Tardiff v. Knox County

365 F.3d 1 (2004) | Cited 39 times | First Circuit | April 9, 2004

Before us are orders certifying class actions in two jailhouse strip search cases. The two cases, arising before two different judges in the same court (Judges Carter and Hornby), are targeted at two different Maine counties (Knox and York). In each case, a named plaintiff seeks to represent others who (it is alleged) were improperly strip searched. Following certification of the classes, we allowed interlocutory appeals and now affirm.

Over the last few decades, a changed popular sensibility has produced a series of decisions curtailing what was once an apparently routine practice in many jails of strip searching arrestees not yet convicted of any crime. 1 Such cases have included, or brought in their wake, lawsuits by those who were searched seeking damages from officials or governmental entities responsible for the searches. E.g., Miller v. Kennebec County, 219 F.3d 8, 12-13 (1st Cir. 2000). These suits are ordinarily framed as Fourth Amendment claims for unconstitutional search and seizure and brought under 42 U.S.C. § 1983 (2000). Id.

Qualified immunity has defeated some of these claims against officials, e.g., Savard v. Rhode Island, 338 F.3d 23, 33 (1st Cir. 2003) (en banc), cert. denied, 124 S. Ct. 1074 (2004), but the doctrine does not protect counties, see Owen v. City of Independence, 445 U.S. 622, 650 (1980). Though not liable under a respondeat superior theory, Bd. of the County Comm'rs v. Brown, 520 U.S. 397, 403 (1997), counties and like entities may be liable under section 1983 not only for their formal official acts and policies, but also, under some circumstances, for practices "so permanent and well settled" as to constitute established custom. Monell v. Dep't of Social Servs., 436 U.S. 658, 691 (1978) (quoting Adickes v. S.H. Kress & Co., 398 U.S. 144, 167-68 (1970)). In the first case before us--brought against Knox County, its sheriff, and unidentified officers--Laurie Tardiff filed such a section 1983 claim, asserting that she had been arrested for witness tampering after having been previously charged with violating an anti-harassment order. Tardiff v. Knox County, 218 F.R.D. 332, 334 (D. Me. 2003). At the jail house, Tardiff alleged that she was taken to a shower area, ordered by a female corrections officer to disrobe, and required to "squat and cough" while the officer could see her vagina and anal area. Id.

Judge Carter, presiding in the case, called this a "strip and visual body cavity search," Tardiff, 218 F.R.D. at 334; Judge Hornby, faced with a similar description by the lead plaintiff in a parallel suit against York County, called the same thing a "strip search," Nilsen v. York County, 219 F.R.D. 19, 22, 25 (D. Me. 2003). For simplicity's sake, we use "strip search" in this decision to cover the described conduct, recognizing that the phrase is used variously in different decisions and that there is a spectrum of possible search practices inflicting differing indignities. Just what happened to Tardiff and others may itself be disputed.

Tardiff sought to represent a class of persons similarly situated and in due course Judge Carter certified a class under Fed. R. Civ. P. 23(b)(3). Tardiff, 218 F.R.D. at 336. The class, as defined by the court, covered:

All people who after November 19, 1996, were subjected to a strip search and/or visual body cavity search without evaluation for individualized reasonable suspicion while being held at the Knox County Jail:

(1) after having been arrested on charges that did not involve a weapon, drugs, or a violent felony; or

(2) while waiting for bail to be set on charges that did not involve a weapon, drugs, or a violent felony; or

(3) while waiting for an initial court appearance on charges that did not involve a weapon, drugs, or a violent felony; or

(4) after having been arrested on a warrant that did not involve a weapon, drugs, or a violent felony.


In the second case, brought before Judge Hornby, a substitute lead plaintiff--Michael Goodrich--was arrested for failing to report for probation and (he alleges) was strip searched with the same strip, squat and cough procedure as Tardiff. Nilsen 219 F.R.D. at 22. Like Tardiff, he alleged that this was done in accordance with a policy or custom of the county that was applied widely to arrestees including those charged with unthreatening offenses. Judge Hornby, acting shortly after Judge Carter, entered a comparable order certifying a class under Rule 23(b)(3). Id. at 19-20. 2

In each case, the defendants sought interlocutory review of the class certification. We have discretion as to whether to entertain such an appeal. See Fed. R. Civ. P. 23(f) and 1998 advisory committee's note to subdivision (f). One reason for review is a threat of liability so large as to place on the defendant an "irresistible pressure to settle." Waste Mgmt. Holdings, Inc. v. Mowbray, 208 F.3d 288, 293 (1st Cir. 2000). Although we thought it quite possible that the certification orders would survive review, the financial and similar information provided by the two counties in this case persuaded us to grant interlocutory review, which we expedited.

Nominally, review of decisions granting or denying class certification is for "abuse of discretion," Smilow v. S.W. Bell Mobile Sys., Inc., 323 F.3d 32, 37 (1st Cir. 2003), but this chameleon phrase is misleading. Express standards for certification are contained in Rule 23, so an appeal can pose pure issues of law reviewed de novo or occasionally raw fact findings that are rarely disturbed. See id. Here, we are concerned with applying general standards to particular facts where the touchstone is "reasonableness" and review is deferential. 3

All class actions certified under Rule 23 must meet certain prerequisites listed in subsection (a): there must be numerosity of class members, common questions of law or fact, the representative must be typical of the class, and his or her representation of the class must be adequate. Fed. R. Civ. P. 23(a). Here, these preconditions are admittedly satisfied. The defendants' attack is instead on the further requirement, for a (b)(3) class, that the court find that common questions "predominate" over individual ones and that the class action be "superior" to other methods of resolving the controversy. Fed. R. Civ. P. 23(b)(3).

Subsection (b)(3), the cute tiger cub that has grown into something unexpectedly fearsome in civil rights and mass tort litigation, is a joinder device for consolidating separate but similar claims--as opposed, for example, to disputes about a common fund, which is covered by subsection (b)(1)(B). The subsection lists non-exclusive factors for making the determination--the most pertinent here is manageability--but the "predominance" and "superiority" labels turn largely on the particular facts and issues presented. Some courts have allowed strip search class actions and others not. 4

What here are the issues? The principal ones--and this is our own very rough cut--appear to be these: what rule, policy or custom as to strip searches was in force in each county during the periods in question; was it lawful as applied to groups or individual class members; if unlawful as to some categories of arrestees, were there still facts as to particular arrestees that justified a strip search; and if liability exists, what are the damages to each person injured? There are obviously other issues (e.g., as to individual defendants); but these four are likely the most important.

On the first issue--what rules, policies or customs prevailed--there is already deep disagreement between plaintiffs and defendants. So the question arises whose version is a court to assume for purposes of deciding whether a class action is even suitable? It is sometimes taken for granted that the complaint's allegations are necessarily controlling; but class action machinery is expensive and in our view a court has the power to test disputed premises early on if and when the class action would be proper on one premise but not another. 5

In this case, whether there was a rule, policy or custom of automatically strip searching all or most categories of arrestees is on its face a common disputed issue--as both sides concede. Nothing here obliged the district courts to do more than view the issue as such and, so viewed, it weighs in favor of class action status. The issue is like any other common predicate to liability for multiple parallel claims--say whether the train engineer was negligent (or acting within the scope of his employment) when the train went off the rails. See, e.g., Am. Trading & Prod. Corp. v. Fischbach & Moore, Inc., 47 F.R.D. 155, 157 (N.D. Ill. 1969).

The second issue, whether the strip search policy is unlawful, presents different complications. The first is that it might be lawful as to some groups of arrestees (say, those arrested on firearm felony charges) and not others (say, those seized for non-payment of parking tickets). Both district courts sought to narrow the potential multiplicity of categories--which could in several respects work against class status--by narrowing the class. In both orders, the class definition excluded those arrested for drugs, weapons or violent felonies--categories for which automatic strip searches might seem easier to defend. Nilsen, 219 F.R.D. at 25; Tardiff, 218 F.R.D. at 336.

Of course, the defendants might still persuade the court that some within the remaining group of arrestees (e.g., those arrested for misdemeanors of certain types) could also be strip searched as a matter of course. But the core notion that some classes of arrestees should not be strip searched without particularized suspicion is now embedded in the case law. 6 Just what groups within the class can or cannot be searched without such suspicion appears, like the rule-policy-custom question, to present a common set of issues proper for a class action.

Turning to the third of the issues we listed, defendants say that individualized suspicion may have existed as to anyone within the certified class, so that every class member necessarily presents a different problem even as to liability. True, a class member arrested for the most minor and non-violent of offenses might arguably warrant a strip search by his individual behavior (violently resisting arrest) or prior record (armed assaults). If such an individual evaluation were necessary in every case, this might doom an efficient class action.

Both district courts sought to finesse the issue by defining the class to include only those who were searched without evaluation for individualized suspicion. Nilsen, 219 F.R.D. at 25 ("without evaluating for individualized reasonable suspicion"); Tardiff, 218 F.R.D. at 336 ("without evaluation for individualized reasonable suspicion"). This is not a complete answer; for one thing, at some point in the case it could still be necessary to determine individually who among those who were searched were so evaluated.

Further, even if all persons searched based on individual evaluations were readily excluded from the class regardless of the soundness of the evaluation, it might still be necessary--at least in some cases--to consider whether individualized suspicion could have justified a search. Even if overbroad categories were used, the defendants will likely argue that the result would have been the same for some class members even if proper individualized judgments had been made--raising a legal issue as to liability on which we offer no opinion. Compare 42 U.S.C. § 2000e-5(g)(2)(B) (2000) (Title VII's same decision defense as to damages). If the number of such instances was large, the feasibility of a class action might be compromised.

Without prejudging disputable issues, we think that such threats of undue complications as to liability are limited in this case. If there was in fact a rule, custom or policy of strip searching every arrestee or a substantially overlarge category, then it is a fair guess that most arrestees so classed were strip searched on this basis. There might yet be some number as to whom defensible individual judgments to strip search were actually made or could have been made--two different situations with different legal implications; but whoever has the burden of identifying such persons, they may well not be numerous. 7

Determinations to certify a class depend on initial prediction and are always subject to revision. Gen. Tel. Co. of the S.W. v. Falcon, 457 U.S. 147, 160 (1982); Smilow, 323 F.3d at 39-40; Mowbray, 208 F.3d at 297 n.6. If a large number of class members turn out to present non-common issues as to liability, the court may have to consider narrowing or de-certifying the class. Mowbray, 208 F.3d at 297 n.2. But a court can proceed based on its reasonable best guess as to what will happen and we share the district judges' apparent view that the individualized-suspicion issue will arise only in a limited number of cases, assuming always that plaintiffs can establish as a background fact the existence of an improperly broad strip search rule, policy or custom.

By contrast, the presence of damage claims does weigh against class status. In certain cases (e.g., some security frauds), formula damages are generally used. See 3 Conte & Newberg, Newberg on Class Actions § 10:8 (4th ed. 2002). Here each class member wrongly strip searched could in principle testify separately as to the circumstances, emotional damage, lost wages, medical treatment, doctor bills, and so on. Indeed, the less potent the available defense on liability, the more the class action might look like a bundle of individualized damage claims wholly unsuited for class resolution.

Yet the need for individualized damage decisions does not ordinarily defeat predominance where there are still disputed common issues as to liability. Smilow, 323 F.3d at 40. If the class action resolved liability even as to some further narrowed class, this would be a legitimate function. At that point, the court would have further options, such as an agreement on modest uniform damages for those not claiming special injury, with masters to determine the (potentially few) serious claims to special injury. E.g., In re Corrugated Container Antitrust Litig., 80 F.R.D. 244 (S.D. Tex. 1978).

Failing some practical solution allowing full resolution of all class damage claims in a single case, the court could enter a judgment of liability, leaving class members to pursue damage claims in separate law suits. Smilow, 323 F.3d at 40-41; accord. Blihovde v. St. Croix County, 219 F.R.D. 607, 621 (W.D. Wis. 2003); Maneely v. City of Newburgh, 208 F.R.D. 69, 79 (S.D.N.Y. 2002) (strip search cases). In addition, if and when liability is established and the remaining dispute is only the amount of damages, it is common experience that a great many claims settle.

It remains to discuss the superiority vel non of the class action. The counties insist that individual law suits are feasible. We think that only the limited number of cases where serious damage ensued would ever be brought without class status and that the vast majority of claims would never be brought unless aggregated because provable actual damages are too small. This is a conventional argument for a class action, Amchem Prods., Inc., v. Windsor, 521 U.S. 591, 617 (1997), and it applies here.

Whether the law should encourage the bringing of very small claims that would otherwise not be brought is a different matter; law suits serve purposes beyond compensation, and the balance of cost and benefit doubtless varies from case to case. It is enough for the superiority determination here that for most strip search claimants, class status here is not only the superior means, but probably the only feasible one (one-way collateral estoppel aside), to establish liability and perhaps damages. E.g., Blihovde, 219 F.R.D. at 622; Mack v. Suffolk County, 191 F.R.D. 16, 25 (D. Mass. 2000).

It follows that the orders certifying the class actions must be sustained at this initial stage without prejudice to their further alteration. In further proceedings in the district court, little weight should be given to our own glancing discussion of the merits; although necessary background in reviewing the certification orders, our assessment can only be preliminary, lacking the benefit of a fuller record and analyses yet to come.


1. The trend began with the Supreme Court's decision in Bell v. Wolfish, 441 U.S. 520 (1979). In this circuit, the leading cases are Wood v. Hancock County Sheriff's Dep't, 354 F.3d 57 (1st Cir. 2003); Miller v. Kennebec County, 219 F.3d 8 (1st Cir. 2000); Swain v. Spinney, 117 F.3d 1 (1st Cir. 1997).

2. Judge Hornby's order, Nilsen, 219 F.R.D. at 25, defines the class as: All people strip-searched at the York County Jail after October 14, 1996 under a policy or custom of conducting strip-searches without evaluating for individualized reasonable suspicion: (1) while waiting for bail to be set or for a first court appearance after being arrested on charges that did not involve a weapon or drugs or a violent felony; or (2) while waiting for a first court appearance after being arrested on a default or other warrant.

3. Cf. Stanton v. Berkshire Reg'l Sch. Dist., 197 F.3d 574, 577 (1st Cir. 1999) (attorneys' fees under IDEA); Mass. Food Ass'n v. Mass. Alcoholic Beverages Control Comm'n, 197 F.3d 560, 567 (1st Cir. 1999) (adequacy of representation under Fed. R. Civ. P. 24(a)).

4. Compare Blihovde v. St. Croix County, 219 F.R.D. 607, 622 (W.D. Wis. 2003), Bynum v. District of Columbia, 217 F.R.D. 43, 50 (D.D.C. 2003), Maneely v. City of Newburgh, 208 F.R.D. 69, 76-79 (S.D.N.Y. 2002), Mack v. Suffolk County, 191 F.R.D. 16, 25 (D. Mass. 2000), and Smith v. Montgomery County, 573 F. Supp. 604, 613 (D. Md. 1983)(all certifying strip search class actions), with Klein v. DuPage County, 119 F.R.D. 29, 31-32 (N.D. Ill. 1988)(denying certification for strip search class action). There are also several unpublished opinions coming out both ways.

5. Compare Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 177 (1974)(discouraging any preliminary inquiry), with Gen. Tel. Co. of S.W. v. Falcon, 457 U.S. 147, 160-161 (1982)(endorsing such an inquiry). The circuits are also divided. Compare Caridad v. Metro-North Commuter R.R., 191 F.3d 283, 291-93 (2d Cir. 1999), and J.B. ex rel. Hart v. Valdez, 186 F.3d 1280, 1290 n.7 (10th Cir. 1999), with Johnston v. HBO Film Mgmt., Inc., 265 F.3d 178, 186-89 (3d Cir. 2001), and Szabo v. Bridgeport Machs., Inc., 249 F.3d 672, 675-76 (7th Cir.), cert. denied, 534 U.S. 951 (2001).

6. See Miller, 219 F.3d at 12; Swain, 117 F.3d at 12. If anything, some of the rhetoric might lead the reader to believe, quite wrongly, that every arrestee must be independently evaluated based on his or her own circumstances. In reality, generalizations lie behind most so-called individual decisions and to endorse the generalization is, in substance, to endorse some categorical rules. See Roberts v. Rhode Island, 239 F.3d 107, 112 (1st Cir. 2001).

7. A number of district courts say that the burden of identification is on the defendant, e.g., Blihovde, 219 F.R.D. at 622; Mack, 191 F.R.D. at 24; see also Int'l Bhd. of Teamsters v. United States, 431 U.S. 324, 360 (1977)(Title VII), but the various burden issues have not been briefed in this case, and we do not resolve them.

Back to top