For my research I'm examining whether applying a certain promotional intervention improves the amount of misshapen products bought. The study contains of two parts. In part 1 Scanner data from supermarkets will be taken collected from each intervention (1,2,3,4). So in the end you end up with 4 counts:
1 - The amount of products bought of normal shape with intervention
2- The amount of products bought of normal shape without intervention
3- The amount of products bought of misshapen products with intervention
4- The amount of products bought of misshapen products without intervention.
And I am interested in measuring whether or not the intervention increased the amounts bought of misshapen products. What statistical analysis applies best here? I have been told ANOVA in some cases, but others argue this might not be the correct one since the variabel is count data and counts can't be negative.
In part 2 the same study will be replicated in a different country. It is hypothesized that the results in both countries will be quite similar. How do I conduct this analysis? Should I replicate the analysis as in study 1 and if the same conclusions can be made on those results conclude that the same mechanism applies in both countries?