Statistical Test and Power of Movie Rating Data
See the assignment_3.pdf for the full report
This report summarises the statistical modelling and analysis results associated with movie ratings of Adventure, Comedy and Drama genres. The goal of this report is to check if the rating distributions across the genres are statistically similar. The data “movielens” analysed has been taken from dslabs package in R. The study uses Anova and Kruskal-Wallis Rank Sum Test, together with power and test size computation by simulation. The report results in having statistically different rating distributions in the similar to “movielens” simulated data among these movie genres. The simulation suggests to use 1500 samples with the original effect size, for the null hypotheses to be correctly rejected 90% of the time.