Note on the use of social-psychological theories in (travel behavior) research


Social-psychological theories, like the theory of planned behaviour (TPB) or the technology acceptance model (TAM), take prominent place in behavioural science research. These theories generally assume that a limited number of psychological constructs can explain variation in behaviour. For example, the TAM consists of three main constructs that are thought to influence behaviour, namely perceived usefulness, perceived ease of use and intention to use. In a similar fashion, the TPB assumes the existence of four constructs (attitudes, subjective norms, and perceived behavioural control and intention) that are assumed to influence behaviour.

In empirical research, cross-sectional surveys are typically administrated to measure each construct with multiple (standard) questions, as well as the behaviour in question. These data are then used to estimate a structural equation model (SEM) to test the hypotheses that the considered constructs indeed explain variation in behaviour (typically, for each relationship a hypothesis is formulated). Indeed this is generally found to be the case, in fact, much variation is usually explained by the included factors (40-50%). Based on the findings recommendations are typically formulated for policy makers to influence the considered psychological factors in such a way that desirable behavioural outcomes are achieved.

Given the widespread occurrence of studies described above, it seems that relevant scientific knowledge is indeed being accumulated. In this short note, we question this idea and argue against continued research efforts in this direction. We contend that the empirical approach does not lead to scientific progress and (consequently) is practically limited. The arguments presented in this note are not new, they have been reported in previous articles (Ogden, 2003; Benbasat and  Barki, 2007; Sniehotta et al., 2014; Silva, 2007; Weinstein, 2007). Here, we bring some of them together.

Firstly, the empirical approach as described above does not lead to scientific progress. Arguably, scientific progress can only be made when the theories that are being tested can be falsified. This does not seem to be the case for studies that apply social-psychological theories like TPB and TAM. In SEModels, the chi-square test (and relative fit indices) can be used to assess the fit of the model (i.e. how well the model reproduces the observed correlational pattern). But, it is important to recognise that a (good) model fit cannot be used to establish support for the assumed directions of causation. In fact, other model structures -which may not have been explored- could provide equal well, or even better. Moreover, given that it is usually computationally impossible to test all possible model structures (and we have a finite number of data points), it is even likely that the true model structure is different from the one that is deemed best. In other words, no amount of replication studies (which may all report good fit) can provide definitive proof that theories (like the TPB) are indeed valid.

Consequently, the empirical approach is also practically limited. Indeed, if the theory is wrong (which could be the case), the policy recommendations that follow from the results of models based on them, will simply not achieve the desired effects. Consider the application of the TPB in travel behaviour research (or other applications of social-psychological theories). In empirical applications some factors are found to correlate more strongly with the behaviour in question than others. Typically, researchers then recommend that information campaigns should be set up that target those psychological factors that correlate most strongly with behaviour. But what if the processes that give rise to the  correlations are due to entirely different causal processes than assumed by the model? In the best case, no effects may be achieved, in the worst case, adverse effects (if the sign of the estimate implied by the model is opposite to the real effect). While many researchers are aware of the adage correlation is not causation, often they are still drawn into making policy recommendations that assume causal interpretations (Chorus and Kroesen, 2014).

In addition to these fundamental considerations, it may be questioned to what extent conclusions of the kind ‘if people perceive a certain technology as more useful they are more inclined to use it’ are actually helpful in formulating effective policies, as it is still unclear how the technology in question can be made more useful. The same goes for constructs in the TPB. For example, suppose a study finds that perceived behavioural control correlates positively with cycling, it is still unknown how to positively influence this construct by implementing certain policies. Ideally, the effects of possible policies should directly be put to the test in experimental designs.

To conclude, empirical contributions that apply theories like TPB or TAM are of limited value both scientifically and practically, and prevent more viable research efforts. As such the EJTIR policy is to be very hesitant in accepting publications of this kind (or sending them out for review). Only when the contribution extends beyond the typical contribution of applying the theory to a novel behaviour or technology or context, will contributions be considered for publication.


Benbasat, I., & Barki, H. (2007). Quo vadis TAM?. Journal of the association for information systems8(4), 7.

Chorus, C. G., & Kroesen, M. (2014). On the (im-) possibility of deriving transport policy implications from hybrid choice models. Transport Policy36, 217-222.

Ogden, J. (2003). Some problems with social cognition models: a pragmatic and conceptual analysis. Health psychology, 22(4), 424.

Silva, L. (2007). Post-positivist review of technology acceptance model. Journal of the Association for Information Systems, 8(4), 11.

Sniehotta, F. F., Presseau, J., & Araújo-Soares, V. (2014). Time to retire the theory of planned behaviour. Health psychology review, 8(1), 1-7.

Weinstein, N. D. (2007). Misleading tests of health behavior theories. Annals of Behavioral Medicine33(1), 1-10.