+ All Categories
Home > Documents > Dipboye.structured and Unstructured Ints

Dipboye.structured and Unstructured Ints

Date post: 03-Jun-2018
Category:
Upload: fahad-mushtaq
View: 229 times
Download: 0 times
Share this document with a friend

of 45

Transcript
  • 8/12/2019 Dipboye.structured and Unstructured Ints

    1/45

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    2/45

    80 ROBERT L DIPBOYEINTRO UOION

    Achieving a good fit between people and their jobs is a primary objectiveof human resources management HRM). Recruitment, training,performance appraisal, and selection are the chief strategies, and in each case,a structured process has been proposed to achieve an optimal person-jobfit. The usual recommendation is to start with a formal job analysis todetermine the knowledge, skills, and abilities required of incumbents andthe criteria for measuring their performance. On the basis of this analysis,as well as a rational consideration of costs and benefits, techniques are chosenfor implementation that are considered to be best suited for the situation.A formal evaluation follows, and those techniques that achieve the desiredoutcomes are retained while those that fail are modified or discarded.Although structuring HRM practices has been shown to have immense valueto organizations e.g., McDaniel, Whetzel, Schmidt, Maurer, in press ,the irony is that so many organizations fail to use a structured process andinstead rely on intuitive, highly subjective approaches. More attention needsto be given to understanding why some HRM methods that have been shownto be effective are largely ignored, while methods that have never beenscientifically evaluated or that have been thoroughly discredited areenthusiastically embraced.

    In considering this issue, the present paper focuses on perhaps the mostnotable example of how HRM practice and the empirical research diverge:the ubiquitous use of unstructured interviews in employee staffing. Thispaper starts with a comparison of structured and unstructured approachesto selection and then shows how the typical unstructured interview isvulnerable to a variety of biases in information gathering, judgment, anddecision making. Evidence that structured procedures can enhancereliability and validity is next considered, but as already mentioned,organizations have been reluctant to adqpt these procedures. The remainderof the paper examines the possible re11sons that unstructured interviewscontinue to be the dominant method qf selection, promotion, andplacement, despite the evidence that favors struct\)red approaches. In theinterest of parsimony, I will refer to selection, even though the discussionapplies to placement and promotion ~ well.) My central argument is thatHRM needs to broaden its focus to take into account not only whether thebest person is chosen for the job, but also whether other important functionsare fulfilled, such as achieving a good person-job context fit, acquiringpower maintaining justice socializing new hires conveying the v lues ofthe organization, and satisfying the needs of those who implement the HRMprocedures.

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    3/45

    Structured and Unstructured election nterviewsUNSTRUCTURED ND STRUCTURED

    PPRO CHES TO SELECTION

    8

    The present discussion views the sequence o events involved in selectionplacement and promotion decisions as a process that involves severalsources of information on applicants e.g., applications, reference checks,tests), with the interview as a core component of this process AmericanSociety of Personnel Administration, 983; Blocklyn, 988; Bureau ofNational Affairs, 1988; Shackleton Newell, 1991). Indicative of theimportance of the interview is that the information generated by otherprocedures often influences the final decision only after it has been filteredthrough interviewer judgments. Two general strategies in handling theinterview can be distinguished and are outlined in Figure : the structuredapproach which is formal and research guided, and the unstructuredapproach which is informal and guided by intuition. The comparison inFigure I is not meant to imply that all selection, placement, and promotionactivities fall into one of these two categories. Rather, these two approachesare presented as prototypes to aid in understanding, much as comparisonshave been made by decision theorists between intuitive and rationaljudgment processes e.g., Hammond Hamm, Grassian, Pearson, 1987).

    The Unstructured Approach to SelectionMost selection processes in organizations are probably semistructured inthat they possess elements of both the structured and unstructuredapproaches, but I would argue that the typical interview more closelyresembles the unstructured than the structured process in Figure IInterviewers are guided by their personal, often idiosyncratic views of what

    is required in the job. The gathering of information and the subsequentjudgment and choice of applicants are highly subjective and influenced byvague impressions of overall fit. Finally, the evaluations of how wellinterviewers have performed in choosing applicants is based on casualobservations rather than systematic investigation. Although typical of theselection process, the research on interviews suggests that unstructuredprocedures are vulnerable to a variety of biases that can lower the qualityof decisions. I now review this research, as well as recent evidence thatstructuring the process is a way to eliminate these biases and improvedecision making.General, Undifferentiated Theories o the Ideal pplicant

    With a structured approach to selection, decisionmakers start withdetailed information on the knowledge, abilities, and skills required in the

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    4/45

    8

    STRUCTUREDINTERVIEW

    Formal job AnalysisIdentifying KSAs

    Standardized SearchFor InformationDiagnostic of KSAs

    Quantitatively EvaluateSeparate KS s

    Choice Based onRational Considerationof Costs and Benefits

    Empirical Validationof judgments AgainstJob Criteria

    --

    f

    f

    FE TURES COMMONTO SELECTION PROCESS

    ecide What Traits -re Required in the jobather Information on

    Applicant s Traits and 1Qualifications

    udge Applicant onTrait Dimensions.

    ake SelectionDecisions

    tEvaluate InterviewProcess on PerlormanceIn Selecting Applicants

    ROBERT L DIPBOYE

    UNSTRUCTUREDINTERVIEW

    Personal Beliefsof the Interviewer

    UnstandardizedFor InformationBiased by Impressions

    Categorize OnGeneral Traits

    Noncompensatoryand Intuitive Judgmentof General Fit

    Subjective, CasualObservation ofHiree Performance

    figur 1 A summary of the major components involved in the structuredand unstructured approaches to selection interviews.

    job. Information is gathered and evaluated in reference to this profile ofrequirements. On the other hand, decisionmakers in the unstructuredprocess (Figure I) rely on their beliefs about the attributes of the qualifiedapplicant. Unlike the detailed information on knowledge, skills, andabilities that would come from a formal job analysis, these beliefs are moregeneral and diffuse, and are often peculiar to the particular decisionmaker.The research on occupational stereotypes has shown that people holdstrong beliefs about the attributes of individuals in occupations (CrowtherMore, 1972; More Suchner, 1976; Triandis, 1959a, 1959b; WestbrookMolla, 1976). Moreover, the conceptions of the ide l applicant appear

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    5/45

    Structured and Unstructured election nterviews 83diffuse and undifferentiated (Hake Schuh, 1971; Shaw, 1972 . Indeed,a study of over 5,000 interviewers revealed a good applicant" profile (e.g.,cooperative, trustworthy, dependable) that was common across diverseoccupations and that bore a striking resemblance to the profile of a goodboy scout (Hake Schuh, 1971). Experienced interviewers do not appearto differ markedly from inexperienced interviewers in their descriptions ofeither the ideal or typical applicant (Hake , Hollmann, Dunnette, 1970;Imada, Fletcher, Dalessio, 1980; Marks Webb, 1969; PaunonenJackson, 1987). I would hypothesize that the lack of informationcharacteristic o unstructured procedures leads interviewers to categorizeon the basis of these prior conceptions and, as a consequence, lowers thevalidity of their decisions.There are some caveats that should be mentioned in posing the abovehypothesis. There is some evidence that beliefs about the personality traits ofa typic l incumbent (as opposed to an ide l incumbent) are reasonably accurate(Jackson, Peacock, Holden, 1982 . Also, experts generally have morecomplex, reliable, and accessible knowledge structures than nonexperts (FiskeKinder, 1981; Kozlowski Kirsch, 1987; Lurigio Carroll, 1985; SmitherReilly, 1987 ; so it is possible that expert interviewers will not show as largea decline in validity as nonexpert interviewers. Despite these caveats, it seemsunlikely that the conceptions held by an interviewer, even an experienced one,can match the accuracy and specificity of a well-

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    6/45

    84 RO ERT L DIP OYEOne type of bias that can intrude into an unstructured procedure s thetendency for interviewers' opinions of the applicant to leak into theirnonverbal and paralinguistic behavior (Dougherty, Turban, Callender, 1992;Farina Feiner, 1973; Matarazzo Wiens, 1972; Word, Zanna, Cooper,

    1974). Impressions can also influence the time spent with the applicant(Anderson, 1960; Phillips Dipboye, 1989; Tullar, 1989). Perhaps the mostimportant aspect of the interviewer's conduct that is subject to bias s thequestioning of the applicant. Interviewers appear less likely to ask questionsif they are favorably impressed with the applicant than if they have negativeimpressions (Sydiaha, 1961; Tengler Jablin, 1983). Also, a tendency towardan inverted funnel in the sequencing of questions has been shown in whichinterviewers ask closed questions at the beginning of the session, followed byopen-ended questions, possibly as an attempt to test stereotypes of the idealapplicant (Tengler Jablin, 1983). Snyder (1984) found some evidence of aconfirmatory bias in questioning in which interviewers asked questions thatwere slanted in the direction of prior hypotheses about the personality of theinterviewee. Subsequent research has shown that initial impressions can biasthe phrasing of the questions asked of applicants, although not necessarily ina confirmatory direction (Binning, Goldstein, Garcia, Scattaregia, 1988;Eccher, Shes key, Levalle, Binning, 1988; Macan Dipboye, 1988; Radefeld,Williams, Binning, 1990).The linkages between interviewer opinions of the applicant and the conductof the session remain unexamined for the most part, but the evidence gatheredso far is fairly clear in showing that how well the applicant performs in thesession reflects to some degree the interviewer s own behavior. Generallyapplicants appear to perform better in the session and provide moreinformation when interviewers show support and positive regard in theirnonverbal and paralinguistic behavior (Keenan, 1976; Matarazo, Wiens,Saslow, 1965), ask about job-related matters as opposed to personal matters(Matarazzo, Wiens, Jackson, Manaugh, 1970), ask open-ended questions(Tengler J ablin, 1983), and ask about information not in the application(Daniels Otis, 1950). There is also some evidence that the interviewer sconduct of the session can influence the quality as well as the quantity of theinformation gathered. Martin and Nagao (1989) compared an interviewer whoacted warm with an interviewer who acted cold in their nonverbal behaviorduring the session. Although fewer of the interviewees with a warm intervieweroverreported their SATs and grades, these same interviewees tended to showa larger m gnitude of overreporting than those who had a cold interviewer.In another experiment, Dipboye, Fontenelle, and amer (1984) found thatallowing interviewers to form preinterview impressions led to the generationof more information, but also led to more variability in the conduct of thesessions and consequently lowered the reliability and accuracy of finaljudgments.

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    7/45

    Structured and Unstructured election nterviews 85In summary, the interviewer's behavior can influence the quality of theinformation that is gathered, but the specific mechanisms by which this effectoccurs needs further explication. Elsewhere, I have proposed three alternativeprocesses that can emerge in the gathering of information--confirmatory,

    disconfirmatory, and diagnostic and the possible moderators of theseprocesses (Dipboye, 1992). Although much remains to be known about thistopic, I would propose that biases such as those discussed here are more likelyto creep into the conduct of the session, and lessen the diagnosticity of theinformation gathered, when the interview procedures are unstructured.Categorical and iased udgments

    Throughout the interview process, interviewers are involved in judging thecharacteristics of the applicant and the fit of these characteristics to therequirements of the job. In the absence of clear information, interviewers relyon their own personal and possibly erroneous conceptions of the job and theapplicant. Cognitive categorization theory provides a framework for describingthis process (Rowe, 1984). For instance, an interviewer might believe that allor most philosophy majors lack common sense, are manipulative, and possessliberal political attitudes. Upon discovering that an applicant is a psychologymajor, the interviewer categorizes that person, assigning other traits believedto be characteristic of philosophy majors (i.e., no common sense manipulative,and liberal). Moreover, the interviewer judges the applicant on the general fitof the individual to the conception of the ideal applicant. The specific categoriesthat influence the perception of the applicant and the job can be activated ina serendipitous manner by a variety of factors, including the mood of theinterviewer (Isen Baron, 1991) and the particular mix of applicants(Cleveland, Festa, Montgomery, 1988; Heilman, 1980).An implication of cognitive categorization theory is that the manner in whichinterviewers judge applicants can occur automatically without much self-insight into the underlying reasons (Feldman, 1981). A theory such as that ofFiske and Neuberg (1990) would suggest that unless interviewers encounterdisconfirming information, their search for and processing of information isdone in an unthinking manner that has the effect of confirming their initialcategorizations. Perhaps suggestive of this automaticity, when interviewers areasked why they evaluated applicants as they did, they often appear to graspfor an explanation. The most common response is that the applicant does ordoes not provide a good fit. Caplow and McGee (1960) found this in theirclassic study of faculty hiring in universities. One of their sources commentedthat among the most important considerations was whether the candidate fitthe department and was someone they could live with: We take a good lookat their letters and then when they're down here we look at them and talk tothem and then we take a good look into our crystal ball and pull out the best

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    8/45

    86 RO ERT L DIP OYEman. In other words, we're completely subjective about the whole thing (p.123). More recently, Judge and Ferris (1992) reported that numerous interviewswith college recruiters about what they look for in a qualified candidaterevealed an amazing convergence across decisionmakers on a statement thatgoes something like this: 'I can't articulate it, but I1l know it when I see it(p. 3).A variety of rating effects have been found that possibly reflect thetendency to categorize applicants on the basis of the vague andundifferentiated conceptions of the ideal applicant. These effects include halo(Dipboye, Gaugler, Hayes, 1990; Hake , 1971; James, Campbell,Lovegrove, 1984; Knicki Lockwood, 1985; Kinicki, Lockwood, Hom,Griffeth, 1990; Shahani, Dipboye, Gehrlein, 1992), negativity (Rowe,1989), primacy (Farr, 1973; arr York, 1975; Johns, 1975), and contrast(Burkhardt, Weider-Hatfield, Hocking, 1985; Cesare, Dalessio,Tannenbaum, 1988). Numerous studies in both the laboratory and the fieldhave shown similar-to-me effects in which the interviewer gives morefavorable evaluations to the extent that the applicant is similar to theinterviewer o background characteristics education attitudes and otherfactors (Graves Powell, 1988; Milstein, Burrow, Wilkinson, Kessen,1976; Orpen, 1984). Related to these effects, personal liking for the applicanthas been found to bias judgments of qualifications (Anderson Shackleton,1990; Graves Powell, 1988; Keenan, 1977; Raza Carpenter, 1987). Thehighly subjective nature of interviewer evaluations also is revealed by theeffects that the applicant's styl of self-presentation can have on interviewerjudgments. Applicants are viewed as more qualified if they show positive,responsive verbal and nonverbal behavior, including enthusiasm, warmth,good eye-contact, smiling, head nodding, voice modulation, energy, handgestures, and vocal expressiveness (e.g., Anderson Shackleton, 1990;Baybrook, 1985; Forbes Jackson, 1980). Moreover, style is usually foundto be more important than objective information on the applicant (GilmoreFerris, l989a; Kinicki Lockwood, 1985; Parsons Liden, 1984;Rasmussen, 1984; Sharf, 1970; Tessler Sushelsky, 1978).Further evidence of the categorical nature of interviewer judgments comesfrom research showing that interviewers are influenced by the race, disability,gender, and age of the applicant. Professional interviewers as well asinexperienced interviewers can be influenced by these factors (Barr Hitt,1986; Hitt Barr, 1989; Turner, Fix, Struyk, 1991; Van Vianen Willemsen,1992), although questions remain as to the consistency and strength of theeffects (Dipboye, 1992; Stone, Stone, Dipboye, 1992). The evidence is muchclearer with regard to another irrelevant factor, the physical attractiveness ofthe applicant. The applicant's facial attractiveness has been shown to influencethe interviewer's selection decisions at all phases of the selection process(Carlson, 1967; Carroll, 1966; Kinicki Lockwood, 1985; Raza Carpenter,

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    9/45

    tructuredand Unstructured election lntetviews 871987; Rynes Gerhart, 1990; Shahani, Dipboye Gehrlein, 1993; Springbett,1958). In one of the more recent demonstrations of the effects of attractiveness,Rynes and Gerhart (1990) obtained ratings of the physical attractiveness ofMBA applicants and then examined the correlations of these ratings withrecruiters' evaluations of the applicants. Attractiveness was more importantas a correlate of recruiter evaluations of firm-specific fit than were objectivecharacteristics of the applicants such s GPA sex business experience majorand extracurricular activities. n addition to these field studies, numerousexperiments have shown an attractiveness bias for both organizationaldecisionmakers and college students (e.g., Morrow, 1990; Morrow, McElroy,Stamper, Wilson, 1990; Morrow McElroy, 1984; Rothblum, Miller,Garbutt, 1988).Finally, research has shown that interviewers are subject to a variety ofattributional biases (Herriot, 1989). Perhaps the most important of these isthe tendency of interviewers to underestimate the influence of their ownconduct of the session on the applicant, and to overestimate the importanceof the applicant's traits. The interviewer's prior expectations can moderate thiseffect Belec Rowe, 1983; Phillips Dipboye, 1989; Tucker Rowe, 1979).Internal factors are seen as more important causes of the applicant's behaviorwhen the behavior of the applicant is consistent with the interviewer's initialimpressions. External factors are seen as more important when the behaviorof the applicant violates these expectations.

    n summary, interviewers appear prone to a variety of judgmental biases.These biases appear to reflect an automatic, unthinking categorization in whichthe applicant is judged as fitting or failing to fit a diffuse and undifferentiatedconception of the ideal applicant. Also, there is reason to believe that they aremore likely to occur in the context of an unstructured than a structuredprocedure (Dipboye Gaugler, 1993).iases in Decision MakingAfter gathering information on applicants and judging them against therequirements of the position, a decisionmaker must eventually decide on whom

    to hire promote or place. n an unstructured process decisionmakers areunlikely to base their final choices of applicants on a careful analysis of benefitsand costs associated with each applicant. That decision making in organizationsdeviates from a rational model has been thoroughly documented in theorganizational literature (Cohen, March, Olsen, 1972; Lindblom, 1959;Mintzberg, Raisinghani, Theoret, 1976; Weick, 1979), and there is littlereason to believe that decisionmakers in the interview are n exception.Beach's (1990) image theory provides a framework for describingnonrational decision making in the unstructured interview and is a rich sourceof ideas for future research. Beach noted that most decisions are made quickly

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    10/45

    88 ROBERT L DIPBOYEand simply, on the basis of 'fittingness,' and only in particular circumstancesare they made on the basis of anything like the weighing and balancing of gainsand losses that is prescribed by classical decision theory " p. xiii). Decisionmaking is guided by cognitive structures called value images that determinethe rightness or wrongness of a decision. The decisionmaker frames the decisionby calling upon those images that seem most relevant to the situation, andthen uses two tests in choosing from among alternatives. Compatibility testinginvolves assessing whether the option is incompatible with the framed image.f an option exceeds the threshold for rejection, then the option is rejected.This is a noncompensatory decision, in that compliance with some facets of

    the image cannot compensate for violations of other aspects of the image.Moreover the decisionmakers are often not even conscious of what they aredoing. Instead, they act in a "rapid, smooth process that can be called "intuitive"in that it requires minimal cognitive processing and control" (Mitchell Beach,1990, p. 14). f a single candidate survives the compatibility testing, then thisis the choice and no further deliberation occurs. f more than one candidatesurvives, then the decisionmaker engages in profitability testing, which sconscious and deliberate. The purpose of this process is to pick the option thatprovides the most benefits relative to the plans, goals, and principles of thedecisionmakerImage theory has yet to be tested in the context of interviewer decisionmaking, but there is some evidence that supports the notion that interviewersmake decisions rapidly in a noncompensatory fashion. Perhaps the mostfrequently cited of these findings are those of Springbett (1954), who foundthat interviewers decided on applicants after only four minutes into the session.Although subsequent research has failed to replicate the "four-minute-todecision" finding (Buckley Eder, 1988; Hake , 1982), the evidence suggeststhat interviewers make their decision well before the end of the interview(Tschirigi Huegli, 1979; Tullar, Mullins, Caldwell, 1979; Tucker Rowe,1977). The specific time interviewers decide is perhaps not as important as thetendency of interviewers to short-circuit the process and reach a decision beforeall the data have been gathered. This finding seems to reflect a process similarto compatibility testing, but the research has yet to be conducted that appliesimage theory to decision making in the interview.The Failure to Learn From xperience

    In the final step in the process in Figure I, interviewers evaluate theirdecisions by tracking their hits and misses, and then adjusting their selectionpractices on the basis of the results. In most situations, decisionmakers do nothave the information that would allow them to check on how well they haveperformed in picking winners and losers. Moreover, they may tend toremember judgments that were correct, while forgetting their errors. As a

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    11/45

    Structured and Unstructured election nterviews 9consequence those making selection decisions f il t le rn from experienceGarb, 1989). This problem has its source in both the structure of the decisiontask and the basic cognitive limitations of the judge Einhorn Hogarth, 1978).Decision tasks are often structured so that people never receive adequatefeedback on their judgments. This is especially true of selection interviews,where feedback is almost always limited to those who were hired and is usuallyanecdotal. Take, as an example, a decisionmaker who is convinced that malecollege graduates from small private liberal arts colleges are the best candidates,and who then only hires those who fit this profile. Obviously, the decisionmakerin this instance would be unable to test his theory because he would only knowabout those who fit the profile and were hired, and would know nothing aboutthe nonliberal arts majors from large state universities who were rejected.Decisionmakers could evaluate their intuitive judgments following the sameprocedures contained in structured process, but this almost never happensbecause they lack the necessary information.The failure to learn from experience also results from the cognitivelimitations of decisionmakers and their bad judgment habits Faust, 1984).There is evidence that people in general, and interviewers in particular, areprone to seek information that confirms their hypotheses and unable orunwilling to use disconfirming information Snyder, 1984). Where theinterviewer later supervises or works with the applicant, the impressions formedin the interview may even carry over to the workplace in a manner similarto the escalating commitments that have been observed in performanceappraisal Bazerman, Beekun, Schoorman, 1982; Schoorman, 1988). Relatedto confirmatory biases are illusory correlations. In the first demonstration ofthis bias, Chapman and Chapman 1969) found a systematic tendency tooverestimate the frequency of occurrence of word pairs if the words werenaturally associated e.g., hat and head or knife and fork), or if both wordsin a pair were distinctive i.e., they were atypically long). In a similar fashion,interviewers in n organization might perceive correlations that do not existbetween personal characteristics of employees and their success in theorganization. f an interviewer observes tall employees achieving outstandingsuccess, the observed co-occurrence of these two relatively distinctive eventscould lead interviewers to perceive a correlation of height to success when thereis none.Further compounding these errors, interviewers are prone to hindsight biaswhen reflecting on past decisions in that they believe that what has alreadyhappened was inevitable Fischhoff, 1975). Imagine an interviewer whoobserves a short applicant who is hired and then fails. Subsequently, a ruleis derived from this experience i.e., short people will fail on the job) that isthen applied in future interviews. This bias reflects a tendency to reconstructpast events so that they make more sense, even if the reconstruction bears littlerelation to reality. Overconfidence is a final tendency that is both a consequence

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    12/45

    90 ROBERT L DIPBOYEand an antecedent of the aforementioned biases. While the accuracy ofprediction declines with increases in the redundancy and amount ofinformation available to those making the predictions (Kahneman Tversky,1973; Oskamp, 1962, 1965), the confidence of these individuals in their ownpredictions tends to increaseBecause of the structure of the selection task as well as cognitive limitationsand habits, interviewers do not have the information that they need to fullyevaluate how well they have performed in selecting employees. The perversityof the highly unstructured situation is that increased experience with theseprocedures may only lead to stronger beliefs in the validity of one's judgments.As aptly stated by Dawes 1991, p. 256), The basic problem with experienceis that you have no idea what would have happened i you or someone elsehad done what you or they didn't do.

    Structured Approaches to Selection Produce Higher Quality DecisionsIn contrast to the unstructured process depicted in Figure I, structuredselection is rational in that it uses the scientific method and containsaccuracy checks at every step. The first step is to conduct job analyses todetermine the knowledge, skills, and abilities required to perform the tasksin the position. On the basis of the analysis, interview questions and ratingscales are crafted to measure these KSAs. The actual conduct of theinterview is highly standardized so that each applicant is asked the same

    questions. The choice among candidates is a rational decision based onanticipated worth to the organization of each candidate's predictedperformance (Cascio, 1991). Finally, and perhaps most important researchis conducted to evaluate the validity of the predictors used in selection.f a criterion-related strategy is used, predictors are validated againstmeasures of what formal job analyses have shown to be the critical jobcriteria

    As already indicated, much of the research showing the biases injudgment and decision making in the interview has used unstructuredprocedures. In the typical study, few formal guidelines for gathering andevaluating information were provided to those in the interviewer role.Moreover, they were often given global rating scales that were unrelatedto specific job requirements. There has been surprisingly little researchexplicitly testing this hypothesis, but, as already mentioned, the biases andintuitive processes described earlier seem likely to flourish to the detrimentof the quality of interviewer judgments in unstructured interviews(Dipboye Gaugler, 1993). Structuring the selection process shouldreduce the magnitude of these biases, thereby enhancing the reliability andvalidity of interviewer judgments.

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    13/45

    Structured and Unstructured Selection nterviews 91Meta Analytic Support for th Use of Structured Interviews

    Support for the contention that structuring improves the quality ofinterviewer judgments comes from several recent meta-analyses. Wright,Lichtenfels, and Pursell 1989) reviewed six unpublished and seven publishedvalidations of structured interviews and found a mean observed validitycoefficient of .29 and a mean corrected validity of .39. In a much more extensivereview, Wiesner and Cranshaw 1988) reviewed a total of 150 validitiesinvolving a total sample size of 51,459. The mean observed validity was .26.Correcting for criterion unreliability and range restriction yielded a validityof .47. A comparison of structured and unstructured interviews revealed muchhigher validities for the former than the latter. Unstructured individualinterviews had a mean corrected validity of .20, whereas structured individualinterviews had a mean corrected validity of .63. Unstructured board interviewshad a mean corrected validity of .37, whereas the structured board interviewshad a mean corrected validity of .60.The most comprehensive of the meta-analyses is that of McDaniel et al. inpress). Their analysis included 106 validity studies that used job performanceas the criterion and involved a total N of 13,011. A mean observed validityof .23 was obtained, and after correcting for criterion unreliability and rangerestriction, this coefficient increased to .45. A further subgrouping of thesevalidities revealed that structured interviews in comparison to unstructuredinterviews had higher observed .24 vs 21) and corrected .47 vs . .40) validities.Because the unstructured interviews used in this study were probably morestructured than the typical unstructured interview, McDaniels et al. in press)suggested that their meta-analysis may have actually underestimated thedifferences in validity.The complexity of the issues involved in using interviews is greater than forother selection techniques such as mental ability tests or biodata and consequently,there are ambiguities in the interpretation of the results of these meta-analyses.A particularly important issue is the variation among individual interviewers inthe criterion-related validity of their judgments that has been observed withunstructured procedures Dipboye et al., 1990; Dougherty, Ebert, Callender,1986; Kinicki et al., 1990). These studies show that, even in the context of anunstructured procedure, some individual interviewers are capable of highly validjudgments whereas others are abysmally low in the validity of their judgments.Despite the questions that remain, one conclusion of these meta-analyses is quiteclear: structured interviews yield better results than unstructured interviews.Characteristics of the Structured Interview

    What are the attributes that account for the higher validities of structuredinterviews? Answering this question is difficult, given the differences that exist

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    14/45

    92 ROBERT L DIPBOYEamong so-called structured procedures. There are several formats that havebeen described as structured, including the patterned behavior descriptioninterview (Janz, 1982), the situational interview (Latham, Saari, Pursell,Campion, 1980), the multi-modal employment interview (Schuler Funke,1989), as well as other varieties developed for use in specific firms or industriesArvey, Miller, Gould, Burch, 1987; Brown, 1980; Landy, 1976; Yonge,1956). The prototypic example is Campion et al. s 1988) highly structuredinterview technique. The interview format that they propose is intended toeliminate the subjectivity of the typical interview procedure and contains thefollowing basic attributes.First, the interview questions are constructed based on the knowledge,skills, and abilities found through formal job analyses to be required in thejob. A variety of question types are permitted, including questions about jobknowledge, how the applicant would handle specific situations, anddemographic information such as past experience and education. However,only questions that pertain to KSAs found to be critical prerequisites areallowed. A second feature is that exactly the same questions are asked ofall candidates, with no variations allowed. Although questions can berepeated, interviewers cannot ask follow-ups or prompts A thirdch r cteristic is th t interviewers r te the nswers to the questions on sc lesthat behaviorally anchor the good, marginal, and poor answers. Theseanchors can be developed either through having experts provide actualanswers they have heard in interviews that exemplify each point on the scale,or through having persons familiar with the job brainstorm potentialanswers. A fourth attribute is that it involves a. panel of interviewers whoare familiar with the job requirements.

    The fifth, and perhaps most important, attribute of the highly structuredinterview technique is that the panel conducts the interviews and evaluatesapplicants consistently across all applicants. There are several safeguards toensure consistency. The panel members are trained prior to the session on howto use the procedures and to avoid bias, and then independently rate theapplicant s answers. Also, information on applicants is withheld frominterviewers prior to the sessions, the same panel is used across all candidates,and discussion is delayed until all interviews have been conducted. The ratingsof panel members are averaged, and only in the case of large discrepanciesis there a discussion of the applicants. Candidates are not allowed to askquestions during the interviews, but are provided this opportunity in a separate,nonevaluative session. Finally, throughout the process, special attention isgiven to ensuring compliance with legal guidelines and professional standardsof testing as set forth in the EEOC s Uniform Guidelines 1978) on employeeselection and the validation principles of the Society for Industrial andOrganizational Psychology (1987).

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    15/45

    Structured an d Unstructured election Interviews 93easons that Structured nterviews Might mprove the uality of judgmentsBecause of variations among the various structured formats, it is difficultto pinpoint the exact source of the higher validities and reliabilities (DipboyeGaugler, 1993). Nevertheless, there are at least three categories of featuresthat serve as possible explanations.One likely source of advantage is that structured interviewing proceduresare based on formal job analyses and are consequently more job-related thanunstructured procedures. In support of this contention, Wiesner and Cranshaw1988) found that higher validities were associated with procedures that werebased on formal job analyses than those based on informal or no job analyses.Formal job analysis may have these benefits because it leads to better samplingof behavior that is part of the criterion domain (Wernimont Campbell, 1968).

    For instance if customer rel tions is identified as crucial dimension oeffective performance then this provides the basis for asking questions abouthow the applicant would interact with and relate to customers. Still anotherreason that job analysis may benefit the selection process is that it providesdecisionmakers with a common perspective on the knowledge, skills, andabilities required in the job. Previous research has shown that such informationon specifications can improve reliability (Langdale Weitz, 1973), reduce theeffects of primacy (Peters Terborg, 1975) and irrelevant information (WienerSchneiderman, 1974), and increase the discrimination between qualified andunqualified applicants (Osburn, Timmereck, Bigby, 1981).A second potential advantage of the various structured formats is the mannerin which these formats deal with ancillary data, such as test scores, biographicaldata, reference checks, and school transcripts. Interviewers usually haveancillary information, and their final impressions of an applicant can beinfluenced to some extent by these data (Dipboye, 1989; McDaniel et a ., inpress). Structured interviews either do not allow interviewers to previewancillary information Campion et al., 1988), or else provide for a morestructured preview of this information than found in the typical unstructuredinterview (Brown, 1979).A key feature in structured interviews that possibly accounts for theirreliability and validity is that they standardize the process by which informationis gathered on an applicant by holding interviewers to the same line ofquestioning. In the case of highly structured interviews (Campion et al., 1988)and situational interviews Latham, 1989), exactly the same questions are askedin the same order with no follow-ups or probes. Other structured procedures,such as the patterned behavior description interview Janz, 1982), the selectioninterview blueprint (Brown, 1979), and the multimodal employment interview(Schuler Funke, 1989), hold the interviewer to the same pattern ofquestioning but allow some degree of freedom in pursuing independent linesof questioning.

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    16/45

    94 RO ERT L DIP OYEStructured interviews incorporate a variety of procedures in the judgmentphase of the process that have been shown in previous research to enhancethe accuracy and reliability of judgment. These procedures typically encourageinterviewers to delay their evaluation of the applicant until after the session,

    thus separating information gathering from the final integration and evaluationof information. Well-defined rating scales (e.g., behaviorally anchored scales)are provided Maas, 1965; Vance, Kuhnert, Farr, 1978), rather than the lesswell-defined graphic and trait scales so often found in unstructured interviews.Ratings are often decomposed, meaning that the interviewer rates theapplicant's qualifications on separate dimensions, in contrast to the holisticratings of the applicant's overall qualifications so often found in theunstructured approach. Note-taking is permitted and encouraged in somestructured procedures (Janz, 1982; Mayfield, Brown, Hamstra, 1980), thuspossibly reducing errors in recall (Macan Dipboye, 1986; Schuh, 1980). Theend result of the standardization and job-relatedness of these variousprocedures is that the interviewer has a firmer basis for inferring dispositionsfrom the statements and behavior of the applicant (Herriot, 1989).Finally, in forming a final judgment of the applicant's qualifications,structured procedures statistically combine the information gathered on theapplicant, usually through simple averaging or summation. In contrast, theglobal evaluations formed with unstructured procedures allow interviewers tocombine their impressions of applicants using intuitive and idiosyncraticcombinations. Numerous studies have compared the accuracy of predictionusing intuitive combinations of information with the accuracy achieved withstatistical combination of the same data. In the typical study, judges are giventhe same set of information and predict an event (e.g., predict bankruptcy frominformation in a loan application). A comparison is made of the accuracy ofthe judges' intuitive forecasting with the accuracy achieved through statisticalprediction, usually from a linear model. Decades of this type of research hasshown that statistical combinations almost always outperform predictionsbased on judges' subjective combinations of the same information (Faust, 1984;Sawyer, 1966). Meehl 1986) has observed that There is no controversy insocial science that has shown such a large body of qualitatively diverse studiescoming out so uniformly in the same direction as this one (p. 374).he ifficulties o Implementing Structured Selection ProceduresMany questions remain as to why structuring the selection process mightlead to higher quality decisions (Dipboye, 1992; Dipboye Gaugler, 1993),but the evidence is compelling that the inclusion of formal job analyses,standardized questioning of applicants, and behavioral rating scales improvethe validity and reliability of interviewer judgments (Dipboye, 1992).Regardless of the evidence, however, unstructured interviews remain the

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    17/45

    Structured and Unstructured election nterviews 95method of choice in organizations. Structured interviewing appear to havegained more acceptance in recent years (Solomon, 1989), but the frequencyof its use pales in comparison to the overwhelming amount of unstructuredinterviewing that is still used in the initial selection, placement, and promotionof employees. Take, for example, academic settings. I am unaware of universitydepartments that have implemented anything remotely resembling a structuredselection process in the hiring of faculty. Indeed, a faculty member who hasthe audacity to suggest such an approach is likely to be hooted down. Becauseof affirmative action and EEOC pressures, more employers are usingqu ntit tive ev lu tion sheets nd are more sensitive to questions th t areillegal and unethical, but relatively few organizations appear to base theirprocedures on a formal job analysis, hold interviewers to the same line ofquestioning, impose behavioral rating scales, or use decomposed ratingprocedures. Probably the least likely aspect of the structured process to beimplemented is the statistical combination of data on applicants. Even whereobjective measures are included (e.g., such as aptitude tests, personalitymeasures, and quantifiable biographical data), the final decision to hire is mostfrequently based on an intuitive combination of these data.

    Nat only are structured procedures less frequently used, I would furtherhypothesize that once implemented, they become progressively unstructuredover time. It is important to remember that most of the validation researchon the interview consists of one-shot studies evaluating relatively short-termeffects, and only rarely are there replications in which validity is evaluated atseveral points in time. Hulin, Henry, and Noon (1990) have found thatcriterion-related validities of tests show a systematic and large decline over time.They attribute these decrements to changes in the actual abilities of theindividuals as they acquire proficiency on their tasks. I suspect that if followup studies were conducted with interviews, one might find even largerdecrements in validity. In testing this supposition one would not only needto examine validity as a function of the time between the collection ofinterviewer evaluations and the criteria, but also as a function of the time thestructured interview h s been in use

    n predicting a large decline in validity of interviewer evaluations, I amproposing that when an interviewer's judgment of the applicant's abilities isthe predictor, there are not only changes in the abilities of those hired, as shownby Hulin eta . (1990), but also changes in the predictor itself. As time passes,a process of destructuring occurs in which interviewers stray from the originalguidelines in the direction of more intuitive, less structured procedures. Oneof the few studies providing evidence of this tendency was reported by Lathamand Saari 1984) in a validation of the situational interview. They found avalidity coefficient that was much lower than typically found r .14) withthis type of interview. Closer examination revealed that the interviewers haddeviated from the structured procedures which dictated rigid adherence to a

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    18/45

    96 ROBERT L DIPBOYEpattern of questioning and instead, used the questions as a guide to formingan overall impression. The authors suggested that in the future researchersshould be present when the data are being collected (p. 573). I suspect thatthe destructuring observed in this study was not an anomaly but a frequentoccurrence. Researchers obviously cannot continue to monitor the int rvi windefinitely and interviewers eventually revert t more informal intuitivpractices.The reluctance of employment interviewers to use structured informationgathering and decision procedures appears to reflect a more general resistanceto structured approaches by decisionmakers. The reluctance of physicians touse statistical equations or computer-assisted diagnoses is well documented(Komaroff, 1982; Schwartz, 1979). In research with managers, Isenberg(l984)found that even those who had been trained to use decision aids seldom did,and if such structured procedures went against their intuition, they rejectedthese aids. In a survey of clinical psychologists, Wade and Baker (1977) foundthat only 8.6% were concerned with reliability or validity of their assessmentpractices. Moreover, 81 of the respondents in this same survey indicated thatthey used personalized procedures in evaluating the results of tests used intheir practice, and a mere 18.5% reported the use of standardizedinterpretations. he results of another survey show that psychologists involvedin individual assessment in industry show an overwhelming preference for aninterview (93.8%), with only 15.2% saying that they follow a strictly structuredprocedure and only 39 stating that they conduct a formal job analysis priorto their assessments (Ryan Sackett, 1987).

    Evidence of destructuring also comes from decision-making research that hasrevealed what appears to be an irresistible urge to go beyond simple decisionrules, even when these rules clearly aid the accuracy of decisions. Twoexperiments conducted by Arkes, Dawes, and Christensen (1986) provideinteresting demonstrations of this. In the first experiment, undergraduatespredicted whether each of 40 students made the honor roll or not on the basisof grades in three randomly selected courses. They were given a rule to followin which they were to predict honors when the student had 2 or 3 A's and nothonors when A's numbered 0 or l The subjects were told that following thisrule would allow a 70 hit rate. One group was encouraged to try and beatthis accuracy rate, whereas another group was warned to stick with the rule.The group that was explicitly warned to not deviate from the 70 rule did betterthan the other groups, especially when they were given no incentives. Also,subjects who were paid incentives for the accuracy of their judgments tendedto go beyond the decision rules they had been given and subsequently mademore errors than those who were not paid. In a second study, subjects weregiven items of information and predicted whether a baseball player would bechosen the most valuable player in the National League on the basis of thisinformation. hey were told that by choosing the player whose team finished

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    19/45

    Structured and Unstructured election lnte iews 97highest in the standings th t they would be correct in bout 75 of their choices.Although both high nd moderate knowledge groups fell short of the level ofperformance they could have achieved if they h d conformed to the rule, themoderate knowledge subjects actually outperformed the high knowledge group.

    Although the Arkes et al. ( 1986) research did not directly address selectiondecisions, I would suggest th t similar effects occur as interviewers embellishand add to structured procedures. I would go a step further to suggest thatit may be virtually impossible to induce decisionmakers to adhere rigidly toan optimal strategy, especially if they are experts and re positively impressedwith their own perspicacity.

    FORCES THAT WORK G INST STRUCTURED INTERVIEWSWhen attempts have been m de to explain why decisionmakers fail to followa structured process the most common explanations have een in terms ocognitive biases. A frequent suggestion is to use training to debias users(Fischhoff, 1982). The reasons for reliance on unstructured procedures rundeeper, however, than the cognitive biases that have been the focus of muchof the discussion in recent years. Figure 2 depicts the centrifugal forces t workin the organization th t can dr w decisionmakers toward the unstructured.The most immediate pressures come from the needs of the persons who mustactually implement the process. A second force is the desire to broaden the

    Cultural Assumptionsand Values

    Politics of the OrganizationNorms for Procedural and

    Distributive JusticeAttempts to Provide a

    Good Person Context Fit

    Personal Needs ofthe Interviewer

    Structured lnverview.Procedures

    figur 2 Organizational forces that work against structured interviewprocedures.

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    20/45

    98 ROBERT L DIPBOYEscope of selection to provide a better fit of the hiree to the context of the job.A third consists of the norms for fairness and justice that exist in the subunitor the organization as a whole. Political pressures are another force that canact against structured procedures while sustaining unstructured processes.Finally, unstructured procedures are used because they communicate thecultural values associated with the organization or subunit.

    The Personal Needs of the Decision makerInterviewers must not only know how to properly use a structured processbut they must also have sufficient motivation to use the procedures as theywere intended to be used. A reason that unstructured procedures persist while

    structured procedures are resisted is that the former can better fulfill thepersonal needs of those who must implement them. The greater potential ofunstructured procedures to allow personal gratification derives from thefreedom that these procedures allow in expressing preferences, the basic taskcharacteristics involved the lower cognitive costs associated with intuitivejudgment, and the self-image and personality traits of the interviewer.

    Perhaps the major reason that unstructured procedures are preferred is thatdecisionmakers can express their idiosyncratic and particularistic leanings toa greater extent with this format. They can show favoritism toward those whoare similar, physically attractive, and likable. Bigots can let their prejudicesrun rampant, while the more socially conscious can give advantages to thedisadvantaged. Although often unfortunate and unfair, unstructured selectionprocedures are perhaps more enjoyable because they provide the opportunityto vent these preferences.

    A second source of personal satisfaction are the task characteristics involvedin selection. Hackman and Oldham (1980) in their Job Characteristics Modelpredict that jobs having autonomy, skill variety, identity, feedback, and tasksignificance are more likely to be satisfying than jobs that lack these samedimensions. To some extent, an unstructured procedure is higher on all of thesecore characteristics. Unstructured interviews are high on autonomy in that theyallow the decisionmaker discretion in the topics that are discussed, the contentand order in which questions are asked, the criteria that are used in the finalevaluations nd the pace at which e ch interview is conducted. For instancean interviewer might choose to spend most of the time on the academic recordwith one applicant while concentrating on the work record with anotherapplicant. This is not optimal in making valid, reliable evaluations, but it maybe more gratifying to the interviewer who feels a sense of control over theprocess. The unstructured interview allows a variety of skills to be applied tothe task that may include report writing, conversational skills, recruiting?bservation, and above all skills in integrating information to form a holisticJUdgment of the applicant s qualifications. Again, the opportunity to express

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    21/45

    Structured and Unstructured election nterviews 99these skills is not always conducive to making good decisions, but the diversityof activities involved makes the typical unstructured interview less boring, lessroutine, and more challenging. Because of the two-way communicationallowed, unstructured interviewing appears to contain more feedback frominterviewee to interviewer. Unstructured interviews could be seen s havingmore task significance relative t structured procedures that in some casesrelegate the interviewer to the role of observer and recorder. To the extentthat unstructured procedures involve the interviewer in the entire selectionprocess (writing the ad, recruiting potential applicants, interviewing, evaluatingthe information, deciding among applicants, negotiating salary), the processalso contains more task identity.Not only are unstructured procedures more likely to gratify important needsof the decisionmaker, but there is also some degree of intrinsic satisfactionassociated with intuitive decision processes. This possibly reflects the tendencyof persons to take the path involving the least cognitive effort. According toimage theory, compatibility testing requires limited cognitive capacity, is oftenunconscious, and feels effortless (Mitchell Beach, 1990). As long ascompatibility testing proceeds without an alternative passing the threshold forrejection, the choice of alternatives is accompanied by positive feelings anda sense of conviction. n contrast, when the decisionmakers are forced to paycloser attention to the alternatives, as would be the case in a highly structuredprocedure, the process is more effortful and more likely to arouse negativeaffect.All of the above factors suggest that interviewers will tend to findunstructured procedures more personally satisfying, but there are likely to beindividual differences that moderate these preferences. The job characteristicsmodel of Hackman and Oldham 1980) posits that the core job characteristicsinduce high levels of job satisfaction mainly for those who are high in growthneeds. f their reasoning can be extended to the interviewing task, theninterviewers who are high on growth needs should be most satisfied withunstructured interviews and least satisfied with structured interviews.Interviewers who are experiencing a high degree of role conflict and who placelittle importance on the interviewing role seem likely to prefer the structuredprocedures. Holland's (1985) occupational classification scheme suggestspersonality differences that may moderate these preferences. Interviewers whoare social types gain satisfaction from social interaction and should preferunstructured interviews, whereas those who are conventional types are drawnto record keeping, filing, and other clerical activities and should preferstructured interviews. The self-image of the interviewer is another potentialmoderator (Meehl, 1986, p. 374). Interviewers who view themselves as highlyperceptive and skilled in reading people seem more likely to prefer theunstructured interview than those who view themselves more modestly on thisdimension. Predictions also can be drawn from the widely used, albeit

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    22/45

    100 RO ERT L DIP OYEcontroversial, Myers-Briggs Type Indicator. Interviewers who are high on theintuitive and feeling scales should show the strongest preference forunstructured procedures, whereas those on sensing and judging should showthe least preference.

    Providing a Good Fit of the Applicant to the job ContextThe various motives that possibly influence the use of unstructured

    interviews are too numerous to fully enumerate but are crucial determinantsof whether structured interview procedures are used or avoided and whetherthey stay structured. As important as they may be, however, we need to gobeyond the individual level to consider additional forces at the level of he groupand the organization that pull decisionmakers toward unstructured procedures.One of these is the pressure on decisionmakers to achieve a good fit betweenhirees and the ontext of the job.A common criticism of structured selection procedures is that they focusnarrowly on the requirements of the job itself and ignore the requirements ofthe various contexts of the job, such as being a team player or a goodorganizational citizen. Several researchers have suggested broadening thepredictor and criterion domains to include the context (Bowen, Ledford,Nathan, 1991; Borman Motowidlo, 1993; Campbell, 1990). Consistent withthese suggestions, recent research has demonstrated the positive outcomesassociated with a good fit between the job-context and the person's goals(Vancouver Schmitt, 1991) and values (Chatman, 1989; O'Reilly, Chatman,

    Caldwell, 1991). According to Schneider's (1987) Attraction-SelectionAttrition (ASA) model, fit is achieved through a cycle that involves selection,recruiting, and socialization. I now show how the unstructured interview mightcontribute to each of these phases of the process.In selecting for fit to the context, values, personality traits, and goals areparticularly important (Borman Motowidlo, 1993), and it has been proposedthat interviews are the best means of assessing these attributes. Dobbins, Cardy,

    and Carson (1991) suggested that ._. the traditional employment interview,much maligned when using typical validation strategies, may in fact be foundto be more appropriate as a predictor of fit within the organization (p. 18).Similarly, Chatman (1991) noted that the reason interviews continue to be usedis that they fulfill a primary function of selection processes by allowing theorganization t assess how well a person s values fit the organization's valuesand norms (p. 343). Recently Judge and Ferris (1992) stated that calls forstructured interviews as a way to improve the validity of the interview maybe misplaced if the true goal, and utility, of the interview lies not in selectingthe most technically qualified, but the individual most likely to fit into theorganization (p. 23). They recommend using tests and other objectiveprocedures to evaluate fit of the applicant to the KSAs of the job while using

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    23/45

    Structured and Unstructured election nterviews 101the interview to determine fit of the applicant to the goals and values of theorganization. One could go a step further to suggest that in assessing contextualfit unstructured interviews are preferred over structured interviews. Theprimary reason s that unstructured interview procedures allow the interviewerto go beyond KSAs and explore attributes that are ignored in structuredapproaches, but are crucial to predicting contextual performances.Just as an unstructured procedure might allow the organization to pick theapplicant that best fits the job context, the same type of procedure also mightallow the applicant to make better choices. A recent experiment suggests thatapplicants can make choices among organizations that provide a fit of theorganization to their personalities Bretz, Ash, Dreher, 1989). However, noneof the research conducted so far has evaluated the hypothesis that unstructuredinterviews allow applicants to make better decisions than structured interviews.Still, there are several reasons that this might be the case. With an unstructuredprocedure, the applicant is able to ask questions about whether there is a matchto his or her abilities, interests, goals, values, and needs. The interviewer canprovide realistic previews of the job and the context. Bargaining andnegotiation can even occur nd as a consequence the interviewer and applicantmay accommodate to some of the demands of the other in return forconcessions. n contrast, the highly structured interview provides limitedopportunity for the applicant to gather information or influence the conditionsof employment.To provide a good fit to the context, the right types of persons not onlymust be selected but they also must be attracted to join the organization. Thebelief that structured selection procedures interfere with the recruitment ofapplicants, while unstructured procedures facilitate recruitment, providesanother reason for the continued use of unstructured interviews. Perhaps themost consistent finding in the research on recruiting is that applicants are morefavorably disposed to interviewers who are attentive, warm, thoughtful, sociallyperceptive, and likable in their conduct of the session Campion, 1980; HarnThornton, 1985; Harris Fink, 1987; Keenan, 1978; Schmitt Coyle, 1976;Young Heneman, 1986). On the basis of these findings, one could surmisethat unstructured interviews are more effective in recruiting because theinterviewer can add a more personal touch to the process, and can better conveyattributes such as warmth and empathy. Also, the flexibility of the unstructuredprocedure allows the interviewer to shift away from assessment when theapplicant is highly qualified and concentrate instead on selling the job andorganization.The evidence gathered so far on the relative impact of structured andunstructured interviews on the success in recruiting applicants has yieldedmixed findings. In support of structured procedures, Taylor and Bergmann1987) found that the more structured that recruiters perceived their interviewsto be, the more likely applicants were to say they would accept a job offer.

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    24/45

    102 ROBERT L DIPBOYEAlso, there is some evidence that applicants respond favorably to recruitersif they ask job-related questions (Alderfer McCord, 1970; Taylor Sniezek,1984). Not surprisingly, the most negative reactions occur in response to shoot-from-the hip questions about inappropriate topics such as marital plans andchildrearing (Craig Greenberg, 1986; Rynes, 1993), conduct that should beless likely with structured interviews.There is also evidence to support the use ofunstructured procedures. One studyfound that students perceived a particular type of structured interview to beboring, fakable, and, possibly irrelevant (Rynes, 1993). In perhaps the only directcomparison of structured and unstructured interviews, Latham and Finnegan(1987) found that college students rated the unstructured interview as more likelyto favorably influence them to accept a job offer than either a patterned orsituational interview. The research of Martin and Nagao (1989) suggests that thereaction to structured procedures may depend on the position for which theinterviewee is applying. Undergraduate students role played an applicant foreither a desk clerk or a management trainee. The subjects were interviewed bymeans of a paper-and-pencil form, computer, or a face-to-face session with aninterviewer who conducted the session in either a cold or warm manner. Subjectswho interviewed for a high status job (manager) felt more resentful toward theprocedures when the session was conducted by means of a paper-and-pencilinstrument or a computer than when it was face-to-face. This greater resentmentwas expressed v n when the interviewer in the face to face session cted in acold manner. The actual effectiveness of interviewing procedures in attractingpotential employees is an issue that remains unresolved. The fear that structuredprocedures may actually tum off prospective employees seems prevalent anddeserves much more attention than it has received so far.

    A primary means of ensuring that individuals fit the job context issocialization, which consists of those actions taken by the organization toinfluence its members to adopt the behaviors, attitudes, and values appropriateto their roles (Fisher, 1986; Van Maanen Schein, 1979). Selection proceduresare rarely discussed as tactics of socialization, but as the first significantencounter with the organization, the selection process can be part of theunfreezing of the potential new hire, in which the individual is shaken loosefrom previous attitudes, values, and norms possibly through stressfulexperiences such as ridicule and intimidation. The interview and other selectionprocedures can also motivate applicants by signaling that they will have adifficult task ahead of them if they are hired, or by leading them to believethat membership in the organization is reserved for the fortunate few andshould be highly valued (Trice, Belasco, Alutto, 1969). Deal and Kennedy(1982, p. 12) alluded to these functions when they described interviews atTandem Corporation as a type of 'inquisition' in which applicants were calledback several times. The message conveyed to prospective employees is 'we takelonger, and take care of people we hire-because we really care.'

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    25/45

    Structured and Unstructured Selection nterviews 103Also important to the socialization of a new hire is the information thatan applicant can acquire as interviewers spin tales about their experiences inthe organization and engage in other chit-chat. The perceptive applicant canlearn a variety of lessons from these conversations, such as who is subservient

    to whom, how employees should dress, what topics are open to discussion andwhich are off limits, who are the favored employees, and who is in the outgroup. Thus, unstructured procedures allow the organization to beginsocializing the potential new hire even before the selection decision has beenmade. Although selection is unlikely to have as large an impact as laterexperiences in the organization (Chatman, 1991), the selection process can setthe stage for subsequent attempts to socialize the new hire.In summary, unstructured interviews can be part of the overall effort toprovide a good fit between the hiree and the context of the job. Theseprocedures can provide the basis for selecting, recruiting, and socializing hireesso that they conform to what is considered to be right type. Whetherunstructured procedures are actually more effective in this regard will needto be examined in future research. The more important consideration for thepurpose of this paper is that they are perceived by organizational membersto be important to attaining a good person-context fit, thus providing one more

    reason that they are preferred over structured procedures.Maintaining Procedural and istributive justice

    An additional force that needs to be taken into account in explaining theprevalence of unstructured procedures is the fairness of the selectionprocedures. Generally, selection techniques are more likely to be used if theyare consistent with the distributive rules for allocating outcomes (i.e., allocatingjobs to applicants), and the procedural rules for making the decisions. Althoughstructured procedures actually may be fairer in most circumstances, I amproposing that decisionmakers frequently use unstructured interviews in theattempt to ensure both distributive and procedural justice.

    From a distributive justice perspective, one could predict thatdecisionmakers will judge the fairness of a procedure on the basis of howoutcomes are distributed. There are at least three rules that might be used(Deutsch, 1975 . n equity rule would lead decisionmakers to view a fairselection decision as one in which the most qualified candidates are hired(Adams, 1965 . f he distributive rule was based on need, then job offers wouldgo to those who need the job the most. For instance, preference might be givento applicants with dependents because they have a greater need for money,or to women and minorities because they have suffered past discrimination.Finally, an equality norm would dictate that no distinctions be made and thatall candidates have an equal chance of being hired, perhaps as the result ofa random draw. Research on the effects of these alternative rules shows that

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    26/45

    104 RO ERT L DIP OYEequity is the dominant rule, but that the other two can be salient under somecircumstances (Cohen, 1987; Bierhoff, Cohen, Greenberg, 1986). Theadvantage of unstructured interviews in maintaining distributive justice is thatthey allow the decisionmaker the flexibility to implement whichever rulehappens to fit the dominant norms in the situation. Moreover, unstructuredprocedures allow interviewers to distort their perceptions of the selectionprocess so that decisions that actually deviate from the dominant rule are seenas complying with this rule.From a procedural justice perspective, the distribution of outcomes is notas important as the fairness of the procedures leading to the distribution(Folger, 1977). Research has suggested several procedural rules (FolgerGreenberg, 1985; Tyler Bies, 1990). Decisions are perceived as fairer ifthe participants are: (I) allowed to voice their opinions, 2) their opinionsare not muted, (3) once opinions are asked, they actually influence thedecisionmaker, ( 4) the decisionmakers are neutral and suppress their biases,5) decision making criteria are consistently applied across employees, 6)timely feedback is provided after the decision, and 7) procedures areexplained and justified.More important than these rules, according to Tyler and Bies 1990), is thequality with which decisionmakers interact with others in implementingprocedures and allocating outcomes. The importance of the interactionalquality as a determinant of perceived fairness was demonstrated in a studyconducted by Bies and reported in Tyler and Bies (1990, pp. 81-82). MBAstudents recalled past interviews in which they felt fairly or unfairly treated.Selection procedures were seen as fairer when the MBAs had the opportunityto present their credentials and qualifications to the interviewers, were treatedwith honesty, courtesy and respect, were given timely feedback on the results,and no illegal or unethical questions were asked regarding their race, sex, ormarital status. ost important, Bies s results suggested that higher quality ofinterpersonal treatment in the interview led to perceptions of fairness even whenthe student had been turned down for employment. Bies and Shapiro 1988)conducted a follow-up to this research in which they manipulated some of thefactors identified in the earlier survey of MBAs. They were assigned to oneof four conditions which differed on whether the interview procedure hadallowed voice, and whether justification was provided for the final decision.In the mute condition, subjects were told that the interviewer had not askedthe candidate any questions concerning the resume, and that the candidate wasgiven no opportunity to ask the interviewer about the job and the company.In the voice-procedure condition, subjects were told that the candidate hadan opportunity to ask questions about the job and the company, and that therehad been an opportunity for the applicant to present his credentials. Subjectsperceived a decision to not hire as fairer when voice was allowed than whenvoice was not allowed (mute condition).

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    27/45

    Structured and Unstructured election nterviews 105f interviewers attempt to be fair to applicants by incorporating a higherquality of interaction, as suggested by Bies and colleagues, then the effect maybe to destructure the selection process. Possible testimony to the success ofthese efforts to be fair is the very low frequency of discrimination lawsuits in

    which the interview has been the focus of the complaint. Indeed, Campion andArvey 1989) found that the interview was n issue in fewer than I% of themore than 8,000 cases reported between 1979-1987. Although these statisticssuggest that unstructured interviews are unlikely to provoke a suit, it isimportant to note th t once a lawsuit is filed, structured interviews appear easierto defend in court. Gollub nd Campion (1991), in n analysis of91 FederalDistrict Court cases, found th t the decision was more likely to be in favorof the employer if the interview was job related, contained specific behavioralcriteria, and was standardized. Other features that enhanced defensibility werethe use of a panel, a formal employment decision system, systematiccombination of ratings, and record keeping.In summary, unstructured procedures are a two-edge sword. By allowingdiscretion in the questions asked, there is greater risk that the interviewer willengage in improper behavior. I believe, however, that as long as the interviewerdoes not engage in such behavior n unstructured interview will be seen asfairer as the result of having more open and informal communication. Thisperceived fairness could be still another reason that unstructured interviewsare more frequently used th n structured procedures.

    Acquiring and Maintaining PowerAlthough it seems at odds with pressures to be fair in the selection process,

    politics are a reality of organizational life. Indeed, politics offer a possibleexplanation for why the criterion-related validities of the interview are muchlower when they are used to select police th n when they are used in otheroccupations (McDaniel et al., in press). The strong pressures from competingconstituencies and internal power struggles th t can occur in police selectionmay lower the reliability nd validity of interviewer judgments in this type ofsituation.Despite the damage done to the validity of judgments, the attractiveopportunities that unstructured interviews offer for wielding power is a primaryreason that they are preferred over structured procedures. Pfeffer (198la) hasdefined power as the ability to achieve desired outcomes and politics as theactivities taken within organizations to acquire, develop, and use power (p.7). In his theory, political activity arises from scarcity, interdependence,heterogeneous goals, and heterogeneous beliefs bout the means of c h i e v ~ n ggoals. Among the ways decisionmakers acquire power is by reducmguncertainty, controlling the decision process, and building coalitions (Pfeffer,198lb). Moreover, skilled politicians seek situations that are sufficiently

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    28/45

    106 RO ERT L DIP OYEambiguous that their actions cannot be closely scrutinized and monitored.From the perspective of this theoretical framework, unstructured interviewsoffer several advantages to decisionmakers seeking power and influence.Power is acquired by reducing uncertainty, and there are few situations thathave as much uncertainty associated with them as the selection of a newemployee (Graanovetter, 1974; Kanter, 1977; Pfeffer, 198la). A lack ofagreement regarding the requirements of the position or doubts about theapplicant s qualifications present opportunities to the power seeker. Althoughit seems paradoxical, interviewers may prefer the looseness associated withunstructured interviews because it allows them to manage the uncertainty thatsurrounds selection decisions. With unstructured procedures they have moreopportunity to persuade others as to who should and should not be hired thanwith the tight guidelines of a structured interview. Unstructured proceduresalso allow both interviewer and applicant to engage more freely in impressionmanagement tactics. A skilled self-presentation by applicants can convey theirqualifications (Baron, 1989; Gilmore Ferris, 1989b), and as a consequence,can also help interviewers to reduce the uncertainty surrounding the hiringdecision. Unstructured procedures can allow the interviewer to engage inimpression management as well, thus enabling them to convince others of thequalifications of their preferred candidates and their own ability to judge thesequalifications.

    An additional advantage of the unstructured interview is the control itprovides over the decision process. he unstructured interview provides anopportunity for the interviewer to control the alternatives, and the informationon these alternatives. Also interviewers can stress those criteria that f vor theirpreferences while de-emphasizing or ignoring the criteria that do not providethem an advantage. Salancik and Pfeffer 1976) demonstrated how this mightoccur in an experiment in which subjects chose from among applicants, andthen wrote justifications for their decisions. They found that when asked tojustify their choices, the judges stressed the socially-sanctioned criteria ofgrades, experience, and recommendations. In no case did the judges mentionthe similarity of the applicant, although similarity had, in fact, stronglyinfluenced their evaluations. Manipulating the criteria to justify decisions islikely to be difficult in a structured procedure where standards are explicit andmuch harder to bend to the decisionmaker s preferences. In contrast, theambiguity of an unstructured procedure allows the power seeker to uselanguage and symbols to rationalize their decisions and to hide their trueintentions.

    Unstructured procedures make it easier to use what can be the most blatantpower tactic, coalition building. Pfeffer (198la) has described several examples.In one case, executives at GM promoted obviously unqualified candidates toposts to win their loyalty and to build a power base. A second case involveda large retailing organization that tended to promote people on the basis of

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    29/45

    Structured and Unstructured election nterviews 107their retailing background despite their lack of expertise in the areas they wereto head. Given that the primary objective was to ensure that the company wouldmaintain the dominance of the retailing department, having experience in thearea being managed was a drawback, not an asset. Obviously, selectionprocedures that build on careful analysis of job requirements and that arestandardized would make such coalition tactics hard to accomplish withoutrevealing the underlying motivation of those using them.Finally, unstructured procedures provide political advantages because theymake it more difficult to evaluate decisionmakers. Regardless of which of thespecific power tactics is used, skillful politicians seek unstructured situationsbecause they are ambiguous and prevent close scrutiny and monitoring(Gilmore Ferris, 1989b). or instance, unstructured interviews can be seenas a way of avoiding inspection by the EEOC, affrrmative action officers, andpotential plaintiffs (Lancaster, 1975; Daniel, 1986). They also can play a rolein the internal politics of the organization as departments attempt to avoidmonitoring of their selection practices by those outside the department,especially the HRM department. There is research to suggest that HRMdepartments use standardized testing as a way of gaining power over otherunits in the organization (Cohen Pfeffer, 1986), but I would suggest thatthe other side of this struggle is the use of unstructured interviewing byoperating departments to retain authority over staffing decisions.In summary, the selection process is one of many contexts in whichindividuals and groups in an organization engage in opportunistic behavior.The ambiguity and lack of accountability present in unstructured proceduresmake it more attractive to those seeking to acquire and maintain power.Additionally, political pressures may lead to a destructuring of a structuredprocedure as interviewers and other decisionmakers shape the procedures totheir advantage.

    The Symbolic Functions o the InterviewThe last and most inclusive reason that unstructured procedures dominatethe selection process is that they are an important means of expressing andmaintaining the values of the group, subunit, and organization (Trice Beyer,

    1984, p. 665). I will start with Schein's (1985) definition of culture as deeperlevels of basic assumptions and beliefs that are shared by members of anorganization, that operate unconsciously, and that define in a basic 'taken-forgranted' fashion an organization's view of itself and its environment p. 6).The values of the organization define what ought to be and are a crucialcomponent of a culture. The selection process, as well as other organizationalpractices, can have the symbolic function of communicating these values, andin so doing, can channel the energy of participants and help maintain the system(Dandridge, Mitroff, Joyce, 1980). Symbols can be instrumental in nature,

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    30/45

    1 8 ROBERT L DIP8 YE

    in that they convey information about some logical aspect of selection, orprimarily expressive (e.g., myths, stories, cocktail parties), in that they dealwith the feelings nd emotional needs of participants (Daft, 1983).When organizations confront poorly understood problems th t requireintuition, symbols can provide a rich source of information that can be usedin understanding the situation (Daft, 1983). Along these lines, Pfeffer (l98lb)noted that symbols are important especially when the preferences foralternative outcomes from organizational actions are unclear or conflicting.In coping with an uncertain environment organizational symbols can helpparticipants find common ground in the interpretation of events can serve asa source o motivation and can reduce the anxiety associated with theuncertainty, Given that the factors that predict success in the organization areoften poorly understood, the selection process seems especially open to the useof symbols. By drawing from the analyses of Daft 1983) nd Pfeffer (l98lb),the prediction can be made that expressive symbols are more likely to be foundin the selection process the more uncertain organizational participants areabout the qualifications of the position. t follows that one could expect tofind more symbolism in the selection process used in filling higher levelmanagerial and professional positions th n for lower level positions. Forinstance, there is often uncertainty and anxiety surrounding the replacementof a high level person, such as a President or CEO. The ceremonial activitiesassociated with the search for such an individual can serve as a visible symbol,proclaiming to those inside and outside the organization th t attempts are beingmade to find a suitable replacement (Pfeffer, l98la).There are several vehicles for communicating symbols that are associatedwith unstructured selection processes. In some cases the selection processconsists of rites, which Trice nd Beyer 1984) defined as unified eventsconsolidating a number of cultural expressions in an elaborate nd dramaticmanner. In other cases they are ceremonials consisting of several rites connectedwith an occasion or event. Other cultural forms that can occur in the selectionprocess re the telling of sagas (i.e., historical narratives about theaccomplishments of the organization nd its leaders), myths (i.e., dramaticdescriptions of imagined events), stories (i.e., descriptions of true eventscontaiqing some fiction and some truth), legends (i.e., an embellishedrecounting of n extraordinary event in the history of the group ororganization), nd folktales (i.e., completely fictitious stories).With repeated use, an unstructured interview and other selection procedurescan become an integral part of the culture (Schein, 1985). In the beginning,selection techniques are implemented s a consequence of decisionmakersvalues, their beliefs about the nature of reality, and their beliefs about howto deal with this reality. Imagine, for instance, that the key people in anorganization value creativity and believe that applicants with unconventionalbackgrounds should be preferred over applicants with conventional

  • 8/12/2019 Dipboye.structured and Unstructured Ints

    31/45

    Structured and Unstructured election nterviews 109backgrounds. This value then leads to the selection of applicants who havetaken unusual routes in their careers, and the rejection of applicants with moreconventional records. f the persons selected on these bases succeed, then thisvalue evolves into a widely shared image that successful applicants areunconventional (Beach, 1990; Borman, 1987 . The image is part of the bodyof knowledge in the organization defining what is correct, and is so deeplyheld that it constitutes a basic assumption of the culture. This is particularlytru if a practice receives continual support in which case the assumptionsunderlying the practice may be taken for granted and alternative practices noteven considered. The cultural functions of organizational practices can be sostrong that they persist even though there is little evidence that they serve theirmanifest goals effectively. In this case, they become ritu ls that reduce anxietiesand convey values but do not fulfill their manifest purposes (Bolman & Deal,1991; Trice et al., 1969). According to Bolman and Deal (1991), tests andinterviews often produce data of doubtful validity, but may bolster the self-confidence of those who are hired and allow those who are not to feel thatthey were treated fairly (p. 265).The idea that interview procedures become empty symbols, devoid ofinstrumental value, is consistent with institutional theory ( D. According toMeyer and Rowan (1977), perhaps the leading advocates of IT, the formalstructures and practices of many organizations reflect the myths of theirinstitutional environments instead of the demands of their work activities (p.341 . From the perspective of this theory, selection procedures have ceremonialvalue in demonstrating to participants in the environment of the organization(e.g., the federal government, potential customers, stockholders, competitors)that legitimate techniques are being used in the hiring of employees. Therefore,unstructured interviews are more frequently used in the selection of managerial,professional, and other higher level employees because they are seen as morelegitimate than structured approaches by important constituents of theorganization. For example, a university department might avoid structuredselection procedures (e.g., weighted application blanks, personality testing) inhiring faculty because such procedures would look bad to the administrationand faculty of other universities.Once a procedure such as an unstructured interview becomes institution-alized, IT proposes that a decoupling occurs in which the interview is evaluatedsolely for its ceremonial value rather than on the basis of how it performs inachieving organizational goals. A primary use of myths and ceremonies thatmasquerade as rational selection practice is to decouple the actual performanceof the work from external evaluation. For instance, using a procedure thatis widely seen as legitimate, such as an unstructured interview, makes it moredifficult for the EEOC or potential litigants to scrutinize the selection practicesuf an organization. Thus, the unstructured int


Recommended