Bijlsma-Frankema, Sitkin, and Weibel: Distrust in the Balance
18 Organization Science, Articles in Advance, pp. 1–22, © 2015 INFORMS
theorizing, we have created a model that extends the interrelatedness of actions and cognitions is central to
current understanding of what goes on between groups the understanding of distrust and trust as relational
in a mutual distrust relationship. This approach also phenomena.
reveals how perceptions and behaviors coevolve, empha-
sizing trust and distrust as psychological states that man- Acknowledgments
ifest through interlocking cycles of action. Our results The authors thank Denise Rousseau for her helpful ideas and
provide a springboard for future studies in which the encouragement during the first stages of this study.
Appendix. Summary of Studies Examining the Trust/Distrust Distinction
Author (year) Research question/aim Context/methodology Summary of results
Chang and Fang To explore differences in Online survey; AVE and AVE analyses show that trust and distrust are
(2013) trust and distrust formation SEM analyses distinct constructs.
and asymmetric behavioral Sample: 1,193 online Antecedent asymmetry: a disposition to
outcomes. customers of several Web distrust significantly affects distrust but has
stores no effect on trust.
Consequence asymmetry: the negative effect of
distrust is stronger than the positive effect of
trust on high-risk Internet behaviors.
Dimoka (2010) To shed light on the nature, Experiments: buying on Distinct neural correlates (activated brain
dimensionality, distinction, eBay; using psychometric areas) of trust and distrust are identified,
and relative effects of trust measures of trust and lending support to trust and distrust as
and distrust on economic distrust without and with distinct constructs. Trust is found in the
outcomes (willingness to functional neuroimaging reward, prediction, and uncertainty areas;
pay a price premium). (fMRI) distrust is found in the intense emotion and
Manipulation: four seller fear of loss areas.
profiles of high/low (H/L) Prediction of price premiums respondents are
trust and H/L distrust willing to pay: neural correlates of distrust
Sample: 177 respondents in have a stronger (negative) effect (29%) than
behavioral experiments; 15 do correlates of trust (20%), explaining a
in fMRI experiments medium-large difference in R2 (Cohen’s
f 2 = 0018).
Data suggest that the trust–distrust distinction
relates to intentional/autonomic types of
responses, collected/frenzied natures of
responses, and long-term/short-term time
horizons: “These findings are consistent with
the literature that views trust as developing
slowly over time through careful
deliberation, while distrust is quick and
episodic, based on emotional cues” (p. 390).
Ou and Sia (2010) To investigate the Web survey, simulating “The trust and distrust measures show strong
antecedents (evaluation of buying electronic products within-construct convergent and
Web design attributes, at two real Web stores between-construct discriminant validities,
functional perception, and Discriminant validity tested suggesting that trust and distrust are two
motivational perception of in several ways, including distinct and separate constructs” (p. 923).
website) and influences of AVE square roots and Asymmetric effects on trust and distrust are
trust and distrust on PCA factor analysis found: distrust is significantly influenced by
buying intentions, a Structural model analyzed functional perception but not by motivational
theoretical framework is with PLS analysis perception. Trust is influenced by both
proposed. Asymmetric Sample: 324 undergraduate functional and motivational perception.
effects on and of trust and and postgraduate students Functional perception has an asymmetric
distrust are hypothesized. effect on trust and distrust, acting more
strongly to lower distrust than to build trust.
An asymmetric effect of trust and distrust is
found: the effect of distrust on lowering
buying intentions is greater than the
enhancing influence of trust on buying
intentions.
Bijlsma-Frankema, Sitkin, and Weibel: Distrust in the Balance
Organization Science, Articles in Advance, pp. 1–22, © 2015 INFORMS 19
Appendix. (cont’d)
Author (year) Research question/aim Context/methodology Summary of results
Komiak and “This paper theoretically Experimental and qualitative The analysis shows that the pattern of trust
Benbasat (2008) proposes and empirically study; random distribution and distrust processes differs similarly in the
tests the notion that of participants to two personalized and depersonalized conditions.
building trust and distrust conditions: a personalized Distrust is mainly related to a situation when
in online recommendation condition and a participants become “aware of the
agents (RAs) are two sets nonpersonalized RA unknown,” when the RA is judged
of distinct and separate condition. Measurement incompetent, or when the RA is not
processes” (p. 728). and analysis based on matching individuals’ expectations. Trust is
process protocols of more strongly tied to judging RAs as
participants’ thinking competent, when their information sharing is
aloud about RA websites. evaluated in a positive way, and when third
It is proposed that a parties verify the impression of the
personalized RA will respondent. Furthermore, a personalized RA
evoke more trust than evokes a higher proportion of trust and a
would a nonpersonalized lower proportion of distrust than a
RA. depersonalized RA.
Sample: 49 business school
students, average of
25-minute protocols; 1,062
processes of trust building
and 947 of distrust
building
Cho (2006) To determine how consumer Web survey on books and Discriminant validity of trust–distrust
evaluations of an clothing distinction by SEM model comparison (free
e-vendor’s businesses LISREL path analysis and with a covariance constraint) “strongly
operations relate to Sample: 593 book indicated that trust and distrust are distinct
consumer judgment of purchasers, 288 clothing constructs” (p. 30).
e-vendor trustworthiness, purchasers Asymmetric determinants of trust and distrust
consumer trust and distrust found: benevolence impacted trust more
of the e-vendor, and significantly than did competence, whereas
consumer behavioral competence affected distrust more than did
intentions (self-disclosure benevolence.
and willingness to Asymmetric effects of trust and distrust
commit). Asymmetric found: distrust affected self-disclosure more
determinants and effects of significantly than did trust, whereas trust had
trust and distrust are a more significant effect on the willingness
hypothesized. to commit than did distrust.
Benamati et al. To examine the “nature of Survey; data analyzed with Discriminant validity of trust–distrust shown
(2006) trust versus distrust in [an SEM comparison of by a significant difference between both a
online bank] to determine models and AVE free and covariance-restrained SEM model
if they exist as comparison and AVE comparison.
separate—but Sample: 513 university Trust is positively related, and distrust
related—constructs 0 0 0 and students negatively related, to intentions to use online
the relative influence of banking. Trust is a stronger predictor of use
both constructs on user than is distrust.
intentions to transact”
(p. 2).
Clark and Payne “To examine the nature of Facet approach used to “With regard to the distinction between trust
(1997) propensity to trust as generate a definitional and mistrust, evidence of the correlation and
defined by the mapping framework of trust, the SSA configurations suggest that they
sentence and to see exploratory phase: 44 may represent two sub-constructs within the
whether or not the interviews; hypotheses overall concept of trusting relations”
empirical data would testing: survey study, (p. 222).
reflect the structure of the correlations, and smallest “The distinction between trust and mistrust
definitional framework” space analysis (SSA) items is clearly illustrated in the three
(p. 209). Sample: 428 colliery dimensional SSA solution: The Trust and
workmen in the British mistrust items may be seen to form distinct
coal industry regions. This is an interesting discovery, as
the questionnaire items were designed with
the notion that trust and mistrust were
opposite ends of a continuum” (p. 215).
Bijlsma-Frankema, Sitkin, and Weibel: Distrust in the Balance
20 Organization Science, Articles in Advance, pp. 1–22, © 2015 INFORMS
Appendix. (cont’d)
Author (year) Research question/aim Context/methodology Summary of results
Wrightsman (1974) To reexamine the underlying Factor analysis of PHN scale Based on two separate factor analyses of PHN
dimensions of the 84-item items items, two factors emerged, a positive
Philosophies of Human Sample: 530 undergraduates “beliefs that people are conventionally good”
Nature (PHN) scale, factor (10 items) and a negative factor
designed by the author in labeled “cynicism” (10 items). The two
1964. factors were correlated to each other only to
a limited degree (−0.27 and −0.33 in two
samples, respectively), further supporting the
notion of two distinct constructs.
Wrightsman recommends the use of these
shorter scales for measuring trust and
cynicism (distrust), respectively.
Constantinople To “attempt to extend a Survey Validated scale with two subscales for basic
(1969) self-concept measure Sample: 952 college trust and basic mistrust, derived from
based on E. H. Erikson’s students, ranging from Erikson scale that theoretically and
theory 0 0 0 to a measure of freshmen to seniors empirically distinguished trust from distrust.
personality development in Differential effects on trust and distrust in
late adolescence” (p. 357). support of construct distinction: men scored
higher on trust and distrust than women.
Seniors scored lower on mistrust than
freshmen, whereas no class effect on trust
was found.
Note. AVE, average variance extracted; PCA, principal component analysis; PLS, partial least squares; SEM, structural equations modeling.
Endnotes obtain answers to questions the subjects do not themselves
1
Following Festinger (1954), we define values in this context have but feel compelled to answer nonetheless. By starting
as desired end states. Core values we define as values that with open-ended questions, the interviewer can discover what
are relevant to participants in a specific context, such as the the respondent thinks of naturally and spontaneously. By using
work context. The social interactionist approach used in this the respondent’s own terms in follow-up questions, the inter-
study indicates relevance when respondents talk about these viewer can still avoid imposing categories prematurely on
values unsolicited, make comparisons with perceived values the respondent. Only after the respondent has been given the
of others, and react emotionally to perceived threats to these opportunity and no new information surfaces through free
values. recall (called “saturation”) can the interviewer explicitly ask
2
We use the term “administrators” to refer collectively to directive questions in a structured way. In this way, all direc-
members of the general management team: the general man- tive questions are answered while minimizing the chance of
ager, the court’s middle manager, and the managers of the staff imposition of the interviewer’s preconceptions on respondents.
departments. Interviews may take a little longer, but they get as much or
3
According to Woods (1983, p. 9), the concept of perspec- more information as a singularly structured, directive inter-
tive is defined as “the way people define and interpret the view. Kvale also suggests that the unstructured, undirected
situation they are in, which governs the way they behave in portion of the interview focus on specific topics that the inter-
such a situation.” Perspectives are linked to action through viewer creates for the study. The goal in this study was to
strategies. Strategies are not isolated acts but packages of acts discover what the respondents liked and did not like about
interrelated by intention to solve problems thrown up by the the changes, which aspects of the organization and the change
situation in the pursuit of valued aims (Woods 1983). Perspec- they interpreted as positive or negative, and how and why they
tives and strategies are social in nature and are shaped through behaved the way they did regarding problems they reported.
6
ongoing interactions with others—in our case, the members This is done in several iterations, as other authors advocate
of the groups, producing shared meaning-giving and common (Gioia et al. 2010, Pratt et al. 2006).
7
actions. Quotations from the interviews are indicated by a letter
4
The topics used in this study included one’s work situation, (“J” for judges and “A” for administrators) and a number indi-
relations within one’s work unit, the management of one’s cating which specific interviewee. Where a quoted phrase was
work unit, the management of the court organization, the rela- used by more than one source, no specific source identification
tions between work units, the relations between professional is included.
8
groups (e.g., judges, administrators), the relation of the court Lay use of the term “distrust” does not necessarily coincide
to the environment, past changes in the organization, and with the scientific definition and use of the construct. But we
expectations for the future. would suggest that it can be used, as we have here, in con-
5
Kvale’s (1996) rationale is that by beginning with directed junction with other, theoretically based indicators to assess its
questions, the interviewer can create a demand effect and presence.
