Government Accountability Project v. United States Department of Homeland Security

Opinion

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA

GOVERNMENT ACCOUNTABILITY PROJECT,

Plaintiff, Case No. 1:17-cv-2518 (CRC)

v.

U.S. DEPARTMENT OF HOMELAND SECURITY,

Defendant.

MEMORANDUM OPINION

This Freedom of Information Act case is reminiscent of the classic board game Battleship, where players array a fleet of plastic warships on a secret grid and alternate directing “shots” at the opponent’s vessels by calling out precise coordinates. A shot hits its mark only if an enemy vessel is situated on a specified target.

Plaintiff Government Accountability Project (“GAP”) asked for any records in the Department of Homeland Security’s (“DHS”) possession that related to “ideological tests” and “searches of cellphones” at the U.S. border. The agency obliged in a manner consistent with the rules of Battleship. It canvassed its electronic records for direct hits, looking only for records that contained the verbatim language GAP used in its request. For the first search, it used the terms “ideological tests” and “border”; for the second, it used the keywords “search” and “cellphone.” After the searches yielded zero responsive documents, GAP complained that DHS unreasonably omitted additional search terms that quite likely would have generated a more robust return. Because FOIA requests do not operate like a game of Battleship—and for other

2

more technical reasons that follow—the Court agrees and will order the agency to conduct its search anew. I. Background On April 4, 2017, GAP, a non-profit advocacy organization, submitted a FOIA request to DHS. Complaint ¶ 6. GAP requested three species of documents. It sought (1) “correspondence between White House staff and the DHS concerning ideological tests at the U.S. Border”; (2) “correspondence concerning searches of cellphones, the protocols, information about who was searched . . . , search rates, protocols if a search is refused, etc., for citizens and non-citizens, at the U.S. border”; and (3) “any records generated in connection with topics listed above that raised or were responding to compliance of 5 U.S.C. § 2303(b)(8),” a federal whistleblower- protection statute. Id.

Some five months passed with no response from DHS, so GAP filed suit in November 2017 to force the agency’s hand. Id. ¶ 8. Ten days later, on December 1, the agency informed GAP that its search had located no responsive records. Def’s MSJ, Ex. 2, Declaration of James V.M.L. Holzer (“Holzer Decl.”) ¶ 11. But “[o]ut of an abundance of caution,” the agency “conducted an additional broader search,” this time soliciting the assistance of DHS’s Office of the Chief Information Officer (“OCIO”). Id. ¶ 12. It sent “two search taskers” to OCIO. The first asked OCIO to search for “any and all emails” between DHS.GOV and EOP.GOV1 email addresses that used the words “ideological test” and “border.” Id. ¶ 13. The second requested “any and all emails” between the same email addresses using the words “search” and “cellphone.” Id. ¶ 14.

1 DHS believed that “EOP”—which stands for Executive Office of the President—was the appropriate email suffix for White House staffers.

3

OCIO completed its search in January 2018. It informed DHS that the “first tasker yielded zero records” and the “second tasker returned 965 megabytes of records,” which amounted to 807 documents. Id. ¶ 16. DHS then “conducted a document-by-document key word search of all 807 documents” and determined that none of them were responsive to GAP’s FOIA request. Id. ¶ 17.

GAP now contends the agency’s searches for communications regarding ideological tests and cellphone searches were inadequate, that a third search needed to be done for the part of its request seeking whistleblower records, and that at the very least, the Court should conduct its own review of some portion of the 807 documents to test the agency’s claim that the documents the search did turn up were ultimately non-responsive. The parties have briefed these issues on summary judgment, and the matter is now ripe for the Court’s resolution. II. Legal Standard FOIA cases are typically resolved on summary judgment. See Brayton v. Office of U.S.

Trade Rep., 641 F.3d 521, 527 (D.C. Cir. 2011). Summary judgment is appropriately granted if “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(a).

When the adequacy of an agency’s search is questioned, the agency must show “beyond material doubt that its search was reasonably calculated to uncover all relevant documents.” Ancient Coin Collectors Guild v. U.S. Dep’t of State, 641 F.3d 504, 514 (D.C. Cir. 2011) (quoting Valencia-Lucena v. U.S. Coast Guard, 180 F.3d 321, 325 (D.C. Cir. 1999)) (internal quotation omitted). The key question is whether the search itself was reasonable, regardless of the results. See Cunningham v. U.S. Dep’t of Justice, 40 F. Supp. 3d 71, 83-84 (D.D.C. 2014). “Therefore, the adequacy of a FOIA search is generally determined not by the fruits of the

4

search, but by the appropriateness of the methods used to carry out the search.” Francis v. U.S. Dep’t of Justice, 267 F. Supp. 3d 9, 12 (D.D.C. 2017) (alteration and quotation omitted). “Although a requester must reasonably describe the records sought, an agency also has a duty to construe a FOIA request liberally.” Nation Magazine, Washington Bureau v. U.S. Customs Serv., 71 F.3d 885, 890 (D.C. Cir. 1995) (alteration, citation, and quotation omitted). In the end, while agencies need not “scour every database,” Francis, 267 F. Supp. 3d at 12, or “turn over every stone,” they “must conduct a ‘good faith, reasonable search of those systems of records likely to possess requested records,’” Freedom Watch, Inc. v. Nat’l Sec. Agency, 220 F. Supp. 3d 40, 44 (D.D.C. 2016) (quoting Cunningham, 40 F. Supp. 3d at 83).

“An agency may prove the reasonableness of its search through a declaration by a responsible agency official[.]” Id. “Agency declarations, especially from individuals coordinating the search, are afforded ‘a presumption of good faith, which cannot be rebutted by purely speculative claims about the existence and discoverability of other documents.’” Freedom Watch, 220 F. Supp. 3d at 44 (quoting SafeCard Servs., Inc. v. SEC, 926 F.2d 1197, 1200 (D.C. Cir. 1991)). Indeed, courts can award summary judgment solely based on agency affidavits and declarations that are “relatively detailed and non-conclusory.” Id. III. Analysis GAP initially identified three bases why it is entitled to summary judgment, but in response to subsequent search efforts by DHS, now emphasizes just one.2 It maintains that the

2 In its opening brief, GAP also complained that the agency failed entirely to conduct any search for GAP’s third request seeking whistleblower records, and asked the Court to review in camera some sample of the documents that the allegedly inadequate search nevertheless produced, to evaluate whether they truly were non-responsive. Pl’s MSJ at 3-5. But in light of the additional information provided by DHS in a supplemental declaration, GAP dropped its request for in camera inspection of the records and refined its argument with respect to the agency’s search for whistleblower records. Pl’s Reply at 6. For reasons explained in footnote 4

5

agency’s search pursuant to GAP’s first two requests—for information on ideological tests and cellphone searches at the border—was inadequate. GAP contends that the search was lacking in three respects: (1) it relied on too few keywords and excluded obvious synonyms; (2) it failed to consult experts to identify search terms likely to produce responsive records; and (3) it canvassed only official government email accounts rather than other forms of electronic communications. Pl’s Mot. for Summ. J. (“Pl’s MSJ”) at 2-4. The Court agrees with GAP on the first score and stops its analysis there, since that is sufficient grounds to grant summary judgment in its favor and order DHS to search its records a second time.

GAP complains that the agency failed to include “logical variations of the terms used to describe the subject matter being sought.” Pl’s MSJ at 2. Instead, the agency searched its records for the exact words in the FOIA request—and nothing else. GAP’s first request asked for correspondence about “ideological tests at the border,” Compl. ¶ 6, so the agency’s search filtered for “ideological tests” and “border,” Holzer Decl. ¶ 13. Words that might serve as proxies for ideology (like “politics” or “values”) were omitted from the search, as were synonyms for test (like “evaluation,” “screening,” or “vetting”). See Pl’s MSJ at 3. Similarly, when GAP asked for correspondence “concerning searches of cellphones,” Compl. ¶ 6, the agency’s search filtered for “cellphone” and “search,” Holzer Decl. ¶ 14, but did not include variant spellings like “cell phone” or “phone,” Pl’s MSJ at 3. GAP contends that the agency’s cramped search is akin to a game of Battleship: fruitful only if the literal request results in a direct hit, but not calculated to turn up all responsive documents.

below, it is not necessary for the Court to reach what remains of GAP’s whistleblower complaint at this time.

6

The Court agrees. Again, though a “requester must reasonably describe the records sought, an agency also has a duty to construe a FOIA request liberally.” Nation Magazine, 71 F.3d at 890. And ultimately, it is the agency’s burden to show “beyond material doubt that its search was reasonably calculated to uncover all relevant documents.” Ancient Coin Collectors Guild, 641 F.3d at 514 (quotation omitted). Two cases in this district confirm that DHS failed to meet this burden.

The first, Summers v. Department of Justice, speaks to the agency’s search for correspondence about “ideological tests.” In Summers, the FOIA requester asked for “any and all commitment calendars” of former FBI Director J. Edgar Hoover; the FBI searched only for the word “commitment.” 934 F. Supp. 458, 461 (D.D.C. 1996) (alteration omitted). Though the search was “fruitful,” turning up some “daily logs and other appointment calendars,” the court agreed with Summers that the search should have included “two of the most obvious search terms” (“appointment” and “diary”) to better ensure that “all documents responsive to the FBI’s request would be discovered.” Id. While “ideological tests” may not have quite as obvious substitutes as “commitment calendars,” searching for that phrase verbatim was always doomed to return limited results. At the very least, the agency should have used synonyms for “test” and proxies for “ideological” and should not have grouped the two words together—since that risks excluding any email where both “ideological” and “tests” were used, but not in the precise “ideological tests” syntax.

DHS tries to distinguish Summers. It contends that the court ordered an expansion of the search terms only because the FBI’s initial search suggested the additional terms “appointment” and “diary” would have yielded still more returns. Def’s Reply and Opposition at 6. But that is not how this Court reads Summers. The key fact in Summers was that the FBI “failed to use two

7

of the most obvious search terms . . . to ensure that all documents responsive to the FBI’s request would be discovered.” 934 F. Supp. at 461. That the FBI’s initial search revealed particular synonyms that the court said must be included in the subsequent search doesn’t mean that its holding applies only to cases where the first search was at least somewhat helpful. If that were right, the Summers rule would apply only to the class of cases where it is needed the least. Consider Summers itself: if the court had ruled the other way—and said the agency need not redouble its search efforts using additional terms—the requester could have simply filed a new FOIA request asking specifically for diaries and appointments, terms Summers knew would produce responsive documents given the fruits of the first search. But according to DHS’s argument here, if the initial search using only the request’s verbatim language turned up nothing, not only would Summers offer no solace, but the requester has gained no new insights from the initial search that could help him craft a second, more effective FOIA request. The Court will not adopt such a counterintuitive view of Summers; its holding applies here, and it mandates that the agency include logical variations of “ideological tests.”

A second case, Bagwell v. Department of Justice, goes to the adequacy of the agency’s search for correspondence regarding cellphone searches at the border. In Bagwell, the FOIA requester sought documents relating to the investigation of the child sex abuse scandal at Pennsylvania State University. 311 F. Supp. 3d 223 (D.D.C. 2018). Responding to the request, the Executive Office for United States Attorneys searched for documents referencing “Pennsylvania State University”—which tracked the exact wording of the request—but did not include other “names commonly used to refer to the University” like “PSU” or “Penn State.” Id. at 229. This very Court easily rejected the search as insufficient: “Because it is likely that emails concerning the investigation would use ‘PSU’ or ‘Penn State’ rather than the full name of the

8

University, the Department’s search was not reasonably calculated to find all responsive emails.” Id. at 230. That logic applies likewise to DHS’s search for correspondence about cellphone searches. Searching only for the word “cellphone” is inadequate; variants that may well be used in correspondence—like the two-word version “cell phone” or simply “phone”—must also be included. Such variants may be just as likely to be used by the relevant government officials. For proof, look no further than the Supreme Court’s most famous case to date on cellphones—or, rather, cell phones—Riley v. California, 134 S. Ct. 2473 (2014). A ctrl+F word-search for “cellphone” in that case returns zero hits; the same search for “cell phone” returns 117.

FOIA requests are not a game of Battleship. The requester should not have to score a direct hit on the records sought based on the precise phrasing of his request. Rather, the agency must liberally interpret the request and frame its search accordingly. Guided by Summers and Bagwell, the Court concludes that DHS neither liberally construed GAP’s request, Nation Magazine, 71 F.3d at 890, nor conducted a search reasonably likely to produce all responsive records, Ancient Coin Collectors Guild, 641 F.3d at 514. Therefore, it will order DHS to conduct an additional search of its records. Before doing so, the Court will order the parties to meet and confer within fourteen days and engage in a good faith effort to arrive at a reasonably limited set of additional search terms that rectify the under-inclusivity of the “ideological tests” and “cellphone” search terms without being too over-inclusive.3 If the parties cannot reach

3 The Court appreciates DHS’s concern that searches for commonly-used words like “cell,” “phone,” and “test” may return too many records for the agency to digest. See Def’s Reply and Opposition Ex. 2, Supplemental Declaration of James V.M.L. Holzer (“Holzer Supp. Decl.”) ¶ 12. But that concern dictates using more sophisticated search techniques, including additional filtering keywords or Boolean operators and connectors, to winnow the results to a manageable level. See Bagwell, 311 F. Supp. 3d at 230 n.3; see also Nat’l Day Laborer Org. Network v. U.S. Immigration & Customs Enf’t Agency, 877 F. Supp. 2d 87, 106-07 (S.D.N.Y. 2012) (giving several examples of how slight changes to search terminology and use of Boolean

9

agreement, they shall file a joint submission setting forth their respective terms and their positions on all disputed terms within seven days thereafter.4 IV. CONCLUSION For the foregoing reasons, the Court will grant plaintiff’s motion for summary judgment with respect to the adequacy of the search and deny defendant’s motion for summary judgment. A separate Order shall accompany this memorandum opinion.

________________________

CHRISTOPHER R. COOPER

United States District Judge

Date: October 12, 2018

connectors yields dramatically different results). It does not justify foregoing a reasonable search.

4 GAP also complains that the agency’s search for whistleblower records was unreasonably confined to official government email accounts. Pl’s Reply at 6. Because the Court is ordering a further search of those email accounts, it will defer resolution of this point to enable the parties to consider the issue in their discussions concerning new search parameters.