26 The Urge to Splurge We conducted a series of three field experiments (Experiments 1, 2a, and 2b) to test the hypotheses. All experiments have a one-factor between-subject design with ‘level of interactivity’ (low versus high) as the independent variable and ‘impulse-visit urge’ and ‘impulse-buying urge’ as dependent variables. We tested the hypotheses at a significance level of .05. The experiments have a total sample size of 436 consumers. The required sample size was calculated a priori with G*Power in the F tests family. The expected effect size (f2) we used to calculate the sample size was .15 since previous research found an average effect size of .17 when measuring the effect of marketing stimuli on impulse-buying behavior (Iyer et al., 2020). As there is just one predictor in our studies (interactivity), a total sample size of at least 89 participants per study was needed according to G*Power tool, a requirement that we have met in all experiments. The first experiment is both confirmative and explorative in nature since some additional exploratory variables were examined next to the hypotheses testing (see Experiment 1, ‘Measurements’). Both the second experiment and third experiment are solely confirmative in nature and were performed to check whether the results of Experiment 1 could be replicated. Replication is often undervalued, although of great importance for gaining confidence in the found results (McEwan et al., 2018). The data for Experiment 1 were collected in one week, the data for Experiments 2a and 2b were collected in two weeks each. All three experiments had an approximately equal number of participants per day, except Sundays, when the shop was closed. All experiments included a manipulation check. Below we describe the method, results, and discussion for each experiment. Overview of Studies An interactive screen (42 inches) was placed in the window of a women’s clothing store located in a well-visited, high-end shopping street in Amsterdam, Netherlands. The content and functionalities of the screen were specifically designed for the three experiments. In line with Liu and Shrum’s (2002) conceptualization, screen interactivity was manipulated through active control, two-way communication, and synchronicity. In both conditions, participants had to press play to start interacting with the screen. In the high-interactivity condition, participants could interact with the screen by swiping through the store’s collection and zooming in on the displayed items (active control). Users could also like products by pressing on a ‘thumbs up’ and interact with the screen by leaving a comment. Both were followed by a short reaction displayed on the screen (two-way communication). Additionally, in the high-interactivity condition, the screen responded immediately to the user’s actions (synchronicity). In the low-interactivity condition, participants had fewer interaction options. After pressing play, they could only reverse the order in which items were shown instead of swiping through them. They could not zoom-in, like, or rate any products or the screen itself, and the response time of the screen on the user’s action was delayed by a full second. The two conditions differed in the interactive features only. All other elements, such as content and size, were identical (see image A). Method. Stimulus material. Experiment 1
RkJQdWJsaXNoZXIy MjY0ODMw