Data week 8 part 1

winniesmith2's version from 2017-11-20 10:45


Question Answer
When to use non-parametric testswhen assumptions of parametric tests are not met such as the level of measurement (e.g., interval or ratio data), normal distribution, and homogeneity of variances across groups. AND if it is not possible to correct for problems with the distribution of a data set. They make fewer assumptions about the type of data on which they can be used. Many of these tests use “ranked” data.
Advantages of non-parametric tests - Used with all scales - Easier to compute - Developed originally before wide computer use - Make fewer assumptions - Need not involve population parameters - Results may be as exact as parametric procedures
Disadvantages of non-parametric tests-May waste information (If data does not permit using parametric procedures. Example: converting data from Ratio to Ordinal Scale). -Requires a larger sample size than the corresponding parametric test in order to achieve the same power. -Difficult to compute by hand for large samples. -Statistics tables are not readily available.
when to use non-parametric tests most commonly when the data is skewed.
Differences Between Several Independent Groups The kruskal-wallis H test
What is the kruskal-wallis H test is the non-parametric counterpart of the one-way independent ANOVA.
The theory of the Kruskal-Wallis H test is very similar to that of the Mann-Whitney U and Wilcoxon test, This is a ‘between subjects’ analysis. It allows you to compare the scores for three or more groups – omnibus test. Scores are converted to ranks and the mean rank for each group is compared. Statistically significant group differences are evaluated based these ranked sums.
Assumptions for the Kruskal-Wallis H test -All groups have the same population distribution e.g. all skewed to left or all to the right -Data measured on any scale -Samples must be random. -Independent observations
What does independent observations mean-Each participant can be counted only once, -participants cannot appear in more than one category or group, -the data from one participant cannot influence the data from another
Effect size calculationcalculate by hand. r= z/ square root n