ABSTRACT
The Kruskal-Wallis test performs an analysis that is very similar to an analysis of variance on the ranks. The test is performed when the assumptions of the parametric ANOVA cannot be made. An example of such data occurs in the following illustration. A researcher was interested in differences in attractiveness and the selection of candidates for jobs. As well as examining female attractiveness a number of experiments were undertaken on male attractiveness. One of the questions considered was whether different types of facial hair led to different judgements of male attractiveness by women. A female personnel officer in a large company agreed to rate photographs of men’s faces on attractiveness on a 0 to 50 scale, with a high value indicating a high level of attractiveness. Out of a large pool of photographs of different men, five men with beards, five men with moustaches and five clean shaven men were randomly selected. (The photographs in the pool had been matched on age, hairstyle and tidiness.) If we examine the data below can we observe an effect of facial hair on the attractiveness judgements?
