of an exception [to disclosure under the FOIA] rests [on] the party claiming it.’’ (Cita- tion omitted; internal quotation marks omitted.) Lieber- man v. Aronow, 319 Conn. 748, 754–55, 127 A.3d 970 (2015). B We begin our analysis of the law enforcement excep- tion with the plain language of the statute. See General Statutes § 1-2z. Subsection (a) of § 1-210 provides in relevant part: ‘‘Except as otherwise provided by any federal law or state statute, all records maintained or kept on file by any public agency, whether or not such records are required by any law or by any rule or regula- tion, shall be public records . . . .’’ Subsection (b) of § 1-210 carves out various exceptions to that rule of disclosure, and the exception at issue in the present case is codified at § 1-210 (b) (3) (D). Section 1-210 (b) provides in relevant part that ‘‘[n]othing in the Freedom of Information Act shall be construed to require disclo- sure of . . . (3) [r]ecords of law enforcement agencies not otherwise available to the public which records were compiled in connection with the detection or investigation of crime, if the disclosure of such records would not be in the public interest because it would result in the disclosure of . . . (D) information to be used in a prospective law enforcement action if prejudicial to such action . . . .’’5 (Emphasis added.) Although the focus of the present appeal is on the first prong of the test—what it means for a document ‘‘to be used in a prospective law enforcement action’’—we briefly address the prejudice prong in part IV of this opinion. 1 The meaning of the first prong of § 1-210 (b) (3) (D) centers around the word ‘‘prospective.’’ Because the FOIA does not define that term, we look to its ordinary meaning. See, e.g., Meriden v. Freedom of Information Commission, 338 Conn. 310, 322, 258 A.3d 1 (2021). Dictionaries in print in the 1970s, when the statute was amended to include the relevant language; see Public Acts 1975, No. 75-342, § 2 (P.A. 75-342); defined ‘‘pro- spective’’ in one or more of three ways. First, prospec- tive can have a probabilistic meaning, such as ‘‘potential,’’ ‘‘anticipated,’’ ‘‘expected,’’ ‘‘likely,’’ or ‘‘pos- sible.’’ See, e.g., The Random House Dictionary of the English Language (Unabridged Ed. 1966) p. 1155 (defin- ing prospective as ‘‘potential’’); Webster’s New Univer- sal Unabridged Dictionary (Deluxe 2d Ed. 1979) p. 1445 (defining prospective as ‘‘anticipated’’); Webster’s New World College Dictionary (2d Ed. 1972) p. 1141 (defining prospective as ‘‘expected’’ or ‘‘likely’’); Webster’s Third New International Dictionary (1966) p. 1821 (defining ‘‘prospect’’ as ‘‘possibility’’); see also, e.g., Ortiz v. State, 93 S.W.3d 79, 86 (Tex. Crim. App. 2002) (prospective also can mean possible), cert. denied, 538 U.S. 998, 123 S. Ct. 1901, 155 L. Ed. 2d 824 (2003). Second, prospective can mean ‘‘in prospect’’ or ‘‘per- taining to a prospect,’’ as in ‘‘prospective buyers.’’ See, e.g., Webster’s New Universal Unabridged Dictionary, supra, p. 1445 (‘‘pertaining to a prospect’’ or ‘‘perspec- tive’’). The term is used in that manner both with respect to identified or pending prospects, as when we say that ‘‘the prospective buyers have been preapproved for a mortgage,’’ as well as with possible or potential pros- pects, as when we say that ‘‘prospective buyers will find the bonus room a nice feature.’’ Third, prospective can mean looking forward in time or effective in the future. See, e.g., Ballentine’s Law Dictionary (3d Ed. 1969) p. 1014 (‘‘[l]ooking to the future’’). This definition is commonly used in distin- guishing statutes that apply retroactively from those that have solely prospective effect; see, e.g., id.; Thomp- son v. Hagan, 96 Idaho 19, 25, 523 P.2d 1365 (1974); and we fail to see how the legislature plausibly could have intended this definition to govern the law enforce- ment exception, insofar as arrests and prosecutions cannot be retrospective or past oriented in the way that laws can be. Cf. Brayman Construction Corp. v. Commonwealth, Dept. of Transportation, 608 Pa. 584, 598 n.10, 13 A.3d 925 (2011) (concluding that ‘‘ ‘effective in the future’ ’’ definition of ‘‘prospective’’ could not be applicable in context of case). A plausible argument can be made, though, that either of the first two definitions is what the legislature intended, and, even within those definitions, there is a range of possible meanings. Accordingly, we conclude that the statute is facially ambiguous. The trial court adopted the first, probabilistic defini- tion, concluding that the statute, by its express terms, requires a prediction as to the probability of a future law enforcement event. Recognizing that this definition of ‘‘prospective’’ includes within it a broad range of probabilities, lying along a spectrum, the trial court rejected a legal standard that lies at either extreme. The court reasoned that requiring the police to demonstrate that a law enforcement action is likely or probable would impose an unreasonable burden but that requir- ing only a speculative possibility would allow the excep- tion to swallow the rule. Accordingly, the court concluded that a prospective law enforcement action is one that is at least a reasonable possibility, which, it explained, is more than theoretically possible, but not necessarily likely or probable, to occur.6 We agree with the trial court that this is the most reasonable reading of the statute. Because § 1-210 (b) seeks to strike a balance between competing public goods, fostering openness and transparency while pro- tecting important governmental functions that demand a degree of confidentiality, it makes sense that the legis- lature would have carved out an exception only for law enforcement actions that are, at the very least, reason- able possibilities. In at least one instance, the commis- sion has articulated that this is the standard that it applies. See Graeber v. Chief, New Haven Police Dept., Freedom of Information Commission, Docket No. FIC 2016-0865 (September 27, 2017).7 The law enforcement exception plausibly could be read to impose an even more exacting standard, such that records would be exempt only if an arrest or prose- cution is pending or likely. That would be consistent with the overarching principle that FOIA exceptions should be narrowly construed in favor of disclosure. See, e.g., Lieberman v. Aronow, supra, 319 Conn. 754– 55. Like the trial court, however, we would require a clear statement of legislative intent before concluding that the legislature wanted to require law enforcement agencies to disclose sensitive and potentially prejudicial investigatory information in the early stages of a homi- cide investigation or when there are promising leads that, although not enough to render prosecution likely, make it eminently possible. At the same time, we agree with the trial court that requiring that the police demonstrate only a remote, speculative, or theoretical possibility of some future law enforcement action would permit the exception to swallow the rule and would fly in the face of the well established principle that FOIA exceptions are to be narrowly construed. We have recognized that ‘‘[t]he overarching legislative policy of the [FOIA] is one that favors the open conduct of government and free public access to government records. . . . The sponsors of the [FOIA] understood the legislation to express the people’s sovereignty over the agencies [that] serve them . . . and this court consistently has interpreted that expression to require diligent protection of the public’s right of access to agency proceedings. Our construction of the [FOIA] must be guided by the policy favoring disclosure and exceptions to disclosure must be nar- rowly construed.’’ (Internal quotation marks omitted.) Stamford v. Freedom of Information Commission, 241 Conn. 310, 314, 696 A.2d 321 (1997). With the thumb on the scale in favor of disclosure, it would be inconsistent with the purpose of the FOIA to conclude that the law enforcement exception applies upon a showing that a law enforcement action is only remotely or theoreti- cally possible. Further support for the reasonable possibility stan- dard can be found in a related statute within the FOIA. General Statutes § 1-215 establishes a default rule that records of arrest are public records subject to disclo- sure pursuant to the FOIA. See General Statutes § 1-215 (b). Included among information that must be redacted from such records prior to disclosure, however, is ‘‘spe- cific information about the commission of a crime, the disclosure of which the law enforcement agency reason- ably believes may prejudice a pending prosecution or a prospective law enforcement action . . . .’’ (Emphasis added.) General Statutes § 1-215 (b) (3). On the one hand, the presentation of the alternatives of a ‘‘pending prosecution’’ or a ‘‘prospective law enforcement action’’ in this related statute supports the conclusion that the legislature viewed the term ‘‘prospective’’ in § 1-210 (b) (3) (D) as signifying a degree of probability that is less than ‘‘pending.’’ Otherwise, the term ‘‘prospective’’ would be superfluous. See, e.g., American Promotional Events, Inc. v. Blumenthal, 285 Conn. 192, 203, 937 A.2d 1184 (2008) (‘‘[i]nterpreting a statute to render some of its language superfluous violates cardinal prin- ciples of statutory interpretation’’); see also, e.g., id. (‘‘[i]n construing statutes, we presume that there is a purpose behind every sentence, clause, or phrase used in an act and that no part of a statute is superfluous’’ (internal quotation marks omitted)). On the other hand, the interpretive principle of noscitur a sociis—a word is known by the company it keeps—suggests that ‘‘pro- spective’’ is not entirely unlike ‘‘pending’’ for purposes of the statutory scheme. Indeed, if a ‘‘prospective law enforcement action,’’ for purposes of § 1-215, meant any remotely possible action, then the word ‘‘pending’’ would become unnecessary. Insofar as the statutory language is ambiguous, we also may consider the legislative history of § 1-210 (b) (3) (D). See General Statutes § 1-2z. ‘‘Like the federal Freedom of Information Act . . . our state FOIA was enacted in the aftermath of the Vietnam War and [the] Watergate [scandal] . . . [when] people were fed up with furtive government and had los[t] faith in govern- ment and politicians.’’ (Citation omitted; internal quota- tion marks omitted.) Commissioner of Mental Health & Addiction Services v. Freedom of Information Com- mission, 347 Conn. 675, 706 n.16, 299 A.3d 197 (2023). Prior to the 1975 amendments to the predecessor stat- ute, the law enforcement exception was broadly worded and exempted any ‘‘investigatory files compiled for law enforcement purposes, except to the extent available by law to a private citizen . . . .’’ General Statutes (Rev. to 1975) § 1-19. Among other changes, P.A. 75-342, § 2, enacted in the wake of, and in response to, the Watergate scandal, dramatically scaled back the scope of the law enforcement exception, adding the current language limiting the exception to information to be used in, and prejudicial to, a prospective law enforcement action. The legislative history of P.A. 75-342, § 2, is replete with statements indicating that the intent of the spon- sors was to increase transparency, give the FOIA sharper teeth, and craft a bill that could stand side by side with the strongest freedom of information laws in the country. For example, Representative Raymond J. Dzi- alo, the sponsor of the bill in the House of Representa- tives, explained that ‘‘[t]he [s]tate’s existing right to know law has been ineffective. It has been unable to protect the public . . . unable to mandate that records are available . . . . This [b]ill before you embodies the significant, no I must say, sweeping changes [to] the present law.’’ (Internal quotation marks omitted.) 18 H.R. Proc., Pt. 8, 1975 Sess., p. 3895.8 Indeed, the authors of P.A. 75-342, § 2, took the extraordinary measure of drafting an informal preamble that, although not included in the FOIA itself, was high- lighted during the legislative debates to emphasize the purpose of the act: ‘‘[T]he [l]egislature finds and declares that secrecy in government is inherently inconsistent with [a] true democracy. That the people have a right to be fully informed of the actions taken by public agencies in order that they may retain control over the instruments they have created. . . . That the people in delegating authority do not give their public servants the right to decide what is good for them to know and that it is the intent of this [l]aw that . . . the records of all public agencies be open to the public except in those instances [in which a] superior public interest requires confidentiality.’’ (Internal quotation marks omitted.) Id., p. 3911, remarks of Representative Martin B. Burke. As the Senate sponsor, Senator Robert L. Julianelle summarized the intent of the amendments: ‘‘[W]e do not have any right to withhold, except with the agreement of the people, any information, any pro- ceedings, anything at all, that we do for the benefit of the people.’’ 18 S. Proc., Pt. 5, 1975 Sess., p. 2325. He further remarked that ‘‘it is the intent of the law that actions taken by public agencies be taken openly and their deliberations be conducted openly and that the record of all public agencies be open to the public except in those instances [in which] a superior public interest requires confidentiality. . . . We exclude the right to public records in only some very serious areas . . . [that are] sacred and should [not be] subject to the law.’’ Id., pp. 2323–24. When this court previously reviewed this legislative history, it observed that, although the FOIA evidences ‘‘a strong legislative policy in favor of the open conduct of government and free public access to government records,’’ it ‘‘does not confer [on] the public an absolute right to all government information.’’ Wilson v. Freedom of Information Commission, 181 Conn. 324, 328, 435 A.2d 353 (1980). Rather, this court concluded, this legis- lative history ‘‘reflects a legislative intention to balance the public’s right to know what its agencies are doing, with the governmental and private needs for confidenti- ality.’’ Id. With respect to the exceptions to the disclosure requirement, the legislative history reinforces this idea that the legislature sought to strike a balance: open access to public records must be balanced against cer- tain compelling public interests in confidentiality. See, e.g., 18 S. Proc., supra, p. 2324, remarks of Senator Julianelle; see also, e.g., 18 H.R. Proc., supra, p. 3910, remarks of Representative Burke (‘‘There should be no mistake about the legislative intent of this bill. . . . [A]ll records of all governmental agencies shall be in the public domain with few and very precise exceptions.’’). Both the text of the statute and its legislative history thus rule out the most extreme readings of the term ‘‘prospective’’—requiring that the law enforcement agency establish that there is a pending law enforcement action or that it demonstrate that there is a mere theoretical possibility of one. With respect to the remaining possi- ble meanings of ‘‘prospective’’; i.e., probable, likely, or possible; we rely on the principle that the FOIA effects a balance between the competing interests and conclude that the legislature intended the law enforcement exception to apply only when a law enforcement agency is able to make the threshold showing that an arrest or prosecution is at least reasonably possible. It need not be probable or likely, but it must be more than only remotely or theoretically possible. This standard effectuates the legislative intent of providing open access to public records without unduly hamstringing ongoing investigations. 2 The law enforcement exception is limited to records ‘‘to be used’’ in a prospective law enforcement action. General Statutes § 1-210 (b) (3) (D). Although the trial court did not discuss this element of the statutory lan- guage, and the parties have not addressed it directly, we briefly touch on it here because the Appellate Court has construed it in a manner that has led the commis- sion to misapply the ‘‘prospective law enforcement action’’ requirement of § 1-210 (b) (3) (D). In Dept. of Public Safety v. Freedom of Information Commission, 51 Conn. App. 100, 720 A.2d 268 (1998), the Appellate Court concluded that ‘‘the statute is not satisfied and, consequently, information is not exempted from disclosure by the mere good faith assertion that the matter to which the information pertains is poten- tially criminal. [Rather], there must be an evidentiary showing that the actual information sought is going to be used in a law enforcement action . . . .’’ (Emphasis added.) Id., 105. Judges of the Superior Court have embraced and relied on this language. See, e.g., Nastro v. Freedom of Information Commission, Docket No. HHB-CV-XX-XXXXXXX-S, 2008 WL 3852748, *3 (Conn. Super. July 23, 2008); see also, e.g., Tuccitto v. Dept. of Public Safety, Division of State Police, Freedom of Information Commission, Docket No. FIC 2004-029 (October 27, 2004) (relying on Appellate Court’s deci- sion in Dept. of Public Safety for proposition that law enforcement exception ‘‘requires an evidentiary show- ing that the records are in fact to be used in a prospec- tive law enforcement action’’). Under this standard, respondents would have to establish that it is at least probable, if not certain, that the subject records would be used for an arrest or prosecution. We agree with the trial court that the Appellate Court’s analysis of the law enforcement exception in Dept. of Public Safety was dictum. That case involved an accidental drowning, and, by the time the document request had been made, it was clear that there had been no criminal conduct. See Dept. of Public Safety v. Freedom of Information Commission, supra, 51 Conn. App. 103. A law enforcement action was neither pending nor even reasonably possible. The statement of the legal standard in Tuccitto was likewise dictum, although for the opposite reason. In that case, a prosecution was pending, and the only question was whether, under the prejudice prong of § 1-210 (b) (3) (D), release of the investigation records was likely to bias potential jurors. See Tuccitto v. Dept. of Public Safety, Division of State Police, supra, Docket No. FIC 2004-029. In both cases, then, the result would have been the same under any plausible standard. In any event, we disagree with the standard annun- ciated in Dept. of Public Safety. It would make little sense for the legislature, having required only that a prospective law enforcement action be reasonably pos- sible, to then impose a far more exacting requirement under which respondents must demonstrate that it is probable or certain that they will use the requested information in such an action. When possible, we must construe statutes to create a coherent and cohesive scheme. See, e.g., State v. Victor O., 320 Conn. 239, 249, 128 A.3d 940 (2016). In the absence of any further guidance, we will presume that the same standard gov- erns both elements of the first prong of the statutory exception, and that the respondents must establish for any requested document only that it is at least reason- ably possible that the information contained therein will be used in support of an arrest or prosecution. To the extent that Dept. of Public Safety imposed a differ- ent standard, it is overruled. C We next consider certain policy arguments advanced by the respondents and/or the Division of Criminal Jus- tice as amicus curiae. The division contends that the reasonable possibility standard should be deemed satis- fied whenever a law enforcement ‘‘investigation is open, a suspect has been identified, and no insurmountable obstacles exist to a future prosecution.’’ In support of this three factor standard, the division describes a hand- ful of cold cases in which the police ultimately were able to make arrests, many years after the crime, when new witnesses or suspects emerged who made state- ments that included details that only the killer or killers could have known. If those case files had been made public, the division argues, prosecutions would have been ‘‘grievously, if not fatally, prejudiced . . . .’’ We recognize that some added leeway must be afforded for investigations of crimes such as homicide. The divi- sion, however, fails to identify any support for its pro- posed rule in the language or legislative history of the statute. Moreover, as discussed, in construing FOIA exceptions, we long have eschewed a categorical approach that fails to account for the specific facts and context of each individual case. See, e.g., Director, Retirement & Benefits Services Division, Office of the Comptroller v. Freedom of Information Commission, 256 Conn. 764, 779, 775 A.2d 981 (2001). The division’s proposed rule would follow that impermissible path, treating the identification of a suspect as dispositive, for example, without considering key factors such as the strength of the evidence pointing to that suspect and whether, with the passage of significant time, the early identification of a suspect may become less mean- ingful. The same can be said for the fact that an investigation remains open and free of insurmountable obstacles to a future prosecution. As a result, we are not prepared to say that those three factors, without more, will always satisfy the reasonable possibility standard. Depending on the context, they may establish no more than a theoretical possibility. As we explained, a theoretical possibility standard would not be consistent with the overarching principles underlying the FOIA, would upset the careful balance the legislature sought to strike, and would be incompatible with what we believe to be the fairest reading of the statutory requirement that the information being sought is to be used in a prospective—that is, at least reasonably possible—law enforcement action. Moreover, to the extent that policy considerations such as those raised by the division are relevant,9 we are concerned that this interpretation of the statute places too much weight on the law enforcement side of the scale. Adopting the rule advanced by the respon- dents and the division would preclude the public from exercising any effective oversight, not only in murder cases, but in all cases in which any applicable statute of limitations has not run. This position also ignores the fact that the legislature through this exception sought to balance law enforcement’s interests in solving crimes and bringing wrongdoers to justice against more than just the public’s interests in transparency and oversight. With the passage of time, it becomes increasingly likely that openness, rather than secrecy, is what will unearth that elusive lead that will help the police solve the case. As the trial court explained, after having reviewed the relevant research, ‘‘a fresh pair of eyes’’—potentially millions of eyes in the age of Internet crowdsourcing—‘‘is far more likely to improve the odds of an arrest and successful prosecution than it is to prejudice that out- come.’’ (Internal quotation marks omitted.) Although the division tells a compelling tale of cases in which the police were able to solve long cold murders because investigatory records were not ‘‘[p]remature[ly]’’ disclosed, we are confident that, for every such case, there is an equally powerful, countervailing pragmatic case for transparency. Some of the most paradigmatic instances in which law enforcement files were inappro- priately shielded from public view, and heinous crimes were solved only when sunlight finally broke through, can be found in the civil rights era of the 1950s and 1960s. The history of that era is replete with incidents in which law enforcement agencies failed to diligently investigate the murders of Black victims and then resisted disclosing investigation records that would have both shed light on their own indolence, if not complicity, and allowed victims’ families and outside organizations, sometimes with the assistance of docu- mentary filmmakers, to help bring the killers to justice. See generally J. McDonald, ‘‘Heroes or Spoilers? The Role of the Media in the Prosecutions of Unsolved Civil Rights Era Murders,’’ 34 Ohio N.U. L. Rev. 797 (2008). These examples demonstrate that a theoretically possi- ble standard poses too great a risk of unfettered and unchecked police discretion. Indeed, as we mentioned in part II B 1 of this opinion, the underlying purpose of the FOIA was the recognized need to increase govern- ment accountability in the wake of Watergate. The legis- lative history of the FOIA makes clear that the legislature has recognized a particular need for robust oversight of local law enforcement agencies.10 The respondents and the division also contend that, as a matter of public policy, we should defer to the expertise of law enforcement agencies and officials in such matters. Once again, however, nothing in the statu- tory scheme or the legislative history gives any indica- tion that the legislature intended that the commission should defer to the expertise of law enforcement agen- cies or officials when construing or applying the law enforcement exception. Quite the contrary, in the con- text of exceptions to disclosure of records, the legisla- ture has been clear when it intends to give the agency possessing the subject records increased deference. Several FOIA exceptions expressly provide that the commission must defer to the determination of the respondent agency or official as to whether the excep- tion is satisfied. See, e.g., General Statutes § 1-210 (b) (1) (exception for preliminary drafts or notes when ‘‘the public agency has determined that the public interest in withholding such documents clearly outweighs the public interest in disclosure’’); General Statutes § 1-210 (b) (18) (exception for correction records that Commis- sioner of Correction has reasonable grounds to believe may result in safety risk if disclosed); General Statutes § 1-210 (b) (19) (exception for other records reasonably likely to result in safety risk if disclosed, as determined by identified public agencies); General Statutes § 1-210 (b) (24) (exception for responses to bids or proposals, so long as public agency certifies that public interest in disclosure is outweighed by public interest in confi- dentiality). The legislature could have drafted § 1-210 (b) (3) in the same manner as those exceptions, granting the law enforcement agency or official broad discretion to determine whether the public interest in disclosure is outweighed by the public interest in confidentiality. It did not. D We next consider how the reasonable possibility stan- dard that we articulated applies to law enforcement records compiled in connection with the investigation of a murder, especially in cases, such as the present case, in which years have passed and the leads are few and far between. Although such cases implicate the same balancing of interests present in any FOIA request for law enforcement records, they also involve distinct considerations. In particular, the fact that there is no statute of limitations means that unsolved murder inves- tigations may remain open, at least nominally, forever. Most crimes, by contrast, feature a statute of limitations of five years or less, which means that, by the time an investigation has gone cold, a law enforcement action will no longer be possible. See General Statutes § 54- 193 (c) and (d); see also footnote 2 of this opinion. We emphasize that the commission has publicly announced its strong inclination to trust the representations of law enforcement officials in these matters.11 At the same time, however, the commission has taken seriously this court’s repeated admonition that ‘‘the claimant of [an FOIA] exemption [must] provide more than conclusory language, generalized allegations or mere arguments of counsel.’’ (Internal quotation marks omitted.) Director, Retirement & Benefits Services Division, Office of the Comptroller v. Freedom of Information Commission, supra, 256 Conn. 773; see, e.g., Hartford v. Freedom of Information Commission, 201 Conn. 421, 434–35, 518 A.2d 49 (1986) (‘‘[t]he commission is not obliged to accept an agency’s generalized and unsupported allega- tions relating to documents claimed to be exempt from disclosure’’ (internal quotation marks omitted)). For murder and, presumably, other crimes that involve lengthy or no statutes of limitations,12 the trial court adopted and applied a nonexclusive, multifactor test to determine whether a future law enforcement action is reasonably possible. Although we agree that the factors identified by the trial court are relevant to the analysis, we caution that individual factors may vary in importance from case to case, and the court’s test is not intended to serve as a complete or mechanis- tic checklist. Indeed, another relevant consideration that the trial court implicitly or explicitly considered, but did not expressly list, is the empirical likelihood that crimes of this sort will be solved after years of fruitless investigation. Pursuant to this multifactor test, the commission must consider, among other things, ‘‘(1) the length of time that has elapsed since the commission of the crime; (2) the length of time that has elapsed since the law enforcement agency last obtained significant new evi- dence or leads, i.e., whether the agency has effectively exhausted all investigatory leads; (3) whether the agency has classified the investigation, even if technically open, as a cold case or the functional equivalent thereof; (4) the number of investigators currently assigned to the investigation; (5) the amount of time investigators cur- rently commit to the investigation; (6) whether the agency has a suspect and, if so, whether the agency’s suspicion is supported by more than speculation; [and] (7) whether advances in [science] or technology, such as advances in DNA analysis, may lead to new evidence or permit the fruitful reexamination of existing evidence.’’ Other courts have considered similar factors in com- parable cases. See, e.g., Chastant v. Prudential Ins. Co. of America, Docket No. 11-CV-626, 2011 WL 4007863, *4–5 (W.D. La. September 8, 2011) (applying Louisiana law); see also, e.g., Dept. of Kentucky State Police v. Teague, Docket No. 2018-CA-000186-MR, 2019 WL 856756, *2 (Ky. App. February 22, 2019). In our view, such factors should be helpful to the commission in carrying out the legislative intention that the commis- sion strike a balance between the public’s right to access government information and the legitimate interests of law enforcement agencies and the state in investigating and prosecuting crimes. In applying these factors, both independently and in relation to each other, sight should not be lost of the two fundamental questions that underlie them. First, does the law enforcement agency continue to actively and earnestly investigate the crime, or has the investiga- tion, although technically still open, essentially gone cold? In the initial months and years following a homi- cide, when significant resources are being dedicated to a case and new leads are regularly pursued, it generally is not difficult for a law enforcement agency to establish that arrest and prosecution are at least reasonably pos- sible. The commission has deferred heavily to the repre- sentations of law enforcement officials under such circumstances. See, e.g., Hoda v. Chief, New Haven Police Dept., Freedom of Information Commission, Docket No. FIC 2007-143 (January 23, 2008) (finding that law enforcement exception was satisfied when FOIA request was made approximately five months after murder and police were actively pursuing leads and suspect); Rouen v. Chief, Groton Police Dept., Free- dom of Information Commission, Docket No. FIC 2006- 064 (January 24, 2007) (finding that law enforcement exception was satisfied when FOIA request was made approximately thirteen months after murder and police were still actively investigating); Poitras v. Chief, Port- land Police Dept., Freedom of Information Commission, Docket No. FIC 1998-085 (August 12, 1998) (deferring to state’s attorney and finding that law enforcement exception was satisfied when FOIA request was ‘‘made very early in the [murder] investigation’’). At some point, however, when years have passed, solid leads have dried up, and the case has been classi- fied (whether formally or de facto) as cold, it may no longer be reasonable for the commission to defer to representations of law enforcement agencies, without more. The commission has long recognized this as well. See, e.g., Estate of Mazzotta v. Chief, Middletown Police Dept., Freedom of Information Commission, Docket No. FIC 2012-033 (November 14, 2012) (requiring disclo- sure of records in homicide investigation that had gone dormant after nine years); Gura v. Chief, New Haven Police Dept., Freedom of Information Commission, Docket No. FIC 2001-147 (February 13, 2002) (requiring disclosure three years after homicide when respondent merely represented that investigation was ongoing); see also, e.g., Donovan v. Greenwich Police Dept., Freedom of Information Commission, Docket No. FIC 87-173 (February 26, 1992) (following in camera review, requir- ing disclosure eight years after homicide, even though investigation remained active). It is largely in the gray areas between these two sets of cases that specific factors, such as those identified by the trial court, will be most illuminating. Second, once a case has gone cold, does there remain a reasonable possibility that the investigation ultimately will culminate in some law enforcement action? This might be true, for example, if the police recently have unearthed significant new evidence or leads, if they can identify emerging technologies with the reasonable potential to move the case forward, or if other factors, such as the existence of evidence tending to incriminate a particular suspect, suggest a reasonable possibility of a law enforcement action. The commission is free to conclude, consistent with its prior practice, that the longer an investigation has been cold, in name or effect, the greater the burden on the respondents to affirma- tively demonstrate that a prospective law enforcement action remains a reasonable possibility. III We now consider whether the existing administrative record is sufficient to permit this court to apply the newly adopted reasonable possibility standard as a mat- ter of law. The trial court, in affirming the decision of the commission, concluded, as a matter of law, that only one outcome was possible. The court concluded that the case had gone cold and that the respondents had failed to provide any testimony or other evidence, beyond pure speculation, that would satisfy the new standard that it had articulated and to establish anything more than a remote or speculative possibility that the investigation will bear fruit. The court also concluded that, in light of the available statistics, there is no more than a theoretical possibility that the police will bring the perpetrator to justice when a case has been cold for this many years. On the basis of the record, we do not agree. The respondents bore the burden before the commission of establishing not only that there was a reasonable possibility that the investigation will result in a law enforcement action, but also that, for each individual document or set of documents sought to be withheld, it is reasonably possible that the requested files contain information that will be used in such a law enforcement action and that disclosure of that information would be prejudicial. In support of their assertion that the law enforcement exception applied, the respondents established the fol- lowing through Sudock’s testimony: The case remains open, and the police have identified a suspect. Sudock typically works on the case at least once a month and periodically receives new leads, which he then investi- gates. The police department monitors advances in forensic technology, including DNA technology, that could yield a break in the case. Although the commission found this evidence to be insufficient to satisfy the respondents’ burden, we find it significant that the commission’s final decision was predicated, as we explain hereinafter, on the clearly erroneous factual finding that the police department had not identified a suspect and, potentially, on the application of a standard that required the respondents to demonstrate either an actual or pending law enforce- ment action. Because application of the reasonable pos- sibility standard that we adopted in this opinion is fact intensive, we conclude that the case must be remanded to the commission for further proceedings. See, e.g., Great Plains Lending, LLC v. Dept. of Banking, 339 Conn. 112, 131–32, 143, 157, 259 A.3d 1128 (2021) (remand to Banking Commissioner for further proceed- ings was required to determine whether plaintiff was entitled to status as ‘‘arm of the tribe’’ for purposes of tribal sovereign immunity because minimal evidence in administrative record was insufficient to permit court to apply newly adopted standard as matter of law); Ann Howard’s Apricots Restaurant, Inc. v. Commission on Human Rights & Opportunities, 237 Conn. 209, 233, 676 A.2d 844 (1996) (remanding for further proceedings in which complainant would have opportunity to pres- ent additional admissible evidence). We reach this con- clusion for three reasons. A First, we agree with the respondents that the commis- sion’s decision rested in part on clearly erroneous fac- tual findings for which there was no substantial evidence in the record. ‘‘Judicial review of an adminis- trative agency decision requires a court to determine whether there is substantial evidence in the administra- tive record to support the agency’s findings of basic fact and whether the conclusions drawn from those facts are reasonable.’’ (Internal quotation marks omit- ted.) Amaral Bros., Inc. v. Dept. of Labor, 325 Conn. 72, 84–85, 155 A.3d 1255 (2017). In the present case, the commission twice stated in its decision that ‘‘the . . . police [department has] not identified any sus- pects . . . .’’ This finding was without support in the record and, in fact, was contradicted by the undisputed testimony of both witnesses.13 Sudock testified that the police department had enough DNA to develop suspects. He further testified that the department had long been close to solving the crime and that it had a ‘‘number one’’ suspect. He confirmed specific details about that suspect, most notably that the suspect’s cell phone had been unchar- acteristically turned off during a twenty-four hour period around the time of the homicide. Hamburg corroborated Sudock’s testimony. He testi- fied that he had been informed that the DNA evidence in the police department’s possession ‘‘was enough to develop suspects . . . .’’ He recalled that, during the 2019 meeting with Sudock, he was informed that the department had had the same ‘‘number one’’ suspect since March 7, 2010, and he indicated that Sudock had asked repeatedly about one relative’s potential involve- ment in the crime. He also confirmed that Sudock had shared information regarding the suspect’s cell phone records. In the absence of any conflicting testimony or deter- mination that both witnesses lacked credibility, the commission’s statement that the police department had not identified any suspects in the homicide was clearly erroneous. Although the department may not have pub- licly named any suspects, it had identified at least one.14 The lack of evidence to support the commission’s find- ings, however, does not end our inquiry. ‘‘Substantial prejudice [arising from the commission’s clearly errone- ous findings] must be affirmatively shown.’’ (Internal quotation marks omitted.) Lawrence v. Kozlowski, 171 Conn. 705, 714, 372 A.2d 110 (1976), cert. denied, 431 U.S. 969, 97 S. Ct. 2930, 53 L. Ed. 2d 1066 (1977); see also General Statutes § 4-183 (j). Although it is a close call, we agree with the respondents that this standard is satisfied. The commission distinguished the present case from two previous cases in which it had found that release of requested records would be prejudicial to a prospective law enforcement action, largely on the basis that, in those cases, the police department had identified one or more suspects. See footnote 14 of this opinion. The erroneous statement that the department had not identi- fied any suspects thus appears to have been important to the commission’s determination that the department failed to provide nonspeculative evidence that the requested records would be used in, and their disclo- sure would be prejudicial to, a prospective law enforce- ment action. As we noted in part II D of this opinion, the identification of a suspect is one of the key factors that the commission may consider when determining whether a law enforcement action is a reasonable possi- bility. Accordingly, the case must be remanded to give the commission an opportunity to consider what weight and importance, if any, to give to the testimony that the department identified a prime suspect early in the investigation.15 B Second, although it is not entirely clear, it does not appear that the commission applied the reasonable pos- sibility standard in the present case. In its final decision, the commission repeatedly identified as its basis for declining to find that the respondents had established that there was a prospective law enforcement action their failure to present evidence of a ‘‘specific,’’ ‘‘expected,’’ ‘‘identified,’’ or ‘‘anticipated’’ law enforcement action. The commission’s proposed final decision, which was prepared by the hearing officer, went even further, stat- ing that Sudock ‘‘could not identify an actual prospec- tive law enforcement action and could only provide speculation.’’ (Emphasis added.) Although the commis- sion ultimately removed the word ‘‘actual’’ from its final decision, it is unclear whether this represented a sub- stantive change. In short, the commission’s final deci- sion could be understood to require a probability, even a certainty, that the requested records will be used for an arrest or prosecution.16 Such a stringent standard is not consistent with the plain language of the statute. Moreover, although deference is due to the commis- sion’s factual findings, in this instance, its findings, such as that the respondents’ claims were wholly speculative, are conclusory and may well be tied up with the legal standard that it applied. Application of the new standard to the existing record, therefore, would run the risk of this court’s ‘‘substitut[ing] its judgment for that of the agency as to the weight of the evidence on questions of fact,’’ contrary to the requirements of § 4-183 (j). On remand, the commission will have the opportunity to apply the new standard we adopted to assess whether some law enforcement action remains a reasonable pos- sibility at this juncture. C Third, we note that, in concluding that a law enforce- ment action was not reasonably possible, the trial court relied on a body of empirical, statistical evidence sug- gesting that the probability of a prosecution in cold murder investigations such as this ‘‘approaches zero’’ and that, at some point, ‘‘a ‘fresh pair of eyes’ . . . is far more likely to improve the odds of an arrest and successful prosecution than it is to prejudice that out- come.’’ The commission was the finder of fact in this matter, however, and it should have the opportunity to review any available statistical data in the first instance, with the input of the parties and any expert testimony they care to submit, before making relevant findings.17 Likewise, should the commission choose to follow the trial court in adopting a more statistical, probabilistic approach to such cases, it may consider, in the first instance, what sort of showing is necessary to establish a reasonable possibility. See, e.g., Immigration & Natu- ralization Service v. Cardoza-Fonseca, 480 U.S. 421, 431, 440, 107 S. Ct. 1207, 94 L. Ed. 2d 434 (1987) (10 percent chance was sufficient to establish reasonable possibility); Lumataw v. Holder, 582 F.3d 78, 92 (1st Cir. 2009) (same). For these reasons, we conclude that a remand for further proceedings in light of the legal standard that we articulated in this opinion is required. IV There remains the issue of prejudice. The commis- sion determined not only that the respondents had failed to establish that any requested records were to be used in a prospective law enforcement action, but also that they had not established that disclosure of the records would be prejudicial to such action. The respondents challenged both determinations on appeal to the trial court. Having affirmed the commission’s decision on the basis of the former determination, the trial court declined to address the latter. The respondents cannot prevail unless they satisfy both prongs of the statutory exception. See, e.g., Com- missioner of Public Safety v. Freedom of Information Commission, supra, 312 Conn. 545 n.31. So long as the commission’s determination as to a lack of prejudice stands, the respondents can obtain no practical relief with respect to the issue of a prospective law enforce- ment action, and any proceedings on remand would be moot. See, e.g., In re Jorden R., 293 Conn. 539, 557, 979 A.2d 469 (2009). Ordinarily, then, our conclusion that the commission failed to properly analyze the prospective law enforce- ment action prong of the statute would require that we remand the case to the trial court to give it the opportunity to consider the prejudice issue. Cf. Equity One, Inc. v. Shivers, 310 Conn. 119, 125 n.2, 74 A.3d 1225 (2013). Under the circumstances of the present case, however, we conclude that further review by the trial court is unnecessary. First, the two prongs of the statutory exception, although distinct, are not wholly unrelated. Several of the factors that the trial court identified as relevant to the ‘‘prospective law enforcement action’’ prong of § 1- 210 (b) (3) (D), such as the existence of a suspect and the availability of DNA evidence for future testing, also go to prejudice. The commission’s reconsideration of these matters will necessarily bear on prejudice as well. Second, it seems likely that the commission also applied the wrong legal standard as to the prejudice prong. In concluding that there was only speculative evidence that disclosure of the requested information would be prejudicial to a prospective law enforcement action, the commission appears to have (1) relied on its conclusion that the respondents failed to present evidence of a ‘‘specific,’’ ‘‘expected,’’ ‘‘identified,’’ or ‘‘anticipated’’ law enforcement action, thus incorrectly conflating the two prongs of the test, and (2) taken too literally Sudock’s offhand statement that he could ‘‘go on with speculating’’ as to how the requested informa- tion might be used. Sudock testified that some of the requested records contain information that only the perpetrator could know, and that such information, if released, could prej- udice any future prosecution. In past cases, the commis- sion has treated such general representations as sufficient to satisfy the prejudice prong of the test, at least when borne out by the commission’s in camera review. See, e.g., Graeber v. Chief, New Haven Police Dept., supra, Docket No. FIC 2016-0865; Lopez v. Chief, Bridgeport Police Dept., Freedom of Information Com- mission, Docket No. FIC 2015-398 (February 24, 2016). If the respondents’ prejudice showing fell short, it was not because Sudock failed to testify as to exactly where and how the documents would be used—requiring that level of prognostication would be unreasonable. On remand, any documents that the respondents submit for in camera review as containing potentially prejudicial information should be reviewed under the commis- sion’s established standards. The judgment is reversed and the case is remanded with direction to remand the case to the commission for further proceedings according to law. In this opinion the other justices concurred. * This case originally was argued before a panel consisting of Chief Justice Robinson, Justices D’Auria, Mullins and Ecker, and Judge Elgo. Thereafter, Justices McDonald and Dannehy were added to the panel and have read the briefs and appendices, and listened to a recording of the oral argument prior to participating in this decision. 1 Because of the potential for confusion—several members of the Hamburg family are involved in this case, as is the town of Madison and its police department—we refer to Barbara Beach Hamburg as the victim, to Madison Hamburg as Hamburg, and to other members of the Hamburg family by their full name. 2 Sudock testified that, during a cold case review, ‘‘another group of indi- viduals . . . look at a case, evaluate what investigative steps have [been taken, and potentially suggest] a different route to go or other evidence to send . . . to the [state forensic science laboratory] . . . .’’ See Connecticut Division of Criminal Justice, Cold Case Investigations (2024), available at https://portal.ct.gov/DCJ/Programs/Programs/Cold-Case-Investigations (last visited February 16, 2024) (describing cold case unit). Whereas other jurisdictions variously define a cold case homicide as one that has not been solved after a period of time ranging from one to five years after the crime and for which no viable and significant unexplored leads remain to investigate; see National Institute of Justice, Office of Justice Programs, U.S. Dept. of Justice, National Best Practices for Implementing and Sustaining a Cold Case Investigation Unit (July, 2019) pp. 1–2, available at https://www.ojp.gov/pdffiles1/nij/252016.pdf (last visited February 16, 2024); in Connecticut, the Division of Criminal Justice defines a cold case simply as one that is ‘‘unsolved for a prolonged period of time.’’ Connecticut Division of Criminal Justice, supra. 3 At the hearing, Sudock acknowledged that all of the documents at issue were responsive to Niemeyer’s request, but he took the position that none of them should be turned over. His stated understanding was that the police department has a policy of not releasing any records in open investigations. Ultimately, however, the department did release certain documents related to the investigation, at various times, both to Niemeyer and to other indi- viduals. 4 The commission did not order the respondents to provide copies of the signed witness statements, which are protected under a different FOIA exception. See General Statutes § 1-210 (b) (3) (C). Accordingly, the only categories of documents that are at issue in this appeal are the investigatory records, interrogation transcripts and recordings, and crime scene recordings and related documents. 5 The parties do not dispute that the first set of statutory requirements is satisfied; all of the documents at issue are records of a law enforcement agency that are not otherwise available to the public and that were compiled in connection with the investigation of a crime. 6 The trial court twice indicated that it had given particular weight to the decisions of federal courts that, when construing the corresponding—albeit differently worded—provisions of the federal Freedom of Information Act; see 5 U.S.C. § 552 (b) (7) (A) (2018); applied a ‘‘pending or reasonably anticipated’’ standard. (Emphasis omitted; internal quotation marks omit- ted.) In other places, the trial court indicated that it was adopting the ‘‘reasonable possibility’’ standard, which both the commission and Niemeyer had advocated for, and we understand that to be the standard the court ultimately adopted. The trial court never defined that standard, however, other than to say that it sits somewhere on the spectrum between ‘‘a mere theoretical possibility’’ and ‘‘likely or probable.’’ With respect to the federal cases, it is worth noting that, although the United States Court of Appeals for the District of Columbia Circuit has interpreted the analogous federal FOIA exemption to require a ‘‘pending or reasonably anticipated’’ law enforcement action, federal courts hearing FOIA claims have used inconsistent language in their efforts to describe and apply that standard. At times, those courts have applied what is arguably a less stringent standard than we articulate here. See, e.g., Citizens for Responsi- bility & Ethics in Washington v. United States Dept. of Justice, 746 F.3d 1082, 1098 (D.C. Cir. 2014) (stating that ‘‘an ongoing criminal investigation typically triggers [the exemption]’’); Juarez v. Dept. of Justice, 518 F.3d 54, 59 (D.C. Cir. 2008) (concluding that, ‘‘so long as the investigation continues to gather evidence for a possible future criminal case, and that case would be jeopardized by the premature release of that evidence, [the exemption] applies’’). Moreover, whereas Connecticut courts decide FOIA exemption claims based on the facts of each case, federal courts may apply a categorical approach that omits consideration of individual circumstances. Compare Director, Retirement & Benefits Services Division, Office of the Comptroller v. Freedom of Information Commission, 256 Conn. 764, 779, 775 A.2d 981 (2001), with United States Dept. of Justice v. Reporters Committee for Freedom of the Press, 489 U.S. 749, 776, 109 S. Ct. 1468, 103 L. Ed. 2d 774 (1989). For these reasons, federal cases interpreting this FOIA exemption may not be as instructive as they initially appear. 7 As we discuss hereinafter, although the respondents rely heavily on Graeber, in which the reasonable possibility standard was deemed satisfied, the commission reasonably could find that that case is distinguishable from the present case because (1) the respondents in Graeber submitted all responsive records for the commission’s in camera review in order to provide specific, nonspeculative support for their assertions, and (2) the investiga- tion in Graeber remained in a different, more active phase, insofar as the assistant state’s attorney leading the investigation spent at least one day per week on the case, she was actively supervising a team of detectives in that regard, and they continued to interview witnesses to the crime. See Graeber v. Chief, New Haven Police Dept., supra, Docket No. FIC 2016-0865. 8 See also 18 S. Proc., Pt. 5, 1975 Sess., pp. 2323, 2325, remarks of Senator Robert L. Julianelle; id., p. 2327, remarks of Senator Lawrence J. DeNardis; 18 H.R. Proc., supra, p. 3910, remarks of Representative Martin B. Burke; Conn. Joint Standing Committee Hearings, Government Administration and Policy, 1975 Sess., pp. 304–305, remarks of Representative Walter J. Hen- derson. 9 We note that such arguments are better addressed to the legislature, and we continue to encourage the legislature to provide as much clarity and specificity as possible as to how it would strike the balance between the competing interests at play in cases such as this. 10 Our legislature expressed this sentiment even more clearly when amend- ing the sister law enforcement exception, § 1-215, in 2015. See, e.g., 58 H.R. Proc., Pt. 18, 2015 Sess., pp. 5974–75, remarks of Representative Richard A. Smith (commenting that some police departments are less than forthcom- ing and that ‘‘the police cannot hide behind closed doors’’); Conn. Joint Standing Committee Hearings, Government Administration and Elections, Pt. 1, 2015 Sess., pp. 212–13, remarks of Representative Smith (opining that police, left unsupervised, can ‘‘become abusive,’’ that police departments have ‘‘all made mistakes,’’ and that stronger transparency laws ‘‘have the advantage of making sure that the police now know that they’re being watched and they need to do a better job at what they’re doing’’); Conn. Joint Standing Committee Hearings, supra, pp. 214–15, remarks of Representative David Alexander (referencing ‘‘a couple [of] bad actors’’ who were complicit in police brutality and raising concerns about hidden arrests and overly oppressive government). Those amendments are directly relevant to our understanding of § 1-210 because the primary purpose of the amendments was to make clear that, during the pendency of an arrest, law enforcement officials must comply with the broad disclosure requirements of § 1-210 and not only with § 1-215. See Public Acts 2015, No. 15-164, § 1. 11 See, e.g., Conn. Joint Standing Committee Hearings, Government Admin- istration and Elections, Pt. 1, 2015 Sess., p. 208, remarks of Colleen M. Murphy, executive director and general counsel of the Freedom of Informa- tion Commission (‘‘[B]elieve me . . . if a member of the law enforcement community comes in and makes that representation to us, we don’t really [second-guess] it . . . as long as there is . . . any kind of basis for it. . . . [B]ut that law enforcement [agency] must at least come in and say what those reasons are and strike that balance between disclosure and confidentiality.’’). 12 Other crimes without statutes of limitations include all class A felonies, various sexual crimes committed against minors, and certain instances of escape, hindering prosecution, perjury, and motor vehicle violations resulting in the death of another person. See General Statutes § 54-193 (a); see also General Statutes § 54-193 (b) (various crimes featuring statutes of limitations of ten to thirty years). 13 The respondents raised this issue in their trial brief. The trial court did not directly address the claim. 14 We are not persuaded that the commission merely intended to distin- guish the present case, in which no suspect has been publicly named and identified, from prior cases, in which a specific suspect had been arrested and was thus known to the public. The commission’s final decision stated that the respondents failed to identify a suspect in the context of distinguish- ing the present case from two previous decisions. In neither of those deci- sions, however, had the prime suspect been arrested or publicly identified. See Strauss v. Chief, Westport Police Dept., Freedom of Information Com- mission, Docket No. FIC 2010-487 (May 25, 2011); Rouen v. Chief, Groton Police Dept., supra, Docket No. FIC 2006-064. 15 To maintain the integrity of their investigation and techniques, law enforcement agencies often make the required showing by submitting for the commission’s in camera review any documents that allegedly fall under the law enforcement exception or otherwise support their claims. See Regs., Conn. State Agencies § 1-21j-37 (f) (1) (‘‘[a]ny party or intervenor may request an in camera inspection of the records claimed to be exempt from disclosure in a contested case . . . and the presiding officer or the commission may order such an inspection on request, on such presiding officer’s or the commission’s own initiative, or on remand by a court’’); see also Hartford v. Freedom of Information Commission, supra, 201 Conn. 434 n.16 (commis- sion may review records, affidavits, or testimony in camera to allow respon- dents to make sufficiently detailed showing without compromising claimed confidentiality). Indeed, in most of the cases on which the respondents rely, the law enforcement agency submitted the documents at issue to the commission for in camera review. See, e.g., Graeber v. Chief, New Haven Police Dept., supra, Docket No. FIC 2016-0865 (finding that prospective law enforcement action was reasonable possibility when respondents submitted copies of all responsive records for in camera review); Hoda v. Chief, New Haven Police Dept., supra, Docket No. FIC 2007-143 (following in camera review of investigation records, finding that law enforcement exception was established); Rouen v. Chief, Groton Police Dept., supra, Docket No. FIC 2006-064 (same); Cotton v. Chief, Meriden Police Dept., Freedom of Informa- tion Commission, Docket No. FIC 2006-020 (August 9, 2006) (finding that first prong of law enforcement exception was satisfied on basis of in camera review). Although this avenue is not required, any party or intervenor may request an in camera review, and the hearing officer may order such an inspection on his or her own initiative. In the present case, the respondents did not submit any records for in camera review, and the commission did not order them to do so. 16 Notably, the commission’s August 26, 2020 final decision relies on Dept. of Public Safety v. Freedom of Information Commission, supra, 51 Conn. App. 105, which we repudiated in part II B 2 of this opinion, for the applicable legal standard. 17 It is unclear, for example, whether the statistical data support the same conclusions for murder investigations in which the police have identified a suspect. It will be for the commission, on remand, to determine what bearing, if any, broad statistical trends have on the likelihood of success in any particular case.