DAVID A. SUJAK Department of Psychology Northern Illinois University

PETER VILLANOVA JOSEPH P. DALY Department of Management Appalachian State University

ABSTRACT. This study is an investigation of the effects of a mandatory versus probablecause drug-testing program and the level of restrictiveness of confidentiality policy on potential applicants' attitudes and intentions to apply for employment in an organization. One hundred fifty-six American students were randomly assigned to four experimental groups conforming to a 2 x 2 analysis of variance design: Type of Program (mandatory versus probable cause) x Level of Confidentiality (low versus high). A less restrictive confidentiality policy was associated with lower intentions to apply for employment relative to a more restrictive confidentiality policy. There were no perceived main or interactive effects involving a drug-testing program. Perceived procedural fairness partially mediated the effects of confidentiality on application intentions, and, in turn, an applicant's attitude toward the organization completely mediated the effects of perceived procedural fairness. These findings showed that procedural variables may be important determinants of applicant responses toward potential employers that maintain drug-testing programs.

TESTING FOR DRUG USAGE in business organizations in the United States has been on the increase as a response to employee substance abuse. Fifty percent of the Fortune 500 firms used drug testing in 1988, compared with 5% in 1982 (Rothman, 1988). Such programs are expected to result in increased efficiency and productivity, improved attendance, lower turnover, fewer accidents, and higher morale ( Cohen, 1984, cited in Crant & Bateman, 1989).

However, some drug-testing programs have been criticized for engendering negative reactions in job applicants, including the perception of drug testing as harassment of workers and as representing a repressive work environment ( Rothman , 1988; Vodanovich & Reyna, 1988). Such perceptions may lead to self-

selection out of the applicant pool, which could result in fewer qualified applicants available to the organization ( Fraser & Kroeck, 1989; Garland, Giacobbe, & French, 1989). Thus, employers who maintain drug-testing programs ought to know which program characteristics engender negative applicant perceptions.

Currently, the most comprehensive model of employee reactions to drug testing was proposed by Crant and Bateman ( 1989). According to their model, the most significant variable for predicting employee reactions to drug testing is employees' perceived fairness of the drug-testing program. In their model, fairness is viewed as a direct determinant of employees' attitudinal and behavioral responses. In the present study, we used the Crant and Bateman model as our starting point by investigating the effects of drug-testing program characteristics on applicant perceptions of fairness and attitudes toward potential employers.

There is little empirical evidence concerning employee or applicant responses to drug testing. In a partial test of their model, using undergraduate business majors, Crant and Bateman ( 1990) found that potential applicants' attitudes and intentions to apply were affected negatively by whether the company tested for drugs and had a demonstrable need for such a program. In their model, these factors relate mainly to perceptions of distributive justice (i.e., the perceived fairness of outcomes received from a decision).

Stone and Kotch ( 1989) investigated current American employees' attitudes toward a hypothetical company with a drug-testing program. They found that both advance notification of testing and rehabilitative (as opposed to punitive) consequences of detected use correlated positively with acceptance of the program. The consequences of a positive test result are related to distributive justice. However, advanced notification of testing is related to information ( Sheppard & Lewicki, 1987), a proposed standard of procedural justice (i.e., the perceived fairness of procedures used to make a decision).

Konovsky and Cropanzano ( 1991) found that fairness perceptions of drugtesting procedures predicted employee attitudes and job behaviors. Their study was limited, however, by the fact that they did not identify specific components of drug-testing programs that influence these fairness perceptions. They suggested that concrete procedural policies be investigated that might affect fairness perceptions.

Several researchers have suggested two key components of guidelines for an effective drug-testing program: keeping the results of testing strictly confidential and avoiding the use of random and across-the-board testing (Bible & McWhirtyer

tyer, 1990; Cowan, 1987; Crown & Rosse, 1988; Gomez-Mejia & Balkin, 1987; Wrich, 1988). As yet, no research evidence exists to support these prescriptions. In the present study, we examined whether the type of drug-testing program and restrictiveness of confidentiality are related to procedural fairness judgments and potential applicants' responses to a drug-testing program, as suggested by the Crant and Bateman ( 1989) model and the studies reviewed above.

According to Westin ( 1967), "Privacy is the claim of individuals, groups, or institutions to determine for themselves when, how, and to what extent information about them is communicated to others" (p. 7). In employment settings, there is often a conflict between employees' rights of privacy and employers' need for information about individuals in order to operate their organizations effectively and safely ( Schein, 1977). Confidentiality of test results ensures that employers do not abuse the power they have over their employees.

Currently, American employers are protected against liability for defamation by a limited privilege to disclose employee personnel records ( Crown & Rosse, 1988). However, an employer risks liability for defamation if he or she reveals to another employer that an applicant failed a drug test ( Bible & McWhirter, 1990). Even within a firm, such information may be revealed to supervisors and managers only on a "need to know basis" ( Cowan, 1987; Crown & Rosse, 1988).

Participants in two studies perceived a greater invasion of privacy when they neither had control over nor gave permission for the release of information, and when negative consequences (i.e., not receiving a job offer) resulted from such disclosure ( Fusilier & Hoyer, 1980; Tolchinsky, et al., 1981). A survey of American human resource management (HRM) executives also yielded findings indicating that confidentiality is positively correlated with higher perceived drug program effectiveness ( Gomez-Mejia & Balkin, 1987). In addition, Murphy, Thornton, and Reynolds ( 1990) found that American college students maintain more positive attitudes about drug tests when results are kept confidential versus when they are disclosed to others.

In an analysis of the relationship between invasion of privacy and attitudinal and behavioral responses, Bies ( 1993) maintained that violations of privacy expectations frequently erode the harmed individual's trust in the procedures used by the organization (p. 81). Barber ( 1983) argued that trust in American business organizations is upheld through two means: (a) reliable role performance, or trust based on the competence of organizational authorities, maintained by the operations of the market system; and (b) fiduciary trust, or trust based on obligations of morality and fairness that arise from the nature of the relationships existing between business firms and their stakeholders.

In the context of a potential job applicant's responses to the confidentiality provisions of a drug-testing program, Barber ( 1983) notion of trust through reliable role performance is maintained by the applicants' right to refrain from applying to an organization that does not adequately respect their privacy. Fairness is not likely to play a role in judgments of the employer's behavior in such an instance as long as the applicant believes in the fairness of the overarching market system and believes that the procedures used by the employer do not violate market definitions of fairness.

Both applicant and employer are thus acting as parties in an exchange relationship, each exercising his or her own power to influence the outcome of the exchange in his or her favor. The judgments made by the applicant in this conceptualization are independent of fairness. This is consistent with Etzioni ( 1988) contention that people are motivated not only to pursue pleasure and avoid pain, but to fulfill moral commitments and duties, as well. Etzioni argues that these two sources of motivation are independent, and we believe them to be distinct in the context of the present argument. Therefore, we believe that confidentiality of drug-testing procedures has a direct effect on attitudes toward an organization and intention to apply that is not mediated by fairness judgments. Specifically, we expected that there would be more negative attitudes and less intention to apply to an organization that has less rigorous procedures for ensuring confidentiality of drug test results (Hypothesis 1).

We decided to study mandatory random testing of all employees versus testing only when there are reasonable grounds because these are the two principal types of drug-testing programs used in industry ( Cowan, 1987; Gerstein & Grossman, 1989; Masi & Burns, 1986). In random drug testing, employees are asked to submit a urine sample without prior notice. In a true random test procedure, all employees at all levels of the organization may be randomly selected for screening at any time.

Testing only when there is probable cause ("fitness for duty testing") is done when there is a reasonable suspicion of drug use. Most organizations follow procedures that spell out what constitutes reasonable suspicion. Some examples of behavior that may indicate drug use are accidents, violent outbreaks, excessive medical claims, and bizarre behavior ( Cowan, 1987).

Random testing is more controversial because it is done without cause ( Masi & Burns, 1986). If there is no impairment of the employee's job performance, as is required for screening with probable cause, testing is more intrusive and can lead to distrust, resentment, and hostility. Random testing is also more likely to be perceived as a violation of privacy rights ( Bible & McWhirter, 1990). In a study of American HRM executives, random drug testing was the characteristic most associated with lower perceived program effectiveness ( Gomez-Mejia &

Balkin, 1987). Additionally, survey findings of American college students' attitudes toward drug-testing programs indicated that random testing is viewed less favorably than testing on suspicion of drug use ( Murphy et al., 1990).

It follows, therefore, that the type of drug-testing program used should have a direct effect on applicant's attitude and intention to apply for a job. However, the direction of the effect is unclear. Distributive justice is likely to be a mechanism for the effects of the type of program on an applicant's response: An applicant who is selected to undergo drug testing is likely to consider the fairness of that outcome, which will in turn influence his or her response. Nevertheless, a number of standards for judging distributive justice have been proposed, most notably equity ( Adams, 1965) and equality ( Deutsch, 1975).

The equity standard dictates that outcomes be allocated to recipients in proportion to their inputs (i.e., that those most deserving of an outcome should receive it). A probable-cause drug-testing program is likely to be viewed as fair if those selected for the program use the equity standard to judge it. However, the equality standard of distributive justice, which specifies that outcomes should be allocated equally, would, if applied to drug testing, favor a random program in which all applicants have an equal chance of being selected for a test.

These two perspectives suggest that either program might be perceived as more fair from the perspective of potential applicants. Therefore, because competing hypotheses about the direction of the effect are plausible, we decided not to favor one prediction over the other in the present context. But it is a legitimate empirical question to ask whether potential applicants would possess more favorable attitudes toward an organization or have greater intentions to apply in each instance. We expected that potential job applicants' attitudes toward and intentions to apply to an organization would differ as a function of whether the organization uses a mandatory, random drug-testing program as opposed to one that conducts tests only when there is reasonable suspicion of drug usage (Hypothesis 2).

In the employee responses model ( Crant & Bateman, 1989), the effects of procedural characteristics on applicants' attitudes and intentions to apply are mediated by the perceived fairness of the drug-testing procedure. This aspect of their model is consistent with theory and research on procedural justice (e.g., Folger, 1987; Thibaut & Walker, 1975; Tyler, 1984).

Tyler and Lind ( 1992) advanced a relational model of procedural justice predicting that trust will be an important determinant of procedural fairness perceptions. The form of trust described by Tyler and Lind corresponds to Barber ( 1983) fiduciary trust (described earlier) in that it is trust borne of relational expectations (here, pertaining to the relationship between an organizational authority and a prospective applicant). Tyler and Lind maintain that procedural

fairness becomes more important to group members as authorities become more powerful. Similarly, Barber predicts that fiduciary trust assumes greater importance in relationships where one party is more powerful than the other.

Both these arguments are consistent with the contention that procedures such as confidentiality protection and probable-cause testing are important to applicants insofar as they protect their interests in relation to a powerful organization. It follows from the Tyler and Lind ( 1992) model in particular that the effects of the type of drug-testing program and the confidentiality of the results on applicants' attitudes and intent to apply will be partially mediated by procedural fairness. We expected that perceived fairness of drug-testing procedures would partially mediate the effects of the program type and confidentiality on the attitudes of potential applicants and their intentions to apply for a job (Hypothesis 3).

The participants were 156 undergraduate students, 78 of whom belonged to a student organization for accounting majors at Northern Illinois University; the other 78 were enrolled in accounting classes at Elgin Community College. Eighty-two of the participants were men, and 74 were women; the average age was 23.19 years (mean year in school = 2.64).

We conducted a laboratory experiment in the spring of 1992 to test our hypotheses. Our independent variables were two types of testing program (random testing of all employees versus testing only when there is probable cause) and two levels of confidentiality of results (less rigorous versus more rigorous procedures). The resulting 2 x 2 factorial design created four experimental cells. Our dependent variables were the participants' attitudes toward the organization and intention to apply for a job with the organization.

The students were told that the study was an examination of how employment opportunity decisions are made. They were randomly assigned to the four experimental conditions, each consisting of 39 subjects. Participants in all conditions were asked to review a scenario containing a description of a hypothetical company that was hiring entry-level accountants. There were four different versions of the scenario, corresponding to the four experimental conditions. Each version contained a manipulation of information about one of the two types of drug-testing programs (mandatory, random testing for all employees versus test

ing only where there is reasonable suspicion) and the degree to which the results are confidential (less rigorous versus more rigorous procedures). These manipulations were embedded among extraneous information about salaries, benefits, job descriptions, and other conditions of employment. This filler information was the same for all versions.

After they read the scenario, the participants completed a questionnaire containing measures of the two dependent variables (attitudes toward the company and intention to apply), a manipulation check of the two independent variables, items about the perceived fairness of the drug-testing policy, and additional filler items. Upon completing the questionnaire, the participating students were debriefed.

We chose to study the attitudes and intentions of potential applicants rather than those of current employees because we wanted to assess whether potential applicants are being discouraged from applying to companies that have drug-testing programs. Using actual job applicants would have involved a number of ethical, as well as practical, problems for the study. Using current employees would have included a sample inconsistent with the population to which the results of this study should be generalized (i.e., job applicants).

We also decided to use a more specific job description than had been used in previous research, to increase the experimental realism of the study. By using a detailed job description, we thought that the participants in the study would be more involved with the procedure, thus enhancing the study's internal validity. To increase psychological involvement in the study through the use of a more specific job description, we were required to use a sample from a more specific population.

Each of the scenarios included a paragraph about the company's drug-testing program. In half of the versions, the program was described as one in which all current employees are tested twice yearly on a random basis. In the other half, the program was described as one in which testing is done only when there is reasonable suspicion for doing so, such as excessive absenteeism or erratic behavior on the job.

Included in the drug-testing policy paragraph was a statement about the degree of confidentiality of a positive test result. In the less rigorous confidentiality condition, it was stated that should there be a positive test result, minimally, the employee's immediate supervisor would be informed and the result would be noted in the employee's personnel record. Additionally, it was stated that permission would not have to be given by the employee in order for such information to be released by the company, even to subsequent potential employers.

For the more rigorous confidentiality condition, the scenario indicated that usually only the employee's immediate supervisor would know of a positive test

result. Others would be informed only on an emergency, need-to-know basis, and then only after permission had been sought from the employee for the release of the information.

The questionnaire included measures of both dependent variables: attitude toward the organization and intention to apply for a job with the company. Multiple items were used to measure each variable. Five items were used to measure attitude toward the company. Responses were measured using 7-point Likerttype scales, with anchors of very strongly agree and very strongly disagree. A sample item was "I feel that this is a good organization." The five items were summed to form a composite. Intention to apply was measured by three items. Responses to two of the items were measured on 7-point Likert-type scales, anchored similarly. These read, "I would like to apply for a job with this company," and "I would consider sending my resume to this company for this job position." There was also one 11-point item, for which the intervals were from 0% to a 100% probability of applying for a job with this organization. These items were summed to form a composite.

Additionally, items checked on the manipulations of the two independent variables. Examples of such items include "This company tests all employees on a random basis for drug use" and "Positive drug-test results are only given to employees' supervisors on a need-to-know basis." Also, four items asked about the fairness of the drug-testing policy. Fairness perceptions were measured using 7-point Likert-type scales, with the same anchors as mentioned previously. Again, item scores for each measure were summed to form a composite. Alpha reliability estimates for scales in the study were .91 for attitude toward the company, .80 for intention to apply, .88 for the program-type manipulation check, .84 for the confidentiality manipulation check, and .87 for perceived fairness of the drug-testing policy.

We used a 2 x 2 between-subjects factorial design to investigate the effects of the independent variables (i.e., program type and confidentiality) on the dependent variables, attitude toward, and intention to apply to, an organization. Factorial analyses of variance (ANOVAs) were performed on the manipulation check items for each of the independent variables. The results showed a significant main effect for confidentiality, F(1, 148) = 129.31, p 〈 .001, with M = 23.08 and M equals; 13.67 for students in the high-confidentiality and low-confidentiality conditions, respectively (a higher score indicated greater agreement with the high-confidentiality condition and greater disagreement with the low-confidentiality condition). A significant main effect was also found for program type, F(1,

148) = 241.62, p 〈 .001, with M = 21.35 and M = 8.18 for the subjects in the random program-type and probable-cause program-type conditions, respectively (a higher score indicated greater agreement with the random program-type condition and greater disagreement with the probable-cause condition). None of the other effects were significant in either ANOVA, indicating that the manipulations did have the intended effects.

Because the two independent variables were highly correlated (r = 71, p 〈 .001), we used a multivariate analysis of variance (MANOVA) to test the first two hypotheses in order to maintain a nominal alpha level of .05. Using Wilks's criterion, the MANOVA showed a significant main effect for confidentiality: λ = .934, F(2, 147) = 5.20, p 〈 .01. Univariate F tests indicated that only intention to apply contributed significantly to this effect, F(1, 148) = 9.54, p 〈 .01, as the effect on attitude toward the company was not significant, F(1, 148) = 2.29, ns. Cell means indicated that potential job applicants had more intention to apply to companies that had procedures that were high in confidentiality in regard to positive drug-test results (M = 19.96) than to companies that had procedures low in confidentiality of results (M = 18.07; see Table 1). Neither the main effect for program type nor its interaction with confidentiality showed a significant multivariate effect, λ = .993, F(2, 147) = .53, ns, and λ = .982, F(2, 147) = 1.34, ns, respectively.

To test Hypothesis 3, we performed a set of multiple regression analyses consistent with procedures described by Baron and Kenny ( 1986) and James and Brett ( 1984). Because the MANOVA results showed that the only significant main effect was for confidentiality on intention to apply, the regression analyses focused on the mediational role of perceived fairness on this relationship. The analyses included three regression equations. First, the proposed mediator (perceived fairness) was regressed on an effect-coded variable representing the degree of test result confidentiality (less rigorous procedures = -1, more rigorous

procedures = 1). Second, the dependent variable, intention to apply, was regressed on the confidentiality variable, coded as described previously. Finally, the dependent variable was regressed on the effect-coded confidentiality variable and on perceived fairness.

The results in Table 2 support Hypotheses 3, that procedural fairness would partially mediate the effect of confidentiality of test results on participants' intention to apply to an organization. Confidentiality had a significant effect on both fairness and intention to apply; more rigorous confidentiality procedures tended to be associated with higher levels of perceived fairness and greater intention to apply. Also, higher levels of perceived fairness were significantly associated with greater intention to apply. Additionally, inclusion of fairness as a predictor of intention to apply did reduce the significance of the confidentiality-intention to apply relationship.

Because of the high correlation between attitude and intention to apply, as well as between attitude and fairness (r = .40, p 〈 .001), we decided to perform a second set of multiple regression analyses to investigate whether attitude toward the organization might mediate the effects of fairness on intention to apply. This second potential mediation makes sense logically because one's attitudes are a putative cause of intention, and one's attitudes are more temporally proximal to intention than the other variables (i.e., fairness, confidentiality, and program type).

Again, these regression equations were calculated: Attitude was regressed on fairness and confidentiality (intention to apply had already been regressed on those same independent variables). Finally, intention to apply was regressed on attitude, fairness, and confidentiality. Fairness was found to be significantly associated with both attitude and intention to apply. When attitude was included in the final equation, it was significantly related to intention to apply. However, fairness now was not significant in its relationship to intention to apply. This pattern of results indicates that attitude acted as a complete mediator of the effects of fairness on intention to apply (Table 3).

Furthermore, this second model explains more system variance than the previous one, which included only perceived fairness as a mediator and omitted the further mediational role of attitude. This difference is demonstrated by the fact that the amount of the variance of intention to apply accounted for by the final equation in the second mediational analysis was 54%, F(5, 150) = 35.54, p 〈 .001, whereas the variance accounted for by the final equation in the original mediational analysis was only 14%, F(4, 151) = 6.21, p 〈 .001.

Because self-reported frequency of drug use among college students has been related to attitudes toward drug-testing programs ( Murphy et al., 1990), we measured participants' attitudes toward others' drug use, attitudes toward one's own drug use, and one's own drug use frequency. There were no significant main effects or interactions for confidentiality and program type on any of these measures.

The results of this study affirm the importance of confidentiality of positive test results on potential applicants' intention to apply. The significant main effect for degree of confidentiality of positive test results on intention provided partial support for our first hypothesis; less rigorous procedures for protecting the confidentiality of drug-test results caused lower intentions to apply for a job with an organization than one with more rigorous confidentiality procedures. This finding is consistent with the findings of Fusilier and Hoyer ( 1980) and Tolchinsky et al. ( 1981), who found that lack of control over the release of confidential information in employment application situations was perceived as an invasion of privacy. However, in the current study, confidentiality did not have a significant main effect on attitude toward the company.

We expected that there would be differences in participants' attitudes toward an organization and intention to apply as a function of the type of drug-testing

program. However, program type did not have a significant effect on either attitudes toward the organization or intention to apply for the job. Thus, Hypothesis 2 was not supported. The result may be attributable to the fact that this group of students reported strongly negative attitudes toward drug use, and few participants reported having ever used drugs themselves.

These drug-use attitudes and drug-usage frequencies were very low, with means across groups of 10.28 9.04, and 9.37, respectively, out of a potential low of 5 and a high of 35. These generally negative attitudes and low self-reported drug-use frequencies might mean that these participants felt that a random testing program would pose no particular threat to them because they did not use drugs themselves. Additionally, because they did not look favorably on drug use, they would probably have been more supportive of an intrusive type of testing program than individuals with more favorable attitudes toward drug use. This interpretation is supported by the Murphy et al. ( 1990) finding that college students' self-reported frequency of drug use was negatively correlated with the acceptability of employee drug testing.

The results also provide partial support for the predictions on Hypothesis 3; the multiple regression analyses indicated that perceived fairness ratings did partially mediate the effects of the program type and confidentiality on intention to apply for a job.

The results of the second set of regression analyses showed that the effects of fairness on intention were completely mediated by attitude and suggest that the causal linkages proposed in the Crant and Bateman model ( 1989) are somewhat incomplete. Their model hypothesized that perceived fairness would act as a mediator between procedural justice variables (e.g., confidentiality) and employee responses, without specifying a mediational role for attitudes. Our results indicate an indirect path from confidentiality to intent to apply that includes procedural fairness, and in turn, attitudes toward the firm, as mediators of the effects of confidentiality on intention to apply (see Figure 1). The finding of the relationship between procedural fairness with both attitudes and behavioral intentions is consistent with Konovsky and Cropanzano ( 1991) results, but extends their analysis to look at antecedents as well as consequences of procedural fairness judgments in the context of drug testing.

The potential generalizability of these results is limited with respect to actual applicants and to applicants for different jobs. Despite the effort to include a sample with characteristics more proximal to those of actual applicants, there remains the question of how reproducible these results would be in an actual employment context. Also, a very specific job description (Management Accountant I) was used in this study; all of the participants were either accounting majors or currently enrolled in an accounting class.

FIGURE 1. Empirical model including attitude toward the organization as a mediator of the effects of fairness perceptions on intention to apply. Unstandardized regression coefficients appear in parentheses.

Recent research by Murphy, Thornton, and Prue ( 1991) showed that acceptability of employee drug testing is related to actual job content, with the perception of threats to personal and public safety associated with performance of one's job being the single best -- but not the only -- predictor of drug-testing program acceptability. The use of a specific job and a sample of participants for whom the job was specifically relevant should have attentuated any contaminating effects of job content on participant responses. On the other hand, this may also have circumscribed the generalizability of these findings to different jobs. Finally, the generalizability of these results and much of the attitudes toward drug-testing literature is circumscribed because of almost exclusive reliance on samples from the United States.

We justified our arguments for Hypotheses 1 and 3 in part by referring to two different mechanisms of trust discussed by Barber ( 1983): reliable role performance and fiduciary trust. We believe that the finding of a direct relationship here between confidentiality and intention to apply indicates that potential applicants would base their judgments of intent to apply in part on a belief that, in a market setting, they are free to exercise their power to refuse the application opportunity if they find some aspect of the application process objectionable (here, a drugtesting procedure with inadequate confidentiality provisions).

To the extent that potential applicants accept the rules of the market mechanism for governing exchange relationships and base their judgment of intention on such rules, we expected that there would be a direct relationship between confidentiality and intent to apply. To the extent that potential applicants use relational indicators to gauge their trust in an employer, basing their expectations of treatment on considerations of morality and fairness, we predicted that the effects of confidentiality on intention to apply would be mediated by procedural fairness. Our results are consistent with both interpretations of confidentiality effects. Our finding of a direct relationship is consistent with Barber ( 1983) notion of reliable role performance. Our finding of a fairness-mediated effect is consistent with Barber's notion of fiduciary trust and with Tyler and Lind ( 1992) relational model of procedural fairness.However, further research is needed before we can conclude that these two trust mechanisms accounted for our results. In particular, a version of the Crant and Bateman ( 1989) model should be tested with reliable role performance specified as a mediating variable between confidentiality and intent to apply and with fiduciary trust specified as a mediator between confidentiality and procedural fairness.Some research indicates that, among such diverse groups as college students ( Murphy et al., 1990; Thombs & Scaffa, 1990), human resource managers ( Masters , Ferris, & Ratcliff, 1988), and the public at large ( Latessa, Travis, & Cullen, 1988), there is support for drug testing. However, our results show that those most directly affected by drug-testing programs (i.e., job applicants) are often less supportive of drug testing. These findings suggest that the administrators of drug-testing programs ought to pay particular attention to their procedures for protecting the confidentiality of the test results, because our results indicate that confidentiality influences potential applicants' intentions.

Received June 28, 1994