An Empirical Investigation of Interviewer-Related Factors That Discourage the Use of High

Structure Interviews
Author(s): Filip Lievens and Anneleen De Paepe
Source: Journal of Organizational Behavior, Vol. 25, No. 1 (Feb., 2004), pp. 29-46
Published by: Wiley
Stable URL: http://www.jstor.org/stable/4093642 .
Accessed: 10/12/2014 08:58
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Wiley is collaborating with JSTOR to digitize, preserve and extend access to Journal of Organizational
Behavior.

http://www.jstor.org

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

Journal of Organizational Behavior

J. Organiz. Behav. 25, 29-46 (2004)
Published online in Wiley InterScience (www.interscience.wiley.com).

DOI: 10.1002/job.246

An empirical investigation of

interviewer-related factors that discourage

the use of high structure interviews

FILIP LIEVENS1* AND ANNELEEN DE PAEPE2

and Workand Organizational
Psychology,GhentUniversity,
of PersonnelManagement
'Department
Ghent,Belgium
2Securex,
Ghent,Belgium

Summary

High structureinterviewsappearto be less frequentlyused in personnelmanagementpractice

thanmightbe expected given theirgood reliabilityand validity.Althoughseveralauthorshave
speculatedon the factors of resistanceto high structureinterviews,empiricalresearchis very
scarce. Two studies are conducted among experiencedhumanresourcesrepresentativeswho
frequentlyconductemploymentinterviews.The firststudyprovidesa fine-graineddescription
of the degree of structureused in interviews, showing that in most interviewsconstraintsare
placed only on the topical areas to be covered and that scoring is done only on multiple criteria. The second study tests various hypotheses regardinginterviewer-relatedfactors, which
may lead to lower levels of structurein interviews. Results show that when interviewersare
concerned about establishingan informalcontact with interviewees, want to have discretion
over interviewquestions, and want to develop interviews efficiently,they are less inclined to
use higherlevels of structurein interviews.Conversely,people who participatedin interviewing workshops and Conventionaltypes reportusing significantlyhigher levels of structure.
Implicationsfor improving interviewermotivationto use higher levels of structurein interviews are discussed. Copyright? 2004 John Wiley & Sons, Ltd.

Introduction
A remarkable finding in industrial and organizational psychology is that personnel practices which are
not supported by empirical research are often very popular whereas personnel practices which have
been shown to be effective are less frequently used (Dipboye, 1994; Johns, 1993). This inverse relationship between psychometric soundness and popularity is probably most strikingly exemplified by
the attractiveness of low structure interviews. In fact, meta-analytic research has demonstrated that
low structure interviews are considerably worse than high structure interviews in terms of reliability
(Conway, Jako, & Goodman, 1995) and criterion-related validity (Huffcutt & Arthur, 1994; Marchese

* Correspondence
to: FilipLievens,Department
of PersonnelManagement
andWorkandOrganizational
Psychology,Ghent
HenriDunantlaan
2, 9000 Ghent,Belgium.E-mail:filip.lievens@rug.ac.be
University,

Copyright? 2004 John Wiley & Sons, Ltd.

Received 17 June 2002

Revised 5 September2003
Accepted 19 September2003

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

30

F. LIEVENSANDA. DE PAEPE

& Muchinsky,1993; McDaniel, Whetzel, Schmidt, & Maurer,1994; Wiesner& Cronshaw,1988). In

addition,low structureinterviewshave been foundto producehigheradverseimpact (Huffcutt& Roth,
1997) and to lead to less favorablelitigation outcomes (Williamsonet al., 1997) than theirhigh structure counterparts.
Despite this scientificevidence, low structureinterviewscontinueto be frequentlyused in organizations. Moreover,the usage of high structureinterviews is lower than might be expected given their
psychometricsoundness.Accordingto a recent survey (Ryan, McFarland,Baron,& Page, 1999), only
34.6 per cent of U.S. firmsreportedusing the same fixed list of job-relatedquestionsper candidatefor a
specific job. Another survey indicated that more interviewersdisagreed than agreed with common
structuredinterviewing practices (Dipboye & Jackson, 1999). In addition to these surveys, there is
furtherevidence that, even when organizationsadopthigh structureinterviews, these interviewstend
to become less structuredover time (Latham& Saari, 1984).
The divergence between interview practice and researchillustratesthat the adoptionand effective
implementationof high structureinterviewsis not only influencedby psychometriccriteriabut also by
various individual, political, and cultural factors, which to date are poorly understood (Campion,
Palmer, & Campion, 1997; Dipboye, 1994, 1997; Harris & Eder, 1999; Rynes, Barber, & Varma,
2000; Van der Zee, Bakker,& Bakker,2002). For example, Rynes et al. (2000) recently posited that
'because of the importance of interviewer motivation to effective interviewing,
developing a better
....
research
the
the
interviewer's
is
one
priorities for improving
understanding of
highest
of
perspective
interview practice' (p. 270).

This paperbegins to fill this gap in the literature.Two studies are conducted. The first descriptive
study aims to gather informationregardingthe degree of structureused in interviews. An important
extension to previous surveys that treatedstructuredinterviewusage as a unidimensionaland dichotomous variable(e.g., Harris,Dworkin,& Park, 1990; Ryan et al., 1999) is that we relied upon the twodimensionaltaxonomyof Huffcuttand Arthur(1994) to obtaina fine-graineddescriptionof the degree
of structureused in interviews.
The second study aims to examine a range of individual(interviewer-related)factors, which may
reducethe use of high structureinterviews.We focus on these interviewer-relatedfactorsbecause these
factors are generally thoughtto be the firstand most immediategroupof sources of resistanceto high
structureinterviews (Dipboye, 1994, 1997). Recently, Van der Zee et al. (2002) provided empirical
evidence that individual interviewers' attitudes and beliefs towards high structureinterviews were
relatedto theirintentionsto use high structureinterviews.However,they did not investigatethe effects
of specific interviewer-relatedbeliefs. A deeper understandingof these interviewer-relatedobjections
that slow down the more frequentuse of high structureinterviews may have direct implications for
personnelselection practice.This is because, at a practicallevel, an in-depthinsight into these factors
may eventually suggest strategies for improving interviewermotivationand for promotingusage of
high structureinterviews (Harris& Eder, 1999; Rynes et al., 2000).

Context

Societal Context
Belgium is situatedin the heartof Europe.Therefore,many multinationalsand internationalorganizations (NATO,EU) have headquartersin Belgium. In Belgium, three languagesare spoken. The
Flemish partof the countryspeaks Dutch, whereasthe Walloonpartof the country speaks French.
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERSAND HIGH STRUCTUREINTERVIEWS

31

There is also a small part of Belgium in which Germanis spoken. This study was conducted in
privatefirms in the Flemish part of Belgium. Approximately70 per cent of the labor force works
in privatecompanies, whereas 30 per cent of the labor force are employed in the public sector.
Legal Context
The fact that three languages are spoken in Belgium necessitates that there is a strongtraditionof
dialogue, consultation,and compromise. This traditionof social consultationis also prevalentin
employment-relatedareas.This means thatefforts areundertakento strivefor compromisebetween
different partners (e.g., employer, employee, union) without much direct outside intervention.
When conflicts arise, the court is used only as a last resort. The fact that people seldom refer to
the court in selection-relatedareas is also due to the fact that there are few laws in Belgium that
regulate selection practices. For example, there exist no regulationsabout how employmentinterviews should be conductedor who should conduct employmentinterviews.
Organizational Context
In Belgian firms, HR representatives typically develop employment interviews themselves.
This means that in Belgium (as comparedto the situation in other countries) HR representatives
have considerable latitude to influence the level of structurein the employment interviews they
conduct.
It should also be noted that in recent years consultancy firms have begun to certify their own
personnel in order to ensure and promote the quality of their efforts. For the same reason, many
consultancyfirmsin the Flemish partof Belgium have stronglyadvocatedthe use of high structured
interviewingpracticesas a betteralternativeto low structuredinterviews.This was evidencedby the
publicationof practicalbooks and the organizationof trainingworkshopson this subject. Yet, this
does not mean that high structuredhave been widely adopted in the Flemish part of Belgium.
According to recent surveys, high structuredinterviews are used in about 30 per cent of the companies in the Flemish part of Belgium.

Study 1
Background
As alreadynoted, previous surveys about structuredinterview usage (Harriset al., 1990; Ryan et al.,
1999) typically treated use of structuredinterviews as a unidimensionaland dichotomous variable.
This is not congruent with recent conceptualizations.First of all, use of structureis better seen as
a multidimensionalvariable because there are various ways to structurean employment interview
(Campion et al., 1997; Dipboye & Gaugler, 1993; Huffcutt & Arthur, 1994). Recently, Campion
et al. (1997) defined structurein interviews very broadly as 'any enhancementof the interview that
is intended to increase psychometricproperties by increasing standardizationor otherwise assisting
the interviewer in determiningwhat questions to ask or how to evaluate responses' (p. 656) and
distinguished 15 ways to structure an interview. According to Huffcutt and Arthur (1994), two
structuredinterviewing practices are especially importantat an operationallevel, namely interview
question standardizationand response scoring standardization.Huffcutt and Arthur(1994) crossed
these two structuredinterviewingpracticesto obtain a two-dimensionalmodel of interview structure.
Therefore,it is importantthat a survey about structuredinterview usage measures at least these two
dimensions.
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

32

F. LIEVENSANDA. DE PAEPE

Second, as reflected in the aforementioneddefinition, structuredenotes standardizationand therefore is not a dichotomy (structuredvs. unstructuredinterviews). Instead, it is better to use the term
'degree of structure'(low structurevs. high structureinterviews) and to conceptualize structurein
employmentinterviewsas a continuousvariablewith variouslevels. In their model of interview structure, Huffcuttand Arthur(1994) identifiedfour progressivelyhigher levels of interview standardization and three progressively higher levels of scoring standardization. In terms of question
standardization,Level 1 consisted of interviews without any constraintson the questions to be asked
(i.e., the typical unstructuredinterview). The next levels consisted of interviews in which constraints
are placed on the topics to be covered (Level 2) or in which all main and follow-up questionsare determined in advance (Level 3). Finally, the highest level of question standardization(Level 4) contained
interviewswhereinexactly the same questionswith no choice or follow-up are asked. In termsof scoring standardization,the following three progressivelyhigher levels were distinguished:a global evaluation of candidates(Level 1), an evaluationof candidatesalong multiple establishedcriteria(Level
2), and an evaluationof each individualresponse of candidatesaccordingto pre-establishedanswers
(Level 3).
In short,recentconceptualizationshave treatedinterview structureas a continuousand multidimensional variable.Therefore,our firstdescriptivestudy will use the two-dimensionalmodel of interview
structureof HuffcuttandArthur(1994) to obtaina more fine-graineddescriptionof the degreeof structure used in interviews.

Method
Sample and procedure
This study was conducted in the Flemish part of Belgium. On the basis of data provided by the
National Institute of Statistics, there are 1249 large private firms (with more than 200 employees)
in the Flemish partof Belgium. A representative(in termsof geographicalregion and industry)sample
of 94 firmswas drawnfrom the availablelist of large privatefirms.After obtainingcontact addresses,
we made telephonecalls to the humanresources(HR) managementdepartmentof each of these firms.
Because some of the HR representativeseitherreportedthatthey outsourcedemploymentinterviewing
to consultancy firms or refused to participate,the final sample consisted of 76 participants(response
rate= 81 per cent). All of these HR representativesactually conducted employment interviews.
Measure
To measure interview structure,we asked two open-ended questions derived from Huffcutt and
Arthur's(1994) two-dimensionalmodel. The firstquestion ('Would you share with me how you ask
questions to candidatesin the interview?') dealt with the level of interview question standardization.
The second open-ended question ('Would you share with me how you rate candidatesin the interview?') assessed the level of response scoring standardization.If unclear,respondents'answerswere
probedby asking for furtherinformationor by asking for an example.
The phone interviews were transcribedand independentlycoded by the two authorsin terms of
interview question standardization and response scoring standardization.The aforementioned
levels of question standardization'and scoring standardizationof Huffcuttand Arthur(1994) served
'Because none of the interviewswas assigned to the fourthlevel of interview question standardizationof Huffcuttand Arthur
(1994), the results of Studies 1 and 2 deal only with the first three levels of interview question standardization(see Tables 1
and 2).

Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERSAND HIGH STRUCTUREINTERVIEWS

33

Table 1. Usage of levels of interview structurein Study 1

Response scoring standardization

Interviewquestion standardization
Level 1

Level 1
Level 2
Level 3
Total

11(14%)
6 (8%)
0 (0%)
17 (22%)

Level 2
20
22
2
44

(26%)
(29%)
(3%)
(58%)

Level 3
4 (5%)
9 (12%)
2 (3%)
15 (20%)

Total
35 (46%)
37 (49%)
4 (5%)
76 (100%)

Notes:N= 76. Level 1 questionstandardization

was no constraints.
Level2 was constraints
on the topicsto be coveredin
the interview.Level3 was a prioridetermination
of all mainandfollow-upquestionsfromwhichinterviewers
can choose.
Level1 responsescoringwasa globalevaluation
of candidates
of candidates.
Level2 wasevaluation
alongmultipleestablished
criteria.Level3 wasevaluation
of eachindividual
answers(seeHuffcutt&
to pre-established
responseof candidates
according
Arthur,1994).

as coding categories. Inter-rateragreement measured by Cohen's (1960) kappa was 0.75 for
coding interview question standardizationand 0.81 for coding response scoring standardization.
Discrepancies among the coders were resolved upon discussion.

Results
Table 1 presents the results of the phone survey. In terms of interview question standardization,the
majorityof HR representatives(58 per cent) reportedthat they determinedthe topical areasto be covered in the interview in advance. In terms of response scoring standardization,most interviewers(49
per cent) said that they rated applicants on multiple pre-establishedcriteria, although many interviewers (46 per cent) also told us that they gave a global rating. It was unusual (only 5 per cent)
for interviewersto rate each individualresponse of candidatesaccording to example answers (level
3 of response scoring standardization).
When we focus on the cell percentagesinstead of on the totals across interview/responsestandardization, the highest usage percentage(29 per cent) was obtainedfor interviews in which the topical
areaswere determineda prioriand in which the candidateswere ratedalong multiple establishedcriteria. The practiceof determiningthe topics in advanceand globally assessing the candidatewas also
commonly reported(26 per cent).
Because we used a more fine-grainedmeasure of interview structure,this study also shows where
specific bottlenecksexist. In fact, only 3 per cent of the interviewersreportedthat they a prioridetermined all main and follow-up questions and that they rated each individual response of candidates
accordingto example answers.The low usage of such high structureinterviewsis troublesomebecause
there is meta-analyticalevidence that highly structuredinterviews yield incremental validity over
moderately structuredinterviews with corrected validity coefficients increasing from 0.35 to 0.56
(Huffcutt & Arthur, 1994, p. 188). All of this confirms the importance of investigating the interviewer-relatedfactors behind the underutilizationof high structureinterviews. As noted above, this
is the focus of Study 2.
Taken together, this survey indicates that the majorityof interviewersreportedthat they conduct
interviews with a modest degree of structurein terms of question and scoring standardization.This
modest degree of structureis reflectedin the fact that constraintsare placed only on the topical areas
to be discussed and that scoring is done only on multiple criteria.Thus, these results also show that
the underutilizationof high structureinterviews is not an American phenomenonbut generalizes to
Europeancountries such as Belgium.
Copyright( 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

34

E LIEVENS AND A. DE PAEPE

Study 2
Background
Sources of resistance to high structure interviews
Dipboye (1994, 1997) developed a broadconceptual model of the factors that discourage the use of
high structureinterviews.In this model, organization-relatedand interviewer-relatedfactors were distinguished.Organization-relatedfactors involved the normsof proceduraland distributivejustice that
exist in the organization,organizationalpolitics, and culturalassumptionsand values associatedwith
the organization.Interviewer-relatedfactors included the personal needs of interviewers and their
desire to provide a good person-organizationfit. Generally,accordingto Dipboye, low structureinterviews seem to persist in practice because they are more personally satisfying for interviewers.For
instance, low structureinterviews provide interviewerswith the opportunityto express their idiosyncratic preferencesand allow them discretionover the questions to be asked. Low structureinterviews
also do not put interviewersin the role of passive observers.Instead,they allow more open and informal communicationand challenge interviewersto make a global intuitive assessment of candidatefit
with the organization.
In anotherrecent conceptual paper,Harrisand Eder (1999) delved deeper into these interviewerrelatedfactorsthatreducethe use of high structureinterviews. Specifically,they classified the various
interviewer-relatedneeds and concernsenumeratedby Dipboye (1994, 1997) into four specific groups.
First, interviewersmight find it importantthat employment interviews do not reduce their discretion
and control in terms of interviewquestions and scoring. Thereis indeed some empiricalevidence that
interviewersvalue discretionand flexibility in termsof interviewquestions and scoring. A recent survey revealed that more practitionersdisagreedthan agreed with the practiceof asking the exact same
questions across applicantsand of ratingapplicantson criteria(Dipboye & Jackson, 1999). In a survey
among prominentI/O psychologists, discretionover interview questions and scoring also emergedas
one of the main reasons for the continued popularityof low structureinterviews (Church, 1996).
Although these results are insightful, none of these studies examinedthe actual relationshipbetween
discretion when conductingemployment interviews and resistanceto the use of high structureinterviews. Therefore,this study tests the following hypothesis:
Hypothesis 1: If interviewers attach importance to a high level of discretion when conducting
employment interviews, they will be less inclined to use higher levels of structurein interviews.
Second, Harrisand Eder (1999) speculatedthat interviewersmight prefera certainlevel of informality in employmentinterviews.More generally,this refersto the importantsocial function of employment interviews (see Anderson, 1992). Examples of the social dimension of employment interviews
include establishingan informaland personalcontact with the interviewee,having a two-way communicationprocess, or makingthe intervieweefeel comfortable.This social dimensionis especially relevant if the interview is used as a recruitmentdevice (Rynes, 1989). In that case, the central goal
consists of selling the job and the organizationto applicantsand of providinga realisticjob preview.
Hence, it is crucial that the intervieweris seen as warm and friendly because interviewees often use
interviewerbehaviorsas signals for the organization'sculture(Turban,Forret,& Hendrickson,1998).
Some studies revealed that interviewees assign great importanceto the social dimension of interviews. Kohn and Dipboye (1998) found that studentapplicantsreacted more favorably to low structure interviews and that they particularlypreferredinterviews high on interpersonalwarmth,high on
voice, and low on question standardization.Hysong and Dipboye (1999) discovered that student
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERSAND HIGH STRUCTUREINTERVIEWS

35

applicantspreferredinterview practices, which treatedthem as 'people' (e.g., asking personal questions). Conversely,structuredinterview practices such as the interviewercontrollingthe interview or
scoring answers numerically were least preferred. Latham and Finnegan (1993) reported similar
applicantperceptionsregardingemploymentinterviews. Applicantsfavored low structureinterviews
because they allowed them to feel at ease and to say everything they wanted. Although empirical
research on the interviewer's perspective is lacking, we expect that, similar to interviewees, interviewers will be concerned about having personal and informal contact with applicants.In turn, this
concern might impede the more frequentuse of high structureinterviews. Therefore,the following
hypothesis is proposed:
Hypothesis2: If interviewersare concernedabouthaving personalcontact with intervieweeswhen
conducting employment interviews, they will be less inclined to use higher levels of structurein
interviews.
A thirdfactorof resistancemight stem frominterviewers'endeavorsto go beyond assessingpersonjob fit and to evaluateperson-organizationfit (Dipboye, 1994, 1997). Indeed, in the person-organization fit literature,it has often been assertedthat low structureinterviews are ideally suited to assess
person-organizationfit (Cable & Judge, 1997; Chatman, 1991; Judge & Ferris, 1992; Rynes &
Gerhart,1990), even though a recent review study found that person-organizationfit is also often
measuredin high structureinterviews (Huffcutt,Conway,Roth, & Stone, 2001). No empiricalstudies
have linked this interviewer-relatedconcern to the use of structure in employment interviews.
However, it can be hypothesized that, similar to assertionsin the person-organizationfit literature,
interviewersperceive low structureinterviews as very well suited to go beyond the person-job fit
model and to assess person-organizationfit. In turn, this concern might impede the more frequent
use of high structureinterviews. Thus:
Hypothesis 3: If interviewersassign great importanceto the assessment of applicant fit with the
organizationin employmentinterviews, they will be less inclined to use higher levels of structure
in interviews.
Finally, according to Harris and Eder (1999), one of the strongest sources of interviewer-related
resistanceto high structureinterviewsmight stem from the increasedpreparationrelatedto the development of high structureinterviews. High structureinterviews demandmore preparationtime due to
the need for a careful job analysis, the development of a fixed list of job-related questions, and the
constructionof detailed scoring guides. Typically,this developmenttime increases when the level of
structurein employment interviews increases (Huffcutt & Woehr, 1999). Interviewers'ease of preparationconcerns might be at odds with the higher development time of high structureinterviews.
There is some empirical evidence for this interviewer-relatedconcern. Latham and Finnegan
(1993) showed thatinterviewers(N= 59) perceived high structureinterviews as lower on ease of preparation.Although this study suggested that the reducedease of preparationmight serve as an objection for using high structureinterviews, the actual link between interviewers'perceptionsand use of
higher levels of structurein interviews was not examined. Therefore, this study tests the following
hypothesis:
Hypothesis4: When interviewersassign greatimportanceto ease of preparationin interviewdevelopment, they will be less inclined to use higher levels of structurein interviews.
Note thatthis hypothesisdeals only with the time demandsin the preparation(development)of high
structureinterviews. After all, once high structuredinterviews have been developed, their actual use
might be more time efficient than their low structuredcounterparts.
Copyright( 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

36

F. LIEVENS AND A. DE PAEPE

Individualdifferences
Resistanceto high structureinterviewingmight also dependon variousindividualdifferencesvariables
(Dipboye, 1994; Graves & Karren,1999). A firstrelevantvariablemight be an interviewer'sknowledge of structuredinterviewingpractices and their benefits (Terpstra& Rozell, 1997). When people
participatein an interviewingworkshop,for instance, they usually will learn aboutthe academic literaturesupportinghigh structureinterviews. Besides specific interview training,we also expect that
people with a backgroundin industrialand organizationalpsychology will be more likely to use high
structureinterviews.This is because people with a degree in industrialand organizationalpsychology
typically know more about the reliability and validity of high structureinterviewsthan other people.
All of this leads to the following two hypotheses:
Hypothesis5: Interviewerswho have participatedin interviewingworkshopswill be more likely to
use higher levels of structurein interviews.
Hypothesis6: Interviewerswith a degree in industrialand organizationalpsychology will be more
likely to use higher levels of structurein interviews.
Next, several scholars (Harris& Eder, 1999; Campion et al., 1997) emphasized the role of interviewing experience. Harris and Eder (1999) speculated that experienced interviewers might show
more resistanceto high structureinterviewsand theirinherentloss of discretionover questions/scoring
than inexperienced interviewers because experienced interviewers have more ingrained habits.
Furthermore,Campionet al. (1997) posited that experiencedinterviewersmight dislike a high structure interview because they consider it a dull and mindless exercise. Thus:
Hypothesis7: We expect thatmore experiencedinterviewerswill be less likely to use higher level of
structurein interviews.
Apartfrom interviewers'trainingand experience, interviewers'self-image of being good judges of
charactermight also play an importantrole (Dipboye, 1994, 1997). This was confirmedby several
surveys, which indicatedthat many interviewersare confidentabout their personalintuitions of candidates (Bretz, Rynes, & Gerhart,1993; Hakel, 1982) and believe they can readbetween the lines and
detect who candidatesreally are (Dipboye & Jackson, 1999). We expect that interviewerswho perceive themselves as skillful in reading people will be less likely to use high structureinterviews
because they consider structuredinterview practices as unnecessary. This leads to the following
hypothesis:
Hypothesis 8: Interviewerswith a self-image of having good insight into humancharacterwill be
less likely to use higher levels of structurein interviews.
Finally,Dipboye (1994) posited thattwo vocationalpersonalitiesincludedin Holland's(1985, 1997)
RIASEC model might influence the use of high structureinterviews.Although the five-factormodel
(Costa& McRae, 1992;McCrae& Costa, 1996) mightalso be useful in this context,Holland'sRIASEC
model mightbe especially relevantherebecause a key assumptionof this model is thatpeople searchfor
environmentscongruentwith theirinterests(also referredto as vocationalpersonalities).On the basis of
this congruencyhypothesis,Dipboye (1994) assumedthat Social types who are characterizedas sociable, friendly,warm, cooperative,and kind (Holland, 1997) will be more likely to use low structure
interviews as comparedto high structureinterviews because low structureinterviewsenable them to
establish a personaland informalcontact with interviewees.This leads to the following hypothesis:
Hypothesis 9: Interviewershigh on Social will be less likely to use higher levels of structurein
interviews.
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERSAND HIGH STRUCTUREINTERVIEWS

37

Holland (1997) definedConventionaltypes as methodical,orderly,conforming,and unimaginative.

Conventionalpeople also like keeping records,filing materials,andplanningin an orderly,systematic,
and carefulmanner.We expect that Conventionaltypes will have less resistanceagainsthigh structure
interviewsbecause the careful planning of high structureinterviews (e.g., the developmentof a fixed
list of questions, the constructionof detailed scoring guides) is congruentwith their preferredactivities. This leads to the following hypothesis:
Hypothesis 10: Interviewershigh on Conventionalwill be more likely to use higher levels of structure in interviews.

Method
Sample and procedure
Questionnaireswere sent to 190 HR representativesin the Flemish partof Belgium. This questionnaire
sample was generatedin a similarway to the telephonesample. However,respondentswere not necessarily the same (overlap was 20.5 per cent, N= 38; see Discussion section). The questionnairewas
distributeda month after the phone interview through e-mail. Study participationwas voluntary.
We received completedandusable questionnairesfrom 127 HR professionals,yielding a responserate
of 67 per cent. Seventy-fiveof the HR representativeswere female and52 were male. Theirage ranged
from 23 to 54 years, with an average of 33.3 years (SD = 7.5). Ninety-five per cent of the HR representativeshad a universitydegree, of whom 32 per cent had a degree in industrialand organizational
psychology. In Belgium, HR professionalstypically have a middle managementposition because they
reportto the HR managerand are in charge of several HR assistants.Mean full-time working experience of the HR representativeswas 8.7 years (SD = 7.7). The mean interviewingexperience was 6.1
years (SD = 4.8).
HR representativesspenton average28 per cent of theirjob time conductingemploymentinterviews
and they interviewedmost frequentlyforjobs in production(24 percent), followed by technical(17 per
cent), clerical (15 per cent), and sales (14 per cent) jobs. With the exception of one HR representative,
all HR representativesindicatedthatthey had relativefreedomin deciding abouttheir style of conducting employmentinterviews. These last two elements are importantbecause a study on the effects of
interviewer-relatedfactorson the use of the degree of structurein interviewsis meaningfulonly under
the conditionsthat (a) the respondentssampled actually are responsiblefor designing and conducting
employmentinterviewsand (b) they could influence the type of interview that they conduct.
Measures
Interviewerconcerns. With the help of five experiencedinterviewers(threewomen, two men; mean
age = 28.1 years, SD = 2.3 years; mean tenurein consultancyfirm= 4.2 years, SD = 1 year) we constructedtwo items for each of the four interviewerconcerns.Example items are (interviewerconcerns
is parentheses):'An interviewshould allow the interviewerdiscretionwhen asking questions' (discretion), 'There should be a personal contact between the interviewer and the interviewee' (personal
contact), 'An interview is ideally suited to assess whether candidates fit with the organization'
(person-organizationfit), and 'An interview should not requirea lot of preparation'(ease of preparation). Interviewersrespondedto these items using a 5-point scale rangingfrom 1 (Stronglydisagree) to
5 (Stronglyagree). Internalconsistencies for the four scales varied from 0.60 (discretionconcern) to
0.78 (personal contact concern). The lower internal consistency of the discretion concern was
probably due to the fact that one of the items of this scale was negatively worded. Although these
reliabilities may seem low, they are respectablebecause each concern was measuredwith only two
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

38

F. LIEVENSANDA. DE PAEPE

items. We measuredeach concern with only two items to keep the survey as short as possible in order
to maximize the responserateof our field sample. As a standardfor comparison,the Spearman-Brown
correctionformula suggests that scales containing six items instead of two items would all have had
reliability coefficients >0.80.
Individualdifferences. Three items were constructedto measureinterviewers'perceptionsof being
a good judge of character.An example item is: 'I am a good judge of humancharacter.'Interviewers
respondedto these items using a 5-point scale rangingfrom 1 (Stronglydisagree) to 5 (Stronglyagree).
The internalconsistency of this scale was 0.73.
To measurethe vocational personalities,three items were taken from the Social and Conventional
scales of the authorizedDutch adaptation(Hogerheijde,VanAmstel, De Fruyt,& Mervielde, 1995) of
the Self-DirectedSearch(Holland, 1977).2Interviewersrespondedto these items using a 5-point scale
rangingfrom 1 (Stronglydisagree) to 5 (Stronglyagree). Internalconsistencies equaled0.61 (Conventional) and 0.67 (Social). Again, these reliabilitiesare respectablebecause only threeitems were taken
from both the Social and Conventionalscales. If we had measuredSocial and Conventionalwith the
full scale of 10 items, the Spearman-Browncorrectionformula suggests that the scales would have
yielded reliabilitycoefficients of 0.84 and 0.87, respectively.
Besides these scales, the questionnaireincluded questions about respondents'interviewingexperience (in months), educational background,and participationin interviewing workshops (yes/no).
Respondentswrote down their answers (e.g., averageinterviewtime) or checked the appropriatebox.
Dependent variable
We used Huffcuttand Arthur's(1994) two-dimensionalmodel to constructtwo items relatedto interview structure.The first item measuredinterviewers'practicesregardinginterview question standardization.Respondentsratedtheirlevel of interviewquestionstandardizationusing a 5-point scale, with
the following behavioralanchors: 1 (Thereare no constraintsabout the topical areas to be discussed
or the questions to be asked in the interview),3 (1 determinein advance the topical areas to be discussed in the interview),and 5 (1determinein advance all main andfollow-up questionsto be asked in
the interview).These behavioralanchorswere adaptedfrom the differentlevels of interview question
standardizationdistinguishedby Huffcuttand Arthur(1994). The second item measuredinterviewers'
practicesregardingresponse scoring standardizationon a 5-point scale, with the following behavioral
anchors: 1 (I assess candidates globally), 3 (1 assess candidates along multiple established criteria),
and 5 (1 assess each individualresponse of candidates according to pre-establishedanswers). These
behavioralanchorswere adaptedfrom the differentlevels of response scoring standardizationdistinguished by Huffcuttand Arthur(1994).
There was a moderate correlation (r 0.32, p < 0.01) between these two items. This confirms
HuffcuttandArthur's(1994) findingthatinterviewquestionstandardizationandresponsescoringstandardizationare two relativelyindependentdimensionsof interviewstructure.Hence, these items were
not summedto constructa measureof interviewstructure.Instead,we followed the same procedureas
Huffcuttand Arthur(1994) to constructa measureof interview structure.In particular,the ratingson
2Similar
assesseseachof thesix
to theoriginalEnglishversionof theSelf-Directed
Search(Holland,1977),theDutchadaptation
interesttypeson the basisof scalesfor (a) the activitiessomeonelikesto do, (b) the competencies
someonehas,and(c) the
characteristics
on
someonelikestodo.Inaddition,theDutchadaptation
includesa fourthscaleassessingpersonality
occupations
the basisof Holland'sdescription
of the six types(Holland,1985).Becausein this studywe werespecificallyinterestedin
set of scales
vocationalpersonalities,
we considered
thislastset of scalesto be mostrelevant.Therefore,
onlytheabbreviated
(andessentiallythe scalestappingthe SocialandConventional
types)wereused.Prior
referringto personalitydescriptors
research(Hogerheijde,
VanAmstel,De Fruyt,& Mervielde,1995;see also Holland,1999)foundfairlyhighconvergence
andthemoretraditional
to vocationalpersonality
(medianr = 0.53,range= 0.43-0.72)betweenthisscalereferring
descriptors
scalesreferring
to activities,competencies,
andoccupations.

Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERSAND HIGH STRUCTUREINTERVIEWS

39

Table2. Usageof levelsof interviewstructure

in Study2
Responsescoring
standardization
(ratingscale)
1

2
3
4
5

Interviewquestionstandardization
(ratingscale)
1

4 (3.15%)a

4 (3.15%)a

2 (1.57%)a
6 (4.72%)b
0 (0%)c
0 (0%)C

7 (5.5 %)a
11 (8.66%)b
0 (0%)c
0 (0%)c

3
7 (5.51%)b

9 (7.09%)b
49 (38.58%)c
1 (0.78%)c
0 (0%)c

4
2 (1.57%)c

6 (4.72%)c
15 (11.81%)c
0 (0%)d
0 (0%)d

5
0 (0%)c

0 (0%)c
2 (1.57%)c
2 (1.57%)d
0 (0%)d

Notes: N= 127. Interview question standardizationwas rated on a 5-point scale, with 1 (There are no constraintsabout the
topical areas to be discussed or the questions to be asked in the interview),3 (I determinein advance the topical areas to be
discussed in the interview),and 5 (1 determinein advance all main andfollow-up questionsto be asked in the interview).These
three anchorsparallelthe threelevels of questionstandardizationof Study 1. Response scoring standardizationwas ratedon a 5point scale, with 1 (I assess candidatesglobally), 3 (1 assess candidatesalong multipleestablishedcriteria), and 5 (I assess each
individualresponseof candidatesaccording to preestablishedanswers). These three anchorsparallelthe threelevels of scoring
standardizationof Study 1. Cells with the same superscriptswere collapsed to form the measure of interview structure(see
Huffcutt& Arthur,1994).

the two items were crossed (see Table 2). Next, specific cells of Table 2 were collapsed to construct
four progressivelyhigher combinationsof structure.To obtain the lowest level of structure(Structure
1), ratingslower than the midpoint(3) of the scale on both question and scoring standardizationwere
collapsed into one category.Thus, Structure1 representedinterviews with little or no constraintson
questions and scoring. Structure2 was distinguishedby the use of some formal structure.To this end,
ratingsthat were equal to the midpointof the scale on eitherquestion or scoring standardizationwere
collapsed into one category. Structure3 was defined by a relatively high level of structure,although
some variabilityin the process was permitted.This level of structurecontainedratings equal to the
midpoint of the scale on both question and scoring standardizationand ratings higher than the midpoint of the scale on either question or scoring standardization.Finally, Structure4 was characterized
by a high degree of structure.To this end, ratingshigherthanthe midpointof the scale on both question
and scoring standardizationwere collapsed into one category.These four progressivelyhigher combinations of structureare the same as those distinguishedby Huffcuttand Arthur(1994).3

Results
Means, standarddeviations, and intercorrelationsof this study's variables are presentedin Table 3.
Among the interviewerconcerns,establishingpersonalcontactwith the intervieweereceived the highest mean rating (M = 4.32), indicating that interviewersperceived this concern as most importantin
employmentinterviews.To examine our hypotheses, we conducteda multiple regressionanalysis. As
shown in Table4, the interviewerconcernstogetherwith the individualdifferencesvariablesexplained
36.2 per cent of the variance, F (11, 127) = 9.52, p < 0.001. Table 4 also indicates that interviewer
concerns regarding discretion, personal contact, and ease of preparationhad significant negative
regressioncoefficients. This is consistent with Hypotheses 1, 2, and 4, which stated that greaterconcerns about discretion,personalcontact, and ease of preparationin employmentinterviewswould be

3Slight variationsin the levels collapsed to compose progressivelyhigher combinationsof structureproducedsimilarresults to

those in Tables 3 and 4.

Copyright( 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

40

F. LIEVENSANDA. DE PAEPE

Table 3. Means, standarddeviations, and intercorrelationsof Study 2 variables

M

SD

Interviewerconcerns
1. Discretion
3.79 0.76 2. Personalcontact
4.32 0.61 0.23 3. P-O fit
3.92 0.72 0.26 0.29
4. Ease of preparation
2.57 0.82 -0.04 0.09
Individualdifferences
5. Workshopparticipation 0.63 0.48 -0.03 -0.05
6. I/O psychologist
0.32 0.47 0.08 0.05
7. Interviewingexperience 73.81 57.46 0.03 -0.11
8. Judge of character
3.79 0.53 0.13 0.19
9. Social
4.29 0.64 0.09 0.07
10. Conventional
3.86 0.53 0.14 0.25
Dependent variable
11. Level of structure
2.49 0.74 -0.19 -0.28

10

-0.09

-0.10 -0.17
0.13 0.07 0.08
0.09 -0.17 -0.19 -0.18
0.25 -0.02
0.39 0.04 -0.11
-0.07 0.19 0.00 0.20 0.04
0.32 0.01 0.12 0.17 0.09
-0.12 -0.35

0.37

0.07

0.12 0.14 0.17

0.05 -0.15

0.04 0.15

Notes:N= 127.Correlations
atp < 0.05andcorrelations
equalto orabove0.18aresignificant
equalto orabove0.23atp < 0.01.
All measureswereratedon a 5-pointscale,withtheexceptionof use of structure
(fourpossiblevalues,see Table2), workshop
(0 = no, 1 = yes), andI/Opsychologist(0 = no, 1= yes). Theamountof interviewing
participation
experienceis indicatedin
months.

Table 4. Regression of level of interview structureon concerns and individualdifferences variables

Predictor
Discretion
Personal contact
P-O fit
Ease of preparation
Workshopparticipation
I/O psychologist
Interviewingexperience
Judge of character
Social
Conventional

SE (b)

-0.16
-0.27
-0.06
-0.29
0.40
0.12
0.00
-0.11
0.10
0.28

0.08
0.10
0.09
0.07
0.12
0.13
0.00
0.12
0.09
0.12

-0.16
-0.22
-0.05
-0.32
0.26
0.07
0.02
-0.08
0.09
0.20

-2.09
-2.73
-0.60
-4.04
3.28
0.91
0.23
-0.97
1.10
2.39

p
0.04
0.01
0.55
0.00
0.00
0.37
0.82
0.33
0.27
0.02

*p< 0.05; **p< 0.01; ***p< 0.001. R2= 0.36;adjustedR2= 0.31.

relatedto lower levels of structurein the interview.The regressioncoefficient of person-organization

fit was not significant.This result does not supportHypothesis 3.
With regardto the individualdifferences variables,two of our hypotheses were confirmed.In line
with Hypothesis 5, interviewerswho participatedin interviewingworkshopsused significantlyhigher
levels of structurein employment interviews. In addition,Conventionaltypes were found to use significantlyhigher levels of structure,supportingHypothesis 10. Contraryto expectations,interviewers'
educationalbackground,interviewers'self-perceptionsof being good judges of character,and Social
did not have significant effects on the degree of structureused in interviews. We used Baron and
Kenny's criteria(1986) to examine whetherthese individualdifferencesvariableshad indirecteffects
(mediatedby the interviewers'concerns)on the degree of structureused in interviews.No supportwas
found for possible mediatingeffects of interviewers'concerns.
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERS
ANDHIGHSTRUCTURE
INTERVIEWS 41

GeneralDiscussion
Main conclusions
Past researchon high structureinterviewshas been mainly psychometricallyoriented.Granted,it is of
key importanceto increase interview reliability and validity. However, due to the emphasis on measurementissues, researchershave neglected the perspectiveof the interviewers.This study is a step in
shifting the emphasis from psychometricallyoriented issues to more motivationallyorientedissues.
This study lends empirical supportto three interviewer-relatedfactors that discourage the use of
high structureinterviews. When interviewersare concerned about having discretion over questions/
scoring, want to establish personal and informal contact with interviewees, and attach importance
to the ease of preparation,our results suggest that they will be significantlyless inclined to use structure in employmentinterviews.No empiricalsupportwas found for our hypothesisthatthe concernto
assess person-organizationfit would affect the use of high structureinterviews.
The non-significantresultfor person-organizationfit does not confirmassertionsin the person-organization fit literaturethatparticularlylow structureinterviewsare suited to judge person-organization
fit (Cable & Judge, 1997; Chatman,1991; Rynes & Gerhart,1990). The distinctionbetween complementaryfit and supplementaryfit may explain these divergentresults. According to Kristof (1996),
complementaryfit refersto the degree to which persons supply the knowledge, skills, and abilities that
organizationsneed. Conversely,in supplementaryfit, the compatibilitybetween personsand organizations is conceptualizedin terms of more basic characteristicssuch as values, goals, and norms. It is
possible thatpreviousstatementsaboutthe usefulnessof low structureinterviewsto assess person-organizationwere typically made at the level of supplementaryfit. However,this study'sinterviewersmight
have interpretedperson-organizationfit in terms of complementaryfit and thereforemight have perceived high structureinterviewsand the assessmentof knowledge, skills, and abilities as reconcilable.
Another importantfinding is that interviewing workshopparticipationwas strongly related to the
use of high structureinterviews. Thus, regardlessof the specific content and method of interviewing
workshops,they seem to have effects beyond teaching various technical principlesbecause they also
appearto encourage interviewersto use higher levels of structurein interviews. It should be noted,
however,thatthese resultsare correlationaland not causational.Therefore,it is also possible thatinterviewers who initially are more motivatedto use high structureinterviewsaremore likely to participate
in interviewingworkshops.
Finally, this study contributesto the interview literaturebecause we found that interviewers'vocational personalitiesplay some role in interviewermotivationto use high structureinterviews. In fact,
the vocational personality 'Conventional'influencedthe degree of structurein interviews, with more
Conventionalpeople reportingthatthey used higherlevels of structure.In supportof Holland's(1997)
congruencyhypothesismethodical,systematic,careful,and forward-planningpeople seem to be more
likely to use higher levels of structurein interviews.

Limitations
We acknowledge certain limitations of our study. In this study, we focused on the concerns of HR
representativesof largeprivatefirms.As was evidencedby the averagejob time these HR officersspent
interviewing,interviewingplayed a centralrole in theirjob. Given their middle-level position in HR
departments,they also reportedthatthey had some discretionin designing and conductinginterviews.
Futureresearch,however,should examine the generalizabilityof our findingsin otherpopulations.For
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

42

E LIEVENS AND A. DE PAEPE

example, interviewingmight play a more peripheralrole in the job of otherpeople (e.g., typical line
managers who occasionally interview candidates). Accordingly, these people might assign greater
importanceto some of the concerns or might even have different objections against high structure
interviews.Given thatthis studywas situatedin a Belgian context (see contextualsidebar),it is equally
importantthatfuturestudies examine the generalizabilityof our findingsin othercountries.For example, it is possible thatHR representativesin the United Stateshave less freedomin deciding abouttheir
interviewingstyle because of legal considerations,so that interviewer-relatedobjections become less
relevant.Similarly,in settings such as the public sector,the effects of interviewer-relatedfactorson the
use of structurein interviews might also be reducedbecause there might exist very stringentrules in
terms of interview format.
Otherlimitationsresultfrom the use of a cross-sectionalresearchdesign and a surveyinstrumentfor
collecting the data.For example, we tried to keep the survey as shortas possible in orderto maximize
the responserate. However,this reducedsurveylength limited the qualityof the informationcollected.
For example,constructswere measuredwith only two items and abbreviatedversionsof existing scales
were used, leading to attenuated,though still acceptable,internalconsistency values. This may have
underestimatedthe effects found. Because our resultsare based on cross-sectionalself-reportsto a survey, common methodvariancein the formof primingandconsistencyeffects may also be an alternative
explanationfor ourresults.However,we do not believe thatmonomethodbias is a likely explanationfor
our results.Monomethodbias may inflatethe resultsbut it does not explain the patternof relationships
found (i.e., the insignificanteffect of person-organizationfit) (Turbanet al., 1998).
Moreover,the overlapin respondentsacross both samples enabled us to check the quality (reliability) of our data.Specifically,we correlatedthe ratingsaboutthe use of interviewstructurefor the group
of HR representatives(N = 38) thatwere includedin both the telephoneand questionnairesurvey.Ratings across the two surveys correlated 0.75 (p < 0.001) for question standardizationand 0.65
(p < 0.001) for response scoring standardization.These values attest to the quality of the data collected, especially if one takes into considerationthatdifferentmethods(i.e., phone surveyversusquestionnairesurvey,open-endedanswers versus rating scales) were used across the surveys.

Implicationsfor practice and future research

Interviewers'motivationto use high structureinterviews is of key practicalimportancebecause-as
confirmedby this study-low structureinterviews and interviews with a modest degree of structure
are still often used despite their substantiallylower validity.Therefore,the challenge in practiceconsists of developing strategies,which promotehigher levels of structurein employmentinterviewsand
thereforehigher reliability/validity,while at the same time ensuring favorable interviewer/applicant
reactions. Because we found empirical supportfor some of the possible interviewer-relatedfactors,
which reduce the level of structureused in interviews, our results might be informativeas to which
strategiesto choose from.
First,interviewersshouldbe encouragedto participatein an interviewingworkshop.In this workshop/
training,particularattentionshouldbe paid to convincinginterviewersthatlow structureinterviewsmay
often lead to biased informationgatheringand decision making (see Dipboye, 1997, for a detaileddiscussion of these biases) andhence to lowervalidity.Herebywe recommendthatpractitionersgo beyond
a lecture with presentationof validity evidence. This is because researchon social persuasion,belief
change, and attitudechange (e.g., Morley, 1987) demonstratesthat people are often more impressed
by 'vivid' anecdotalevidence (e.g., case studies) thanby abstractevidence (e.g., validity coefficients).
Second, to increaseinterviewers'motivation,we recommendinvolving them in strategicinterviewing issues such as the adoption of high structureinterviewing in organizations.If interviewers are
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERSAND HIGH STRUCTUREINTERVIEWS

43

involved in the decision regardingwhich type of structuredinterviewformatto adopt,we believe they

are less likely to feel that structuredinterviewingpracticeswere imposed on them and subsequently
limit their discretion.
Whereas both former strategies aim to persuadeinterviewersto use higher levels of structurein
interviewswhile leaving the structuredinterviewformatunchanged,a thirdstrategyconsists of subtly
modifying the structuredinterview format. Specifically, interviewers'concerns about discretionover
questions/scoringand about establishing informal contact with interviewees may also be accommodated by including a period of free conversationupon completion of the high structurepart and by
communicatingthis to interviewees at the startof the interview.An example of such an interviewformat is the multimodalinterview (Schuler & Funke, 1989). This interview contains situationalquestions, biographicalquestions, and a period of free interviewerand applicantquestions.
We believe that futureresearchshould be geared at the following four directions.First of all, more
research on the interviewerperspective should be conducted. In this study, we focused on two key
operationaldimensions of interview structure(i.e., interview question standardizationand response
scoring standardization).Future studies should scrutinize interviewer reactions to other structured
interviewingpractices (see Campionet al., 1997; Dipboye & Gaugler, 1993). Examples include conductinga job analysis, note takingduringthe interview,the absenceof a pre-reviewof candidateinformation, or using a panel of interviewers.In addition, we need to know whether interview content
(situationalinterviews versus behavior descriptioninterviews) makes a difference in terms of interviewer perceptionsandresistance.Initialresearchby LathamandFinnegan(1993) indicatedthatinterviewers (N= 59) reacted significantly more favorably to situational interviews as compared to
behaviordescriptioninterviews.
As a second avenue for futureresearch,we need to broadenthe range of criteriaused to evaluate
intervieweffectiveness. In particular,both interviewers'reliability/validityand interviewers'reactions
should serve as criteria.After all, the ultimategoal shouldbe to promotethe reliability/validityof high
structureinterviews while at the same time ensuringinterviewermotivationto use these interviews.
Third,broadertheoreticalframeworksshould be applied to explain people's motivation/resistance
to use high structureinterviews. An example is Ajzen's (1991) theory of planned behavior, which
states thatpeople's intentions(e.g., to use high structureinterviews)are determinedby their attitudes,
subjectivenorms,and perceivedcontrol(see Vander Zee et al., 2002). Anotherexample is the broader
literatureon judges' confidence in their expertise (see Camerer& Johnson, 1991, for a review).
Finally,futureresearchshouldnot only examine the role of interviewer-relatedimpedimentsto high
structureinterviewsbut should also include variousorganizationalfactors. For instance, accordingto
Dipboye (1994, 1997), low structureinterviewscontinue to be used in organizationsbecause they are
bettervehicles for conveying an organization'svalues and culture.In addition,political factors such as
seeking to acquireand maintainpower in organizationsmay make low structureinterviews attractive
because of their ambiguity and lack of accountability.An examinationof these organization-related
factors may give a full insight into the complex dynamics underlyingthe less frequentuse of high
structureinterviews in organizations.

Acknowledgements
We would like to acknowledge Winfred Arthur Jr, Jim Conway, and Scott Highhouse for their
valuablecommentson a previousversionof this manuscript.A previousversionof this manuscriptwas
presentedat the 17th Annual Conferenceof the Society of Industrialand OrganizationalPsychology,
Toronto,Canada.
Copyright? 2004 JohnWiley& Sons,Ltd.

J. Organiz.Behav.25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

44

F. LIEVENSANDA. DE PAEPE

Author biographies
Filip Lievens obtained his PhD in Industrial and Organizational Psychology at Ghent University. Currently, he is working as an Associate Professor at Ghent University. His research interests include
assessment centers, situational judgment tests, and organizational attrractiveness.
Anneleen De Paepe obtained a Master's degree in Industrial and Organizational Psychology at Ghent
University. Currently, she works as a consultant in human resources at Securex, a leading social management firm in Belgium. She is mainly involved in personnel recruitment and selection activities.

References
Ajzen, I. (1991). The theory of plannedbehavior.OrganizationalBehavior and HumanDecision Processes, 50,
179-211.
Anderson, N. R. (1992). Eight decades of employment research:a retrospectivemeta-review and prospective
commentary.European Workand OrganizationalPsychologist, 2, 1-32.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediatorvariable distinction in social psychological
research:conceptual,strategic,and statisticalconsiderations.Journalof Personalityand Social Psychology,51,
1173-1182.
Bretz, R. D., Rynes, S. L., & Gerhart,B. (1993). Recruiterperceptionsof applicantfit: implicationsfor individual
careerpreparationandjob search behavior.Journal of VocationalBehavior 43, 310-327.
Cable, D. M., & Judge, T. A. (1997). Interviewers'perceptions of person-organizationfit and organizational
selection decisions. Journal of Applied Psychology, 82, 546-561.
Camerer,C. F., & Johnson,E. J. (1991). The process-performanceparadoxin expertjudgment:how can experts
know so much and predictso badly?In K. A. Ericsson,& J. Smith (Eds.), Towarda general theoryof expertise:
Prospects and limits (pp. 195-217). New York:CambridgeUniversity Press.
Campion, M. A., Palmer, D. K., & Campion, J. E. (1997). A review of structurein the selection interview.
Personnel Psychology, 50, 655-702.
Chatman,J. A. (1991). Matchingpeople and organizations:selection and socializationin public accountingfirms.
AdministrativeScience Quarterly,36, 459-484.
Church, A. H. (1996). From both sides now: the employee interview-the great pretender. Industrial
Psychologist, 34(1), 108-117.
Cohen,J. (1960). A coefficientof agreementfor nominalscales. Educationaland Psychological Measurement,20,
37-46.
Conway, J. M., Jako, R. A., & Goodman, D. F. (1995). A meta-analysisof interraterand internal consistency
reliability of selection interviews.Journal of Applied Psychology, 80, 565-579.
Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory(NEO-PI-R) and NEO five-factor
inventory(NEO-FFI)professional manual. Odessa: PAR.
Dipboye, R. L. (1994). Structuredand unstructuredselection interviews: beyond the job-fit model. In G. Ferris
(Ed.), Research in personnel and human resources management(Vol. 12, pp. 79-123). Greenwich, CT: JAI
Press.
Dipboye, R. L. (1997). Structuredselection interviews: why do they work? Why are they underutilized?In N.
Anderson,& P. Herriot(Eds.), Internationalhandbookof selection and assessment (pp. 455-473). New York:
Wiley.
Dipboye, R. L., & Gaugler,B. B. (1993). Cognitive and behavioralprocesses in the selection interview.In N.
Schmitt,& W. Borman(Eds.), Personnelselection in organizations(pp. 135-170). San Francisco:Jossey-Bass.
Dipboye, R. L., & Jackson, S. L. (1999). Interviewerexperience and expertise effects. In R. W. Eder, & M. M.
Harris(Eds.), The employmentinterviewhandbook(pp. 279-292). ThousandsOaks, CA: Sage.
Graves,L. M., & Karren,R. J. (1999). Are some interviewersbetterthan others. In R. W. Eder, & M. M. Harris
(Eds.), The employmentinterviewhandbook(pp. 243-258). ThousandsOaks, CA: Sage.

Copyright? 2004 John Wiley & Sons, Ltd.

J. Organiz.Behav. 25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

INTERVIEWERS
ANDHIGHSTRUCTURE
INTERVIEWS 45
Hakel, M. D. (1982). Employmentinterviewing.In K. M. Rowland,& G. R. Ferris(Eds.),Personnelmanagement
(pp. 129-155). Boston: Allyn & Bacon.
Harris,M. M., Dworkin, J. B., & Park, J. (1990). Preemploymentscreening procedures:how human resource
managersperceive them. Journal of Business and Psychology, 4, 279-292.
Harris,M. M., & Eder,R. W. (1999). The state of employmentinterviewpractice:commentaryand extension.In
R. W. Eder,& M. M. Harris(Eds.), The employmentinterviewhandbook(pp. 369-398). ThousandsOaks, CA:
Sage.
Hogerheijde,R. P., Van Amstel, B., De Fruyt, F., & Mervielde, I. (1995). BeroepskeuzeZelf-Onderzoekversie
1995. Een gids voor opleidings- en beroepskeuze. Testboekje-Nederlands/Vlaamse
editie [Dutch/Flemish
adaptation of the Self-DirectedSearch: TestBooklet]. Lisse: Swets & Zeitlinger.
Holland, J. L. (1977). Self-directed search: A guide to educational and vocational planning. Palo Alto, CA:
Consulting Psychologists Press.
Holland, J. L. (1985). Making vocational choices: A theoryof careers. Englewood Cliffs, NJ: Prentice-Hall.
Holland, J. L. (1997). Making vocational choices: A theory of vocational personalities and work environments.
Odessa, FL: Psychologist Assessment Resources.
Holland, J. L. (1999). Why interest inventories are also personality inventories. In M. L. Savickas, & A. R.
Spokane(Eds.), Vocationalinterests:Meaning, measurement,and counseling use (pp. 87-101). Palo Alto, CA:
ConsultingPsychologists Press.
Huffcutt,A. I., & Arthur,W. Jr.(1994). HunterandHunter(1984) revisited:interviewvalidityfor entry-leveljobs.
Journal of Applied Psychology, 79, 184-190.
Huffcutt,A. I., Conway,J. M., Roth, P. L., & Stone, N. J. (2001). Identificationand meta-analyticassessment of
psychological constructsmeasuredin employmentinterviews.Journal of Applied Psychology, 86, 897-913.
Huffcutt,A. I., & Roth, P. L. (1997). Racial group differences in employmentinterview evaluations.Journal of
Applied Psychology, 83, 179-189.
Huffcutt,A. I., & Woehr,D. J. (1999). Furtheranalysis of employmentinterviewvalidity:a quantitativeevaluation
of interviewer-relatedstructuringmethods.Journal of OrganizationalBehavior,20, 549-560.
Hysong, S. J., & Dipboye, R. L. (1999). Individualdifferencesin applicants'perceptionsto employmentinterview
elements. Paper presented at the 14th Annual Conference of the Society for Industrialand Organizational
Psychology, Atlanta,GA, April 1999.
Johns, G. (1993). Constraints on the adoption of psychology-based personnel practices: lessons from
organizationalinnovation.Personnel Psychology, 46, 569-592.
Judge, T. A., & Ferris, G. A. (1992). The elusive criterionof fit in human resources staffingdecisions. Human
Resource Planning, 15, 47-67.
Kohn, L. S., & Dipboye, R. L. (1998). The effects of interview structureon recruitingoutcomes. Journal of
Applied Social Psychology, 28, 821-843.
Kristof, A. L. (1996). Person-organizationfit: an integrativereview of its conceptualizations,measurement,and
implications. Personnel Psychology, 49, 1-49.
Latham, G. P., & Finnegan, B. J. (1993). Perceived practicality of unstructured,patterned, and situational
interviews. In H. Schuler,J. L. Farr,& M. Smith (Eds.), Personnel selection and assessment: Individualand
organizationalperspectives (pp. 41-55). Hillsdale, NJ: Erlbaum.
Latham,G. P., & Saari, L. M. (1984). Do people do what they say? Furtherstudies on the situationalinterview.
Journal of Applied Psychology, 69, 569-573.
Marchese, M. C., & Muchinsky, P. M. (1993). The validity of the employment interview: a meta-analysis.
InternationalJournal of Selection and Assessment, 1, 18-26.
McCrae, R. R., & Costa, P. T. Jr. (1996). Towarda new generationof personalitytheories: theoreticalcontexts
for the five-factormodel. In J. S. Wiggins (Ed.), Thefive-factor model of personality: Theoreticalperspectives
(pp. 51-87). New York:Guilford.
McDaniel, M. A., Whetzel, D. L., Schmidt,F. L., & Maurer,S. (1994). The validity of employmentinterviews:a
comprehensivereview and meta-analysis.Journal of Applied Psychology, 79, 599-616.
Morley, D. D. (1987). Subjective message constructs:a theory of persuasion.CommunicationMonographs,54,
183-203.
Ryan, A. M., McFarland,L., Baron,H., & Page, R. (1999). An internationallook at selection practices:nationand
culture as explanationsfor variabilityin practice.PersonnelPsychology, 52, 359-391.
Rynes, S. L. (1989). The employmentinterviewas a recruitmentdevice. In R. W. Eder,& G. R. Ferris(Eds.), The
employmentinterview:Theory,research,and practice (pp. 127-142). Newbury Park,CA: Sage.

Copyright? 2004 John Wiley & Sons, Ltd.

J. Organiz.Behav. 25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions

46

F. LIEVENSANDA. DE PAEPE

Rynes, S. L., Barber,A. E., & Varma,G. H. (2000). Researchon the employmentinterview:usefulnessfor practice
and recommendationsfor future research.In C. Cooper, & E. Locke (Eds.), I/O psychology: Whatwe know
about theory and practice (pp. 250-277). Oxford:Blackwell.
Rynes, S. L., & Gerhart,B. (1990). Interviewerassessments of applicant 'fit': an exploratory investigation.
Personnel Psychology, 43, 13-22.
Schuler,H., & Funke,U. (1989). The interviewas a multimodalprocedure.In R. W. Eder,& G. R. Ferris(Eds.),
The employmentinterview:Theory,research, and practice (pp. 183-189). NewburyPark,CA: Sage.
Terpstra,D. E., & Rozell, E. J. (1997). Why some potentiallyeffective staffingpracticesare seldom used. Public
Personnel Management,26, 483-495.
Turban,D. B., Forret,M. L., & Hendrickson,C. L. (1998). Applicantattractionto firms:influencesof organization
reputation,job and organizationalattributes,and recruiterbehaviors. Journal of Vocational Behavior, 52,
24-44.
Vander Zee, K. I., Bakker,A. B., & Bakker,P. (2002). Why are structuredinterviewsso rarelyused in personnel
selection? Journal of Applied Psychology, 87, 176-184.
Wiesner,W. H., & Cronshaw,S. F. (1988). A meta-analyticinvestigationof the impact of interview formatand
degree of structureon the validity of the employment interview.Journal of Occupationaland Organizational
Psychology, 61, 275-290.
Williamson, L. G., Campion, J. E., Malos, S. B., Roehling, M. A., & Campion, M. A. (1997). Employment
interview on trial: linking interview structurewith litigation outcomes. Journal of Applied Psychology, 82,
900-912.

Copyright(

2004 John Wiley & Sons, Ltd.

J. Organiz.Behav. 25, 29-46 (2004)

This content downloaded from 193.226.34.229 on Wed, 10 Dec 2014 08:58:42 AM

All use subject to JSTOR Terms and Conditions