Darryl N. Veazey, Plaintiff-appellant, v. Communications & Cable of Chicago, Inc., D/b/a Tci Communications, Inc., Chicago Cable Tv, Tci Chicago or Tci-chicago Cable,* Defendant-appellee, 194 F.3d 850 (7th Cir. 1999)

Darryl Veazey contends that his former employer, LaSalle Telecommunications, Inc., incorrectly sued as Communications & Cable Co. of Chicago, violated the Employee Polygraph Protection Act ("EPPA"), 29 U.S.C. §sec. 2001-09, when it discharged him because he refused to provide the specific tape-recorded voice exemplar¹  his superiors had requested. The district court dismissed his suit under Federal Rule of Civil Procedure 12(b) (6) for failure to state a claim. Because it is possible to hypothesize a set of facts consistent with Veazey's complaint that would entitle him to relief, we reverse the district court's ruling and remand for further proceedings.

In the fall of 1996, Darryl Veazey's employer, LaSalle Telecommunications, Inc., suspected that Veazey, who was employed as an outage coordinator/dispatcher, had left a hostile and threatening anonymous message on the voicemail of another employee at LaSalle.²  Accordingly, LaSalle set up an interview with Veazey concerning the incident. Mike Mason, LaSalle's Customer Fulfillment Manager, and Jack Burke, a "cable troubleshooter," questioned Veazey about the message during a four hour interview. Veazey maintained his innocence at all times.

Despite Veazey's denials, Mason and Burke requested that Veazey read a verbatim transcript of the threatening message into a tape recorder, which would in turn enable LaSalle to create a voice exemplar. Veazey refused to read the verbatim transcript of the message for he was concerned about how the tape might be used and because he thought the message was offensive. In a counteroffer, Veazey agreed to provide a tape- recorded voice exemplar of his reading of a different message. Because of his refusal to provide the specific voice exemplar requested, Mason suspended Veazey without pay. Three days later, Mason and Burke again summoned Veazey to a meeting, and once again Veazey refused to provide LaSalle with a voice exemplar of him reading a transcript of the threatening message. Based on Veazey's continued refusal to provide the requested voice exemplar, Mason discharged him for insubordination.

Thereafter, Veazey filed suit against LaSalle alleging that LaSalle's decision to terminate him after he refused to provide the specific tape recorded message violated the EPPA's prohibition against employers, like LaSalle, from administering lie detector tests. LaSalle responded to Veazey's complaint with a motion to dismiss for failure to state a claim under Federal Rule of Civil Procedure 12(b) (6). The district court granted the motion to dismiss, agreeing with LaSalle that the tape recording requested by Mason and Burke did not qualify as a lie detector test as that term is used in the EPPA. Veazey appeals. We reverse and remand this case for further proceedings.

The issue in this case is whether LaSalle's specific request that Veazey produce a voice exemplar of him reading a transcript of the threatening voicemail message amounts to a "lie detector test" under the EPPA.

We review a dismissal under Rule 12(b) (6) de novo, taking a plaintiff's factual allegations as true and drawing all reasonable inferences in the plaintiff's favor. See Strasburger v. Board of Educ., Hardin County Community Unit Sch. Dist. No. 1, 143 F.3d 351, 359 (7th Cir. 1998), cert. denied, 119 S. Ct. 800(1999); Americanos v. Carter, 74 F.3d 138, 140 (7th Cir. 1996). A complaint should not be dismissed for failure to state a claim "unless no relief could be granted 'under any set of facts that could be proved consistent with the allegations.'" Nance v. Vieregge, 147 F.3d 589, 590 (7th Cir. 1998), cert. denied, 119 S. Ct. 426 (1999) (quoting Hishon v. King & Spalding, 467 U.S. 69, 73 (1984)). In other words, if it is possible to hypothesize a set of facts, consistent with the complaint, that would entitle the plaintiff to relief, dismissal under Rule 12(b) (6) is inappropriate. See Graehling v. Village of Lombard, Ill., 58 F.3d 295, 297 (7th Cir. 1995).

The dissent states that " [s]ince Veazey has not claimed that any recording would have been subjected to a stress analyzer, his complaint cannot survive a 12(b) (6) dismissal." In making this statement the dissent misconstrues the burden a plaintiff is under when defending a motion to dismiss under 12(b) (6). As the dissent must be aware, federal courts have notice- pleading, not fact-pleading, requirements. As this court has stated recently, " [i]n contrast [to fact-pleading], the federal rules follow the notice pleading approach, requiring only a 'short and plain statement of the claim showing that the pleader is entitled to relief.' Rule 8(a) (2). The only function the pleadings must serve is to give notice of the claim; the development of legal theories and the correlation of facts to theory come later in the process." International Marketing, Ltd. v. Archer-Daniel-Midland Co., Inc., 192 F.3d 724, 733(7th Cir. Sept. 23, 1999) (emphasis added). Indeed, under the liberal notice pleading requirements of the federal rules, " [a]ll that's required to state a claim in a complaint filed in federal court is a short statement, in plain (that is, ordinary, nonlegalistic) English, of the legal claim." Kirksey v. R.J. Reynolds Tobacco Co., 168 F.3d 1039, 1041 (7th Cir. 1999). See Bennett v. Schmidt, 153 F.3d 516, 518 (7th Cir. 1998); Duda v. Board of Educ. of Franklin Park Pub. Sch. Dist. No. 84, 133 F.3d 1054, 1057 (7th Cir. 1998). Furthermore, the claim must be sustained "if any facts that might be established within those allegations would permit a judgment for the plaintiff." Duda, 133 F.3d at 1057. Thus, the fact that Veazey's complaint does not specifically state that the recording would have been subjected to a stress analyzer, it does not render his complaint susceptible to a successful 12(b) (6) motion to dismiss. If one keeps the principles discussed above in mind, it becomes apparent that Veazey's complaint should not have been dismissed for failure to state a claim.

The polygraph is composed of a combination of devices which measure certain, specified physical data.³  In 1895, an Italian psychiatrist and criminologist named Cesare Lombroso made the unprecedented claim that he could "detect lies" by monitoring a person's blood pressure and "reading" the changes in it. See Michael Tiner & Daniel J. O'Grady, Lie Detectors in Employment, 23 Harv. C.R.-C.L. L. Rev. 85, 85-86 (1988). Lombroso asserted that by understanding the typical criminal responses and physical characteristics he could distinguish "criminal types" from the rest of society. See id. Over a hundred years later, his claims continue to shape society's perceptions of polygraphs and account for their popularity. See id.; Timothy B. Henseler, A Critical Look at the Admissibility of Polygraph Evidence in the Wake of Daubert: The Lie Detector Fails the Test, 46 Cath. U. L. Rev. 1247 (1997); see also David Thoreson Lykken, A Tremor in the Blood: Uses and Abuses of the Lie Detector (1981). In order to understand the importance of the EPPA, and why LaSalle may have violated it, we believe it is helpful for one to understand the evolution of the lie detector and its impact on employers and employees.

Before the EPPA, federal law regulated the use of polygraph machines mainly in law enforcement contexts, but made no attempt to control or monitor their use in the private workplace. It was not a surprise when private employers took it upon themselves to administer more polygraph examinations than either the federal government or state criminal investigators. See Congressional Office of Technology Assessment, Scientific Validity of Polygraph Testing: A Research Review and Evaluation (A Technical Memorandum), OTA-TM-H-15, 98th Cong., 1st Sess. 5 (1983). In fact, a 1978 survey of four hundred major U.S. corporations found that more than fifty percent of the commercial banks and retailers that had responded to the survey used polygraphs. See Belt & Holden, Polygraph Usage Among Major U.S. Corporations, 51 Personnel J. 80, 86 (February 1978). The survey also noted that these companies were more likely to test all job applicants and employees than to conduct random sampling. See id.

As polygraph machines gained popularity in the American business world, many researchers and defense lawyers began to question the accuracy of the machine that was dictating numerous peoples' employment fate. Several studies concerning polygraph validity were published in the late 1970's and early 1980's, and contributed greatly to the understanding of the lie detector's limitations, motivating the United State Congress to pass the EPPA.⁴ 

Field Studies

1. Benjamin Kleinmuntz & Julian J. Szucko, On the Fallibility of Lie Detection, 17 L. & Soc'y Rev. 85 (1982)

In 1982, Kleinmuntz and Szucko obtained the charts of one hundred polygraph examinations which were performed by the then well-known Reid Polygraph Agency in Chicago, Illinois. The study consisted of fifty charts that had been verified as deceptive by the subsequent confessions of the examinees and fifty charts that had been verified as truthful by the subsequent confessions of other people. Polygraphers from the well recognized Reid agency then independently rescored all one hundred charts, incorrectly classifying 39% of the verified innocent examinees as guilty.

2. Frank Horvath, The Effect of Selected Variables on Interpretation of Polygraph Records, 62 J. Applied Psychol 127 (1977)

In 1977, Horvath published a polygraph validity study using fifty-six polygraph examination charts--all of which had been verified by using subsequent confessions made to police. Ten polygraphers then independently rescored the examination charts. Of the now established innocent examinees, only 51% were correctly scored as truthful when denying their guilt-- hardly better than simply flipping a coin.

3. Barland & Raskin, An Evaluation of Field Techniques in Detection of Deception, 12 Psychophysiology 321 (1976)

In this study, the guilt or innocence of the suspects was determined by an expert panel of one judge, two defense lawyers, and two prosecutors who examined each suspect. Barland then administered polygraph examinations to ninety-two criminal suspects, and Raskin independently scored those charts.

Based on the decisions of the expert panel, Raskin incorrectly classified 55% of the innocent suspects as deceptive when they denied their guilt. Once again, the survey suggests that employees (or employers for that matter) might fare just as well if their fate was determined by a simple flip of a coin.

4. Congressional Office of Technology Assessment (O.T.A.) Summary of Studies (1983)

The O.T.A. reviewed ten field studies of polygraph validity and found that the results of these studies varied widely. See Congressional Office of Technology Assessment, Scientific Validity of Polygraph Testing: A Research Review and Evaluation (A Technical Memorandum), OTA-TM- H-15, 98th Cong., 1st Sess. 5 (1983). O.T.A. summarized its findings as follows:

1. false negatives (incorrectly classifying a deceptive person as truthful) varied from 29.4% to 0%;

2. false positives (incorrectly classifying a truthful individual as deceptive) varied from 75% to 0%;

3. inconclusive results varied from 25% to 0%;

4. correct guilty detections varied from 98.6% to 70.6%;

5. correct innocent detections varied from 94.1% to 12.5%.

The significance of a 90%, 80%, or 70% polygraph validity rate cannot be fully understood unless one understands what that figure means to an individual seeking employment or facing criminal charges. In fact, O.T.A. determined that the mathematical chance of false positives is greatest when polygraphs are used randomly to test large numbers of employees because, according to O.T.A., only a small percentage of screened individuals are actually guilty. For example, if one out of one thousand people is actually guilty and we posit that a polygraph will be 99% accurate in determining truthful statements, then the law of probability would dictate that not only would one person be correctly identified as guilty but so would ten innocent people. Given the results of studies demonstrating validity rates much lower than 99%, the negative impact of polygraph screening on innocent individuals is in reality far greater than the hypothetical would suggest, and the actual consequences of invalid polygraph examinations, therefore, affect a far greater number of people in employment situations. See David Gallai, Poligraph Evidence in Federal Courts: Should it be Admissible, 36 Am. Crim. L. Rev. 87, 98 n.64 (1999) ("' [N]o overall measure [of validity] . . . can be established based on available scientific evidence' and that polygraphs 'detect [ ] deception better than chance, but with error rates that could be considered significant.'" (quoting Congressional Office of Technology Assessment, supra)).

Laboratory Studies⁵ 

In 1975, two researchers from the University of Utah conducted a study of polygraph validity involving seventy-two student volunteers. A mock "crime" involving the theft of a valuable object or of a small sum of money was created--half the students involved were "guilty," while the other half were assigned the role of an "innocent" individual. Each student was then administered a polygraph examination.

Thirty-five percent of the polygraph examinations were inconclusive, and 12% were incorrect. Of the incorrect examinations, two-thirds were false positives, and one-third were false negatives. In all, this study revealed that only 53% of the test subjects were correctly identified by the polygrapher as being either guilty or innocent.

In 1978, an extensive validity study by Podlesny and Raskin required polygraphers to make decisions of guilt or innocence based, not on a polygraph, but on the visual observation of the test subjects. Without listing the specific results of the research, the test revealed that the behavioral observations of the polygraphers were more accurate than the polygraph overall.

In 1981, Barland conducted one of the few validity studies ever done on the use of polygraphs in pre-employment screening by testing military personnel who worked in the intelligence field. For the study, Barland told half of the participants to lie when responding to one of the questions--offering twenty dollars if they could produce a truthful reading during the polygraph examination. Again, without going into the specifics of the study, the O.T.A. stated in its 1983 report that "the results of the Barland study raise serious questions about the usefulness of directed lie control questions in screening procedures as well as, in general, the validity of polygraph testing for pre-employment and counterintelligence purposes, especially if used alone."

4. U.S. Army Land Warfare Laboratory Study (1974)⁶ 

In one of the most complete laboratory studies on the validity of the polygraph and other "truth verification devices," Dr. Joseph F. Kubis of Fordham University, who was commissioned by the U.S. Army Land Warfare Laboratory, divided subjects into three basic roles--thief, lookout, and innocent suspect--in a simulated theft situation. After being present during the subject's examination, experienced polygraphers made incorrect evaluations 24% of the time. When other examiners, without having the opportunity to view the subjects as they were tested, rated the polygraph charts, accuracy dropped below 60%.

Armed with these studies⁷  and perhaps recognizing that jobs were too important and the economy too scarce to allow an inaccurate machine to dictate the fate of millions of Americans, the EPPA was enacted, prohibiting private employers, in most situations, from subjecting job applicants or employees to lie detector tests. Ever since, employers have been desperately seeking other ways to evaluate the honesty and other characteristics of job applicants and employees. See Julie A. Spoh, The Legal Implications of Graphology, 75 Wash. U. L.Q. 1307 (1997) (analyzing the rise in the use of graphology, the alleged science of divining personality from handwriting, by American companies in employment decisions); Stephen F. Befort, Pre-Employment Screening and Investigation: Navigating Between a Rock and a Hard Place, 14 Hofstra Lab. L.J. 365 (1997) (discussing the attempts of employers to adopt pre-employment screening practices in light of various federal and state laws); see also Robert B. Fitzpatrick, Lie Detectors Belong in Museums, Not in Sexual Harassment Trials, SD06 ALI-ABA 889 (1998). As we shall see, LaSalle may have been trying to skirt the EPPA's prohibition on lie detector tests when it requested Veazey to provide a tape recording, and discharged him for refusing to do so.

As stated above, Congress enacted the Employee Polygraph Protection Act in 1988 in response to justified concerns that employers were many times misusing lie detectors or their derivatives (as the case may have been here) and were too frequently relying on inaccurate, inconclusive, or unfounded lie detector results to make employment decisions. See S. Rep. No. 100-284, at 46 (1988), reprinted in 1988 U.S.C.C.A.N. 726, 734. Accordingly, the EPPA makes it illegal, with limited exceptions, for employers to use lie detector tests. See 29 U.S.C. §sec. 2001-09. In particular, the provisions applicable to this case provide, in relevant part:

[I]t shall be unlawful for any employer engaged in or affecting commerce or in the production of goods for commerce--

(1) directly or indirectly, to require, request, suggest, or cause any employee or prospective employee to take or submit to any lie detector test;

* * *

(3) to discharge, discipline, discriminate against in any manner, or deny employment or promotion to, or threaten to take any such action against--

(A) any employee or prospective employee who refuses, declines, or fails to take or submit to any lie detector test

29 U.S.C. § 2002 (1999).

The statute defines the term "lie detector" to include "a polygraph, deceptograph, voice stress analyzer, psychological stress evaluator, or any other similar device (whether mechanical or electrical) that is used, or the results of which are used, for the purpose of rendering a diagnostic opinion regarding the honesty or dishonesty of an individual." 29 U.S.C. § 2001(3) (1999) (emphasis added). We point out the illuminating language in this statute (which is obviously broad enough to preclude the resolution of this case on a motion to dismiss) and query whether today's technology permits a tape recording, if used in conjunction with other devices enumerated in the statute, to achieve the results of a "lie detector." The dissent seems to forget the focus of the case and states that "Veazey's assertion that a tape recorder qualifies as a lie detector has no foundation under the terms of the [EPPA], and simply using a tape recorder to compare voice samples does not violate the Act." The dissent ignores the fact that Veazey never limited his complaint to simply using a tape recorder to compare voice samples. While the dissent is correct that simply using a tape recorder to compare voice samples would not violate the EPPA, that is not what this case is about. Rather, the case is about, as the majority opinion makes perfectly clear, the tape recording LaSalle requested being used in conjunction with other devices that would have allowed LaSalle to directly gauge whether Veazey was telling the truth when he denied leaving the threatening message on another employee's answering machine. Examples of fairly common lie detection methods suggest that it is possible.⁸ 

Initially, LaSalle contends that a tape recorder, the device it wanted to use to make the requested voice exemplar, is not a "lie detector" as defined by the EPPA because it, by itself, does not directly gauge a person's truthfulness; it merely records sounds. LaSalle's characterization of the statutory definition of "lie detector" is, in our opinion, too narrow because it overlooks the significance of the phrase "the results of which are used . . . for the purpose of rendering a diagnostic opinion regarding the honesty or dishonesty of an individual." LaSalle concedes that a voice exemplar, evaluated by a voice stress analyzer or similar device, might be and often is used in the rendering of a "diagnostic opinion regarding the honesty or dishonesty of an individual." We are of the opinion that the application of basic logic necessitates that a tape recorder might very well be considered as an adjunct to a "lie detector" determination under the EPPA because the results of a tape recording (a voice exemplar) can be used to render a diagnostic opinion regarding the honesty or dishonesty of an individual when evaluated by a voice stress analyzer or similar device. Accordingly, a tape recorder, when used in conjunction with one of the devices enumerated in the statute or a similar device, may fit within the definition of a "lie detector" under the EPPA.

Furthermore, the narrow interpretation advanced by LaSalle removes many of the protections afforded by the EPPA. For instance, if an employer could tape record an interview with an employee, then turn around and analyze the tape recording with a voice stress analyzer, the EPPA's prohibition on the use of voice stress analyzers would be eviscerated. As the Supreme Court has stated, it is extremely unlikely that an interpretation that allows a statute to be so easily evaded would be the correct one. Cf. Hathorn v. Lovorn, 457 U.S. 255, 265 n.16 (1982) (refusing to interpret sec. 5 of the Voting Rights Act in a way that would allow covered jurisdictions to easily evade the statute); McElroy v. United States, 455 U.S. 642, 655-56 (1982) (refusing to interpret a provision of a federal statute criminalizing forgery to allow a "patient forger" to easily evade the provision); Wilson v. Omaha Indian Tribe, 442 U.S. 653, 666- 67 (1979) (refusing to limit statute regarding disputes over Native American lands in a way that would allow a party to "escape its reach merely by incorporating and carrying on business as usual"); Northern Ind. Pub. Serv. Co. v. Carbon County Coal Co., 799 F.2d 265, 271 (7th Cir. 1986) (noting that a provision of the Mineral Lands Leasing Act preventing railroads from acquiring coal leases would undoubtedly cover a "facile evasion" of the provision through the use of a dummy corporation).

Finally, the EPPA's legislative history indicates that Congress intended the prohibition on the use of lie detectors to be interpreted broadly. See H.R. Conf. Rep. No. 100-659, at 11 (1988), reprinted in 1988 U.S.C.C.A.N. 749, 750 ("The conferees . . . intend that the prohibition on a lie detector test be construed broadly to include any use of a lie detector." (emphasis added)). By excluding devices that might be used in conjunction with those devices specifically prohibited by the EPPA, LaSalle's narrow interpretation of the definition of "lie detector" contravenes Congress' expressed intention. Accordingly, we conclude that the EPPA's definition of "lie detector" does not, as a matter of law, exclude the use of the specific voice exemplar LaSalle requested Veazey to produce.

Our holding in this case does not mean that a tape recorder invariably must be considered a "lie detector" under the EPPA. A tape recorder would not fall within the statutory definition if it was not used in conjunction with another device that assists in the gauging of a person's truthfulness.⁹  Perhaps realizing this, LaSalle also argues that the recording requested in this case could not be used, even in conjunction with another device, to render "a diagnostic opinion regarding the honesty or dishonesty of an individual" and that therefore, the tape recorder intended to be used in this case cannot be considered a "lie detector." LaSalle continues and states that because the requested recording could be used to gauge Veazey's truthfulness only by way of comparison with the actual message left on Ralph Newcomb's voicemail and not by some independent assessment of the tape recording, the tape recording cannot be used to render a diagnostic opinion concerning Veazey's honesty or dishonesty. We disagree with LaSalle's characterization of how the tape recording might be used.

LaSalle and the dissent¹⁰  are correct to distinguish between devices (and presumably combinations of devices) that can be used to directly gauge a person's truthfulness and those that only indirectly determine whether a person has been truthful. For example, the machines used to analyze DNA samples often tell us whether a suspected perpetrator is being truthful when he denies committing a crime. See generally Steven I. Friedland, Law, Science and Malingering, 30 Ariz. St. L. J. 337, 339-40 (1998); Jon Devlin, Genetics and Justice: An Indigent Defendant's Right to DNA Expert Assistance, 1998 U. Chi. Legal F. 395. In fact, one court has stated, " [DNA testing] constitute [s] the single greatest advance in the 'search for truth,' and the goal of convicting the guilty and acquitting the innocent, since the advent of cross-examination." New York v. Wesley, 533 N.Y.S.2d 643, 644 (Sup. Ct. 1988). Notwithstanding the virtues (even if not as great as Wesley suggests) of DNA testing, the fact remains that these sorts of devices, even if assumed infallible,¹¹  do not supply data that, without more, allow an operator to determine whether someone is lying. At most, the results are only evidence of a historical fact; independently these types of devices can not be used to determine whether a subject is telling the truth. Such devices to date have not been recognized as being able to render a "diagnostic opinion" concerning a person's truthfulness. Moreover, the EPPA's legislative history confirms the fact that those devices which only indirectly indicate whether a person is lying should not be included in the definition of "lie detector." See S. Rep. No. 100-284, at 47 (1988), reprinted in 1988 U.S.C.C.A.N. 726, 734-35 (noting that a telephone used to check application information would not be a "lie detector").

While LaSalle and the dissent are justified in distinguishing between devices that directly determine whether a person is being truthful and those that do so only indirectly, its contention that a tape recording of Veazey reading a verbatim transcript of the threatening message could not be used to directly gauge Veazey's truthfulness cannot support a dismissal under Rule 12(b) (6). LaSalle's contention depends on a proposition of fact, specifically whether current technology is capable of determining if someone is telling the truth based upon a tape recording. On a motion to dismiss under Rule 12(b) (6), the only "facts" favorable to a defendant that a court can consider are those alleged in the plaintiff's complaint. See General Elec. Capital Corp. v. Lease Resolution Corp., 128 F.3d 1074, 1080 (7th Cir. 1997). As stated before and what the dissent ignores is the fact that there is nothing in Veazey's complaint that excludes the possibility that the requested recording could be used by itself or in conjunction with other devices to render a diagnostic opinion concerning Veazey's truthfulness. Accordingly, a dismissal under Rule 12(b) (6) based on the "fact" that the recording LaSalle requested could not be used to render a diagnostic opinion on Veazey's truthfulness is inappropriate.

This is so notwithstanding LaSalle's suggestion that Veazey has waived many of the arguments he raises. LaSalle contends that Veazey's failure to specifically articulate how the requested voice exemplar would or could be used to gauge his truthfulness before the district court judge constitutes a waiver of Veazey's right to speculate on appeal about how LaSalle might use the voice exemplar. This court's decision in Highsmith v. Chrysler Credit Corp., 18 F.3d 434 (7th Cir. 1994), forecloses LaSalle's waiver argument. Highsmith instructs that when this court reviews Rule 12(b) (6) dismissals, "we will consider new factual allegations raised for the first time on appeal provided they are consistent with the complaint." Id. at 439-40 (noting, "' [t]his rule is necessary to give plaintiffs the benefit of the broad standard for surviving a Rule 12(b) (6) motion . . . .'" Id. (quoting Dawson v. General Motors Corp., 977 F.2d 369, 372 (7th Cir. 1992)). Accordingly, to the extent that Veazey failed to specifically articulate exactly how the requested exemplar could be used to gauge his truthfulness before the district court, his failure does not serve to waive the arguments he makes on appeal because these arguments are merely scenarios consistent with his complaint.

There is nothing in the EPPA or Veazey's complaint that excludes the possibility that the specific tape recording LaSalle requested qualifies as a lie detector test under the EPPA. Accordingly, we hold that it was improper for this case to have been dismissed for failure to state a claim under Rule 12(b) (6). We REVERSE the judgment of the district court and direct that this case be REMANDED for further proceedings consistent with this opinion.

MANION, Circuit Judge, dissenting.

I would affirm the district court's dismissal of Veazey's complaint. I have no quarrel with this court's thorough analysis of the shortcomings of polygraphs and their fallibility when used as evidence of truthfulness. But Veazey's assertion that a tape recorder qualifies as a lie detector has no foundation under the terms of the Employee Polygraph Protection Act, 29 U.S.C. § 2002(3), and simply using a tape recorder to compare voice samples does not violate the Act.

EPPA restricts certain employers from administering lie detector tests to their employees. 29 U.S.C. § 2002(3).¹  The statute defines the term "lie detector" as including "a polygraph, deceptograph, voice stress analyzer, psychological stress evaluator, or other similar device (whether mechanical or electrical) that is used, or the results of which are used, for the purpose of rendering a diagnostic opinion regarding the honesty or dishonesty of an individual." 29 U.S.C. § 2001(3). Thus, the EPPA prohibits the use of devices which are designed to determine whether particular assertions or answers to questions are true or false, and nothing more. Accordingly, LaSalle could not have used a voice stress analyzer on a tape recording of a statement by Veazey to determine whether the contents of the recorded statement were true or false. But the Act does not prohibit employers from using other investigative techniques to identify employees who may have violated criminal laws or company regulations. Therefore, the district court correctly held that the EPPA does not prevent employers from seeking voice samples from an employee suspected of leaving harassing messages on another employee's voicemail, in order that the employer might compare the voices. Since Veazey has not claimed that any recording would have been subjected to a stress analyzer, his complaint cannot survive a 12(b) (6) dismissal.

Veazey does assert that the term "lie detector" includes a tape recorder, and that EPPA therefore prohibits LaSalle from firing him for failing to submit a recorded voice sample. See Complaint para. 17. Notably, Veazey provided neither statutory nor case law support for this assertion. "When construing a statute, we must first look to the language used by Congress, giving the words their ordinary meaning. ' [A]bsent a clearly expressed legislative intention to the contrary, that language must ordinarily be regarded as conclusive.'" United States v. Wilson, 159 F.3d 280, 291-92 (7th Cir. 1998) (quoting American Tobacco Co. v. Patterson, 456 U.S. 63, 68 (1982)). Veazey's interpretation is clearly at odds with the statute's definition of the term "lie detector," as a tape recorder is not one of the devices which the statute specifically proscribes. Furthermore, merely recording a voice or comparing recorded voices does not result in a "diagnostic opinion" regarding the honesty or dishonesty of the individual as the term is defined by the statute. Unless subjected to the statutorily forbidden "voice stress analyzer," the tape recording of his voice is permissible. A plain reading of the statute thus does not support Veazey's interpretation.

Because the term "lie detector" means devices such as polygraphs, deceptographs, and voice stress analyzers, and a tape recorder is not similar to these devices (it records sounds; it does not analyze the sounds), a tape recorder cannot be a lie detector as the term is defined by the statute.²  Therefore, an employer may terminate an employee for refusing to provide a voice sample for purposes of comparative analysis without violating the EPPA, much like an employer may terminate an employee for failing to provide a urine sample for a drug test. Gonzalez v. Ingersoll Milling Mach. Co., 133 F.3d 1025, 1034 (7th Cir. 1998) (under the employment at-will doctrine an employee may be fired for any reason that does not violate law or public policy); see Slane v. Mariah Boats, Inc., 164 F.3d 1065, 1066 (7th Cir. 1999) (employer claimed that it fired employee for failing to provide a urine sample).³  As with a urine sample, a tape- recorded voice sample will require further analysis. If a person claims he is drug-free, a positive urine test will prove he is lying. Similarly, if a person claims a recorded threat is not his voice, but a comparative voice graph shows the threatening voice was his, he is lying. But the mechanical devices used to analyze the data in the urinalysis or voice analysis are not "similar devices" to the polygraph, the voice stress analyzer, or the other examples of lie detectors set out in the statute. Accordingly, employers may legitimately use tape recorders in ferreting out sexual harassment or other violations of law or company policy. They merely must refrain from subjecting those recordings to lie detector tests, such as a voice stress test.⁴  Because Veazey's complaint does not allege that LaSalle was attempting to do anything other than a comparative analysis, the district court correctly found that he failed to state a claim under EPPA.