7. Limitations and future research
While our research has valuable contributions, it also has some limitations. First, some limitations are associated with the store-intercepted survey research such as measurement error and interviewer effects. Respondents may have been influenced by the presence of interviewers and then distorted in order not to appear impulsive buyers. Another concern is about pre and post-measurement design that, in particular, can lead to premeasurement effects and mortality effects (Mohan, 2013). However, this approach allowed us to capture measures at two points in time, which was critical and appropriate to the study design. Additionally, another limitation of this research is the generalizability. Our sample, interviewed in regional stores, is probably neither truly random nor necessarily representative of any larger population (Beatty and Ferrel, 1998). But, given our interest in relationships between variables rather than population descriptions, this may not be a major problem. Finally, we have decided to focus on few variables and interactions, even if we are aware that other variables could be considered in our model. For example, we did not consider some situational variables like time and money, which were found to influence positive and negative affect and, also, impulse buying (Beatty and Ferrel, 1998; Mohan et al., 2013). It would be interesting to understand which variables affect the pre-shopping preparation tendency in order to include them into the model.