In search of parsimony: a multiple triangulation approach to antecendents of trust in managers

The methodological aim of this study is to search for a parsimonious model to understand the behavioral cues employees use in pondering on trust in managers. A multiple triangulation approach is used to pursue this aim. Based on open interviews with a stratified sample of hospital employees, using the methodology of Kvale (1996), seven hypotheses were formulated that relate trust in managers and behavioral cues, i.c. monitoring performance of subordinates, appreciation of good work, guidance to improve individual performance, support in case of trouble with others, cooperation-related problem solving, openness to ideas of subordinates, and type of leadership. The hypotheses were tested in a survey, that was administered to all organizational members. Correlation analysis, multiple regression analysis, and a Boolean algebra based pattern analysis (Ragin, 1989, 1994) were used to test the hypotheses. The conclusion is that actors react to a few single behavioral cues instead of many complex ones. In the past decade, issues of trust in interand intra-organizational relationships have been increasing in importance on the agendas of organizational scholars, legitimated by changes in the social structure of societies, economic exchange relations, and organizational forms. Due to deterioration in the binding power of reciprocal obligations (Kramer, 1996), of hierarchical relations (Sheppard & Tuschinski, 1996) Antecedents of Trust in Managers 2 and of social institutions relying on hierarchy to sanction deviant behavior (De Swaan, 1990), other mechanisms seem to be needed to keep the social fabric of society intact, in order to sustain cooperative behavior (Kramer, 1996). In contrast to hierarchical relationships that used to dominate the framing of work relations within firms, lateral relationships and alliances are getting more important (Sheppard & Tuschinski, 1996). Between firms, new linkages are formed to achieve and maintain competitive advantage in the marketplace, that make organizations move toward network forms and alliances (Lewicki & Bunker, 1996), that require high levels of trust to function effectively (Creed & Miles, 1996). Furthermore, increasing conditions of change heighten the relevance of trust to organizational performance and to the well-being of organizational members (Mishra, 1996; Gilkey, 1991). By now it is widely acknowledged that trust works as a lubricant in economic transactions, by smoothing relations between actors and reducing transaction costs (Williams, 1975; Powell, 1990; Creed & Miles, 1996). Within organizations, trust has been related to open communication (Curral & Judge, 1995; Smith & Barclay, 1997), acceptance of influence (Blau, 1969; Smith & Barclay, 1997; Tyler & Degoey, 1996), attribution of positive motives (Kramer, 1996), forbearance from opportunism (Smith & Barclay, 1997), mutual learning (Nonaka & Takeuchi, 1995; Boisot, 1995; BijlsmaFrankema, Prins & Weber, 1999; Janowicz & Noorderhaven, 2002), and to positive outcomes such as high levels of cooperation and performance (Morgan & Hunt, 1984; Gambetta, 1988; Costa, Roe & Taillieu, 2001). These studies support the theoretical idea that trust lubricates a wide array of organizational processes. However, insights in behavioral cues that people use to trust or distrust others are less developed. For Antecedents of Trust in Managers 3 instance, more is known about the beneficial effects of trust in managers than about what managers can do or refrain from to earn trust of subordinates. Some general insights have been produced that can be worked upon. In studies of subordinates’ trust in managers, Kramer (1996) and Tyler and Degoey (1996) conclude that relational issues have a stronger impact on subordinates’ trust in managers than task-focused issues. If subordinates feel they are treated fairly, with respect and with dignity, they perceive their manager as benevolent and thus trustworthy. Yet, respect, dignity, and fairness are quite general terms that do not reveal much about what managers can do to generate these experiences. In a recently published meta-analysis of trust in leadership by Dirks and Ferrin (2002), a first systematic review of empirical evidence for antecedents of trust is presented. Trust in leadership is found to be significantly related to transformational leadership, perceived organizational support, interactional justice, procedural justice, transactional leadership, distributive justice, participative decision making, and meeting expectations of followers (Dirks & Ferrin, 2002, p. 622). These complex constructs, however, are measured as multiitem variables. Therefore, they cannot serve directly as antecedents of trust without violating the widely held assumption that bounded rationality makes people focus on a few single cues instead of many complex ones (Simon, 1955; March, 1978). So, next steps must be taken to find the behavioral components responsible for the effect on trust in leaders. Dirks and Ferrin (2002, p. 622) also conclude that further research is needed to “examine the behavioral cues that employees use to draw conclusions about the character of the leader or whether the relation is one involving care and concern.” So, there should be searched for a Antecedents of Trust in Managers 4 restricted set of simple managerial behaviors that trigger trust in managers. Therefore, the present study was aimed to answer the following question: Which behaviors (or sets of behaviors) of managers are related to trust in managers in the perspectives of organizational members? The data were taken from a case study of a general hospital in change from a divisional form to a network form of organizing. The transition to a form with higher trust requirements, and the conditions of change, heighten the relevance of trust in intra-organizational relations (Gilkey, 1991; Creed & Miles, 1996; Lewicki & Bunker, 1996; Mishra, 1996; Bijlsma-Frankema, 2001a, 2001b). Methodologically, the study is in line with recommendations for future research on trust in managers, made by Kramer (1996), who argues that there is a pressing need for more ‘naive theories’ about trust, that are based on mental accounts of people studied. Naive theories are theories that ”individuals, conceptualized as lay epistemologists, carry around inside their heads. … As such, naive theories presumably play a central role in their attempts to retrospectively make sense of, and learn from their experiences” (p. 238). Survey data alone are not fit to arrive at naive theories because of the framing of questions and answers. “Even though the data from such surveys may reveal quite adequately how individuals weight and prioritize among the variables that have been selected by the researcher, they may reveal very little about the naturally occurring set of categories, dimensions or variables that individuals would spontaneously find salient or invoke” (p. 239). Following Baumeister et al. (1990, p. 995), Kramer (1996) advocates the use of autobiographical narratives and searching for mental accounts, that is “people’s actual accounts of genuine events from their everyday lives” because of the high external validity this method can produce. Since a Antecedents of Trust in Managers 5 weakness of this method is that internal validity can be low due to researcher effects on data gathering and interpretation, triangulation of narrative data and survey data is proposed to reduce this problem (p. 239). Kramer’s argument adequately balances two meanings of validity that seldom both are given full attention within a research project, as Smaling (1992) argues. The first meaning, mostly stressed in quantitative research is avoiding biased and unbiased measurement errors. The second, mostly used in qualitative research is whether the researcher’s representation of the empirical world respects the empirical world under study. As Blumer (1969, p. 22) notes: “The problems set for study need to be critically studied to see whether they are genuine problems in the empirical world; the concepts used and the data chosen need to be inspected to see if in fact they have in the empirical world the character given to them in the study.” To strike a balance between both meanings of validity, a multitriangulation approach was used. In the first phase of our study, following Blumer, an interpretative approach was chosen. In this way a weakness of survey research, the hard to measure and hard to control ‘noise,’ created by answers given to questions experienced as irrelevant by respondents, can be curbed. Topic guided interviews were conducted, based on Kvale’s (1996) interview methodology. The kernel of this method is to ‘pose short questions and getting long answers,’ to ge t as close to the concepts, meanings and relations between phenomena that respondents use to make sense of their experiences and, subsequently, to build their opinions on. Analysis of the interview material resulted in seven hypotheses, each relating a single managerial behavior to trust in managers. These hypotheses were tested in a survey research and administered to the population of organizational Antecedents of Trust in Managers 6 members. In triangulating the interview and survey data, low validity of the interview data due to researcher effects come to the fore if different conclusions must be drawn from the survey data. Besides, two different methods of analysis were employed to analyze the survey data, a regression analysis and a Boolean algebra based pattern analysis, each built on different assumptions and related error threats, adding another possibility of triangulating data. The principle of Boolean analysis is to look for patterns of independent variables that discriminate between outcomes in the dependent variable (Ragin, 1989, 1994). Triangulation was employed to strive after robustness of findings, because, as Webb, Campbell, Schwartz and Sechrest (in De Jong-Gierveld & Van Tilburg,