The data gathered in each cycle were independently analyzed
using the following statistical analyses. First, in order to measure
“Content consumption”,we calculated the percentage of answers in
Q1 and Q2 for the different means of delivery in each cycle.We then
compared these percentages for each of the types of delivery that
were analyzed in each cycle. For “Time spent at the exhibit”, we
assessed the normality of the data (we discovered that it was nonnormally
distributed), so we conducted a non-parametric statistical
test (Wilcoxon Signed-Ranks Test in this case) to analyze significant
differences in time. To measure the “Visitors’ perceived quality of
the experience” we analyzed the percentage of answers given by
the visitors to each of the options in Q3. Additionally, for the visitors’
experience we also ran a t-test and a non-parametric test
(Wilcoxon Signed-Ranks Test in this case) to analyze significant
differences in the ratings of the exhibits provided by the visitors in
their answers to Q4 for the two different conditions that were
studied in the cycle. It is important to note that an extensive study
by de Winter et al. (2010) revealed no loss of statistical power when
conducting the non-parametric Wilcoxon test instead of the parametric
t-test when comparing data from Likert scales.We therefore
followed this suggestion when conducting our study.