A cognitive dissonance interpretation of consistencies and inconsistencies in environmentally responsible behavior

https://doi.org/10.1016/S0272-4944(03)00039-2Get rights and content

Abstract

The issue of consistency and inconsistency in environmentally responsible behavior (ERB), as reflected in the correlations between different ERBs, is discussed in the light of social-psychological theories suggesting that most people desire to behave consistently. It is argued that except in cases where different ERBs are substitutes or at least one of them is totally determined by idiosyncratic conditions the desire to behave consistently should lead to ERBs being positively correlated. However, the correlation may be attenuated by the influence of idiosyncratic conditions (considered “noise” in this context) and measurement error and it is moderated by perceived dissimilarity between the behaviors and by the (moral) importance of behaving in a responsible way towards the environment. These propositions are tested (and confirmed) by means of a mall-intercept survey of ordinary Danish shoppers. Implications for the promotion of a generalized ERB pattern are discussed.

Introduction

In the (relatively short) history of research into environmentally responsible behavior (ERB) one of the persistent, yet still not completely settled, questions has been whether different ERBs are—or should be expected to be—correlated (e.g. Gray, 1985; Stern & Oskamp, 1987; Pickett, Kangun, & Grove, 1993; Berger, 1997; Diekmann & Preisendörfer, 1998; Thøgersen, 1999b). The international literature contains studies that report positive correlations between people's propensities to behave in an environmentally friendly way across domains (e.g. Berger, 1997) as well as studies that fail to find such correlations (e.g. Stern & Oskamp, 1987; De Kruijk & van Raaij, 1991). In particular, the early studies found low correlations between ERBs. Hence, in the mid-1980s a thorough reviewer of the literature had to conclude that “there is very little evidence of transfer of heightened pro-ecological behavior from the target behavior to other behaviors” (Gray, 1985, p. 188). Based on the same evidence others inferred that “environmentally beneficial choices…are not reflective of a general conservation stance, but are instead made on an activity-to-activity basis” (Pickett et al., 1993).

The two latter references also illustrate the main reasons why researchers have taken interest in the correlation between different ERB. Positive correlations between different ERBs should be expected if they share motivational roots (“a general conservation stance”) and a lack of correlation is taken to indicate that each ERB is considered in isolation and that behavior is completely determined by the specificities of each behavior and situation. Which of these possibilities is true has important implications for policy aimed at promoting ERB, of course.

Positive correlations between different ERBs should also be expected if “heightened pro-ecological behavior” is transferred from “the target behavior to other behaviors” or, as others have expressed it, if there is a tendency for ERB to “spill over” into other behavioral domains (Frey, 1993; Thøgersen, 1999b). Spillover of ERB would mean that the effects of interventions promoting specific ERBs are multiplied for the benefit of a more sustainable society overall. Hence, the practical importance of this issue should not be underestimated either. The two reasons for expecting different ERBs to be correlated are hardly independent. Most likely, a tendency for ERB to spill over depends on ERB in different areas having shared motivational roots.

Even those most strongly convinced that, in general, ERBs should be positively correlated will hardly claim this to be universally true. An obvious exception is ERBs that are substitutes,1 such as traveling by bicycle and by public transportation. Neither can it be ignored that ERB is determined by personal and contextual constraints, in addition to the individual's motivation to perform the behavior (e.g. Guagnano, Stern, & Dietz, 1995; Ölander & Thøgersen, 1995). Constraints are typically behavior and sometimes even situation specific. Hence, it is also obvious that the correlation between any two ERBs will be attenuated to the degree that either of them is influenced by idiosyncratic constraints. In the context of the present discussion, attenuation due to the influence of idiosyncratic constraints is best considered noise, which may obscure possible common sources of motivation or the propensity for ERB to spill over into other behaviors.

Another common source of attenuation of correlations between ERBs is the use of fallible measures (typically self-reports). Correlations based on fallible measures usually underestimate the “true” correlations due to uncorrelated errors of measurement (Bagozzi, 1994; Mackenzie, 2001). When attenuation due to measurement error is combined with attenuation due to the influence of idiosyncratic constraints, theoretically meaningful correlations between behaviors may easily disappear, especially when the sample size is small.

Hence, in order to achieve a satisfactory answer to the question whether ERBs tend to be—and should be expected to be—positively correlated it is necessary to explicitly formulate the limiting conditions for this proposition. Important limiting conditions have been suggested above: the ERBs should not be substitutes, not totally determined by idiosyncratic constraints, and measurement error should be accounted for. In addition, the motivational forces assumed to produce a positive correlation between ERBs should be considered.

The expectation that different ERBs should be positively correlated is based on (or receive theoretical backing from) social-psychological theories proposing that most people have a strong desire to be consistent in their attitudes, beliefs, words, and behaviors (Feldman, 1966; Cialdini, 1989; Bator & Cialdini, 2000; for an extensive review, see Eagly & Chaiken, 1993). One of the most prominent among these theories is Festinger's (1957) theory of cognitive dissonance, which claims that we have a need to avoid inconsistencies in our beliefs, attitudes, and behaviors.

Backed by experimental evidence (e.g. Zanna & Cooper, 1974; Zanna, Higgins, & Taves, 1976; Elliot & Devine, 1994), cognitive dissonance theory suggests that it is emotionally disturbing and, hence, unpleasant to hold inconsistent attitudes and/or perceptions or to behave inconsistently (Festinger, 1957). Therefore, the desire to avoid cognitive dissonance is assumed to create a drive to behave consistently (and to hold consistent beliefs about oneself in general). Experimental studies have demonstrated that cognitive dissonance can produce environment-friendly adjustments in behavior (e.g. Kantola, Syme, & Campbell, 1984; Dickerson, Thibodeau, & Aronson, 1992; Aitken, McMahon, Wearing, & Finlayson, 1994). In these cases, dissonance was produced by increasing the salience of the discrepancy between a person's behavior and his or her norms for environmentally proper behavior, i.e. by inducing a sense of “hypocrisy” (Dickerson et al., 1992). However, whether this drive has the power to spontaneously align different ERBs is a matter not yet investigated empirically, as far as I know.

If two volitional behaviors are believed to lead to the same super-ordinate goal, a desire to avoid cognitive dissonance should create a drive to perform either both or neither of them.2 Obviously, what matters is how the actors themselves, not some outside observer, perceive the two behaviors. If actors fail to see that the two behaviors are linked to a common goal, they may not experience the slightest inconsistency when performing one and not the other.

In addition, the individual may perceive that the two behaviors differ in the level of freedom to choose. Experimental research documents that cognitive dissonance is not produced if inconsistency can be attributed to external forces (e.g. Festinger & Carlsmith, 1959; Collins & Hoyt, 1972). Specifically, environmentally relevant behaviors may differ in many ways, for instance with regard to the specific physical acts performed, where and when they are performed, the amount of effort and/or resources expended, and/or the specific outcomes obtained. Whether or not these differences translate into different degrees of external determination and, hence, are legitimate reasons for behaving differently (viewed from the perspective of a super-ordinate goal) is to a high degree a subjective question. Hence, the amount of dissonance produced by acting inconsistently not only depends on the individual's perception of the behaviors’ relationship to the super-ordinate goal, but also on the level of his/her perceived volitional control over the behavior (i.e. whether the reason for performing a behavior is attributed to intrinsic motivation or external forces) (cf. Festinger, 1957).

Hence, the following hypotheses about correlations between ERBs can be derived from cognitive dissonance theory:

H1.1: Volitional ERBs are positively correlated because people strive to be consistent in their performance or nonperformance of behaviors that are connected by a common association to a super-ordinate goal, such as environmental protection.

However, due to lack of clarity of the connection between behaviors and super-ordinate goals, as well as different degrees of external determination, the individual does not necessarily perceive it to be inconsistent to behave in a way that furthers a super-ordinate goal in one behavioral domain while behaving in a way that does not, or perhaps even goes in the opposite direction, in another domain. Hence:

H1.2: The correlation between ERBs depends on how similar or dissimilar they are perceived to be in terms of relationship to super-ordinate goal and external determination because individuals will strive harder to perform two such activities consistently the more similar they perceive them to be in these respects.

Not all inconsistencies are assumed to be equally disturbing and, hence, equally likely to produce cognitive dissonance. Festinger (1957) was not very precise in specifying possible sources of variation in the amount of dissonance produced by inconsistency, but later research has attempted to strengthen the theory on this point. The most influential of these attempts is Aronson's (1997; Aronson & Carlsmith, 1962) suggestion that the feeling of cognitive dissonance depends on whether important elements of ones self-concept (i.e. being competent or possessing high moral standards) are being threatened. Specifically, the desire to perceive oneself as a moral person is regarded a key concern in the self-concept explanation of dissonance (Dickerson et al., 1992; Thibodeau & Aronson, 1992).

It has repeatedly been demonstrated that moral standards (or norms) play an important role in guiding ERBs (e.g. Heberlein, 1972; Black, Stern, & Elworth, 1985; Stern, Dietz, & Black, 1986; Hopper & Nielsen, 1991; Thøgersen, 1999a). These empirical results in combination with the self-concept revision of dissonance theory suggest that the importance of consistency in the environmental field, and, hence, the amount of dissonance produced by behaving inconsistently, depends on the person's moral standards for ERBs. Behavioral inconsistency threatens the individual's self-perception as a morally reliable person only if he or she holds moral norms of some strength for this type of behavior. Hence, for individuals who have no or only weak moral norms for environmentally relevant behaviors it matters little if they perceive their behavior in this domain to be inconsistent. However, for individuals with strong moral norms the desire to avoid cognitive dissonance creates a drive to behave consistently. Specifically we may hypothesize that:

H2: The influence of perceived similarity on the correlation between ERBs (cf. H1.2) depends on how morally important it is for the person to behave in an environmentally friendly way because moral importance influences the amount of dissonance produced by acting inconsistently in the environmental field.

Notice that what is proposed here is an interaction between moral norms and similarity perceptions in determining the correlation between pairs of ERBs. The effect of similarity perceptions on the correlation between behaviors should be stronger among individuals with strong than among those with weak moral norms, and the effect of moral norms on the correlation should be stronger for behaviors that are perceived as relatively similar. In fact, direct effects of the two variables should only be expected if most people in the analysed population share norms for the behaviors of some importance (allowing for a direct effect of similarity perceptions) and/or if most people perceive at least some similarity between the behaviors (allowing for a direct effect of moral norms).

Section snippets

Method

The hypotheses are tested by means of a survey of 309 ordinary consumers, contacted in two shopping malls in Aarhus, Denmark, in June 1998. No log was made of the number of individuals refuting to participate, but since they refused without knowing the subject of the study3 the nonresponse is not likely to cause biases in the hypothesis tests. Of course,

Data analysis

First, the hypotheses are tested by means of simple regression analyses using SPSS10. Next, the hypothesis tests are repeated on a subset of the data by means of multigroup confirmatory factor analysis (CFA) using Lisrel 8.53 (Jöreskog, 1993; Jöreskog, Sörbom, Toit, & Toit, 1999; du Toit, du Toit, & Hawkins, 2001). The main advantage of this procedure is that it allows one to explicitly account for measurement error,8

Regressing correlations between behaviors on similarity assessment

Hypothesis 1.1 states that, in general, voluntary ERBs are positively correlated and Hypothesis 1.2 that the correlation between ERBs depends on how similar or dissimilar they are perceived to be. These propositions can be tested by inspecting pairwise correlations between the behaviors listed in Table 2 and by regressing the pairwise correlations on the similarity ratings of the same pairs of behavior. The regression plot is shown in Fig. 1.

Notice first that the pairwise correlations vary from

Summary and implications

This paper deals with the issue of consistency and inconsistency in ERB as reflected in the correlations between different ERBs. Social-psychological theories, including cognitive dissonance theory, suggest that most people desire to behave consistently. This desire should lead to ERBs being positively correlated, except in cases where different ERBs are substitutes or at least one of them is totally determined by idiosyncratic conditions. However, it is argued here that correlations may be

Acknowledgements

This paper is the outcome of a collaborative research project with Folke Ölander, Aarhus School of Business, in the framework of Center for Social Science Research on the Environment (CeSaM), a multidisciplinary center based on collaboration among scholars from a number of Danish unversities and other research institutions. I am grateful to Tommy Gärling, Berit T. Møller, Folke Ölander, and Joachim Scholderer for helpful comments on earlier drafts.

References (55)

  • J.L Arbuckle

    Full information estimation in the presence of incomplete data

  • E Aronson

    The theory of cognitive dissonanceThe evolution and vicissitudes of an idea

  • E Aronson et al.

    Performance expectancy as a determinant of actual performance

    Journal of Abnormal and Social Psychology

    (1962)
  • R.P Bagozzi

    Structural equation models in marketing researchBasic principles

  • R.J Bator et al.

    New ways to promote proenvironmental behaviorThe application of persuasion theory to the development of effective proenvironmental public service announcements

    Journal of Social Issues

    (2000)
  • I.E Berger

    The demographics of recycling and the structure of environmental behavior

    Environment and Behavior

    (1997)
  • J.S Black et al.

    Personal and contextual influences on household energy adaptations

    Journal of Applied Psychology

    (1985)
  • M.W Browne et al.

    Alternative ways of assessing model fit

  • R.B Cialdini

    Social motivation to complyNorms, values and principles

  • De Kruijk, M.D., & van Raaij, W. F. (1991). Similarities and frequencies of behaviors in an environmental context....
  • C Dickerson et al.

    Using cognitive dissonance to encourage water conservation

    Journal of Applied Social Psychology

    (1992)
  • A Diekmann et al.

    Environmental behaviorDiscrepancies between aspirations and reality

    Rationality and Society

    (1998)
  • Du Toit, M., du Toit, S., & Hawkins, D. (2001). Interactive LISREL: User's guide. Chicago: Scientific Software...
  • A.H Eagly et al.

    The psychology of attitudes

    (1993)
  • A.J Elliot et al.

    On the motivational nature of cognitive dissonanceDissonance as psychological discomfort

    Journal of Personality and Social Psychology

    (1994)
  • C.K Enders

    The impact of nonnormality on full information maximum-likelihood estimation for structural equation models with missing data

    Psychological Methods

    (2001)
  • Feldman, S. (Ed.). (1966). Cognitive consistency motivational antecedents and behavioral consequents. New York:...
  • Cited by (293)

    • Sweet to bitter: Dissonance of scuba diving experience

      2023, Tourism Management Perspectives
    View all citing articles on Scopus
    View full text