The dimensionality of trust-relevant constructs in four institutional domains: results from confirmatory factor analyses

ABSTRACT

Using confirmatory factor analyses and multiple indicators per construct, we examined a number of theoretically derived factor structures pertaining to numerous trust-relevant constructs (from 9 to 12) across four institutional contexts (police, local governance, natural resources, state governance) and multiple participant-types (college students via an online survey, community residents as part of a city's budget engagement activity, a random sample of rural landowners, and a national sample of adult Americans via an Amazon Mechanical Turk study). Across studies, a number of common findings emerged. First, the best fitting models in each study maintained separate factors for each trust-relevant construct. Furthermore, post hoc analyses involving addition of higher-order factors tended to fit better than collapsing of factors. Second, dispositional trust was easily distinguishable from the other trust-related constructs, and positive and negative constructs were often distinguishable. However, the items reflecting positive trust attitude constructs or positive trustworthiness perceptions showed low discriminant validity. Differences in findings between studies raise questions warranting further investigation in future research, including differences in correlations among latent constructs varying from very high (e.g. 12 inter-factor correlations above .9 in Study 2) to more moderate (e.g. only three correlations above .8 in Study 4). Further, the results from one study (Study 4) suggested that legitimacy, fairness, and voice were especially highly correlated and may form a single higher-order factor, but the other studies did not. Future research is needed to determine when and why different higher-order factor structures may emerge in different institutional contexts or with different samples.

KEYWORDS:

• Trust in institutions
• dispositional trust
• confidence
• legitimacy
• justice
• fairness

ACTION EDITOR:

• Peter Ping Li

Acknowledgements

Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of National Science Foundation (NSF).

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Lisa PytlikZillig is a Research Associate Professor at the University of Nebraska Public Policy Center. Her work primarily focuses on public engagement around potentially controversial policy issues and trust in institutions.

Joseph Hamm is Assistant Professor of Criminal Justice & Environmental Science and Policy at Michigan State University. His work focuses on the conceptualisation, measurement, and outcomes of trust in institutions, especially institutions of government.

Ellie Shockley is the Institutional Research Analyst at Bismarck State College. Her research interests include social and political psychology and academic engagement and outcomes.

Mitchel Herian is a faculty fellow with the University of Nebraska Public Policy Center. His recent work has focused on trust-related outcomes in a variety of settings.

Tess M.S. Neal is Assistant Professor of Psychology in the New College at Arizona State University. She studies human inference and decision making, focusing primarily on how experts integrate information to make judgments. She is interested in how trust affects the process of decision making.

Christopher Kimbrough is a Research Scientist with the Amazon Connections Team. His research focuses primarily on jury decision making, psychology of religion, and research methodology and analytics.

Alan Tomkins is professor of psychology and law at UNL, currently on leave. He is Acting Division Director, Division of Social and Economic Sciences, Directorate for Social, Behavioral and Economic Sciences, National Science Foundation. His primary research interests focus on trust in institutions and public engagement to inform policy.

Brian Bornstein is Professor of Psychology and Courtesy Professor of Law at the University of Nebraska-Lincoln, where he directs the Law-Psychology program. His primary research interests are jury decision making, eyewitness memory, and trust in the legal system.

ORCID

Lisa M. PytlikZillig http://orcid.org/0000-0002-0962-8502

Joseph A. Hamm http://orcid.org/0000-0001-9250-3681

Tess M.S. Neal http://orcid.org/0000-0002-9528-8638

Notes

1. McEvily and Tortoriello (2011) and others often refer to ‘dimensions’ but are not using the term dimensions as we do here (i.e. as indicating underlying structural and statistical relationships between the constructs). Many times, authors’ use of the term ‘dimensions’ refers to what we refer to as ‘constructs’ in the present article.

2. Figure 1 orders the models from most to least complex. Our discussion here orders the models somewhat differently, in an order that we feel makes it easier to connect the models to theory. For example, rather than organising our discussion according to model complexity, we discuss the many factor model prior to discussing the various ways of collapsing the many factors of trustworthiness, and beginning our discussion with what we perceive as the most frequently-cited organisation of trustworthiness constructs.

3. Note that institutional trust is not always used this way. For example, it is sometimes used to describe ‘system trust’ or refer to safeguards and policies that encourage trustors to rely upon trustees (Bachmann, 2011; Pennington, Wilcox, & Grover, 2003; PytlikZillig & Kimbrough, 2015), but this is not how we are using the term here.

4. For comparability, the present analysis includes only constructs that were used in at least two of the four studies.

5. As previously mentioned, each study included a somewhat different set of constructs. Also, Study 4 did not include dispositional, direct/unspecified, or loyal trust, and therefore was only able to focus on the dimensionality of the trustworthiness constructs. To distinguish Study 4 models from those including a wider range of constructs, we use an adapted labeling system described later.

6. Whenever we conducted exploratory analyses, we conducted them in multiple ways (e.g. using principal axis factoring (PAF) and principal components analysis (PCA), based on both correlations and covariances, and using Varimax and Promax rotations) and then report the most common grouping of constructs.

7. In addition, although not reported in Figure 1, each successively nested model improved fit to the data according to the same rescaled log-likelihood ratio comparisons. Full results available from the corresponding author.

8. Note that a subset of the constructs evaluated here have been reported elsewhere (Hamm, 2014; Hamm, Hoffman, Bornstein, & Tomkins, in press).

9. It is possible that our student sample, receiving course credit for completing the measures, but perhaps not as interested in expressing their trust-relevant views as our volunteer samples the other studies, were less attentive in their consideration of the different constructs being assessed. However, the quality of the data is supported by the fact that it did result in distinct factors for dispositional trust and for the negative constructs of bias and cynicism.

10. It also may be noteworthy that the respondents in three of our studies were from a single state (Nebraska), which could limit generalisability. However, there was substantial diversity across the three samples (students, rural landowners, citizens of a mid-sized city), and Study 4 included a national sample.

11. We did not conduct a formal test of factor equivalence across studies because of the use of different items, response scales, and inclusion of different constructs across studies.

Additional information

Funding

This research was supported in part by the NSF [grant numbers 0903469, 0965465, 1061635, 1154855, 1228559, 1228937, and 1353980].