Understanding non-parametric tests is crucial in psychological research, especially when data doesn't meet parametric assumptions. These tests, including the Wilcoxon signed-rank and Mann-Whitney U tests, are robust against outliers and suitable for small samples. They are essential for valid statistical analysis in research with non-normal distributions or when dealing with categorical data.
Show More
Non-parametric tests are statistical techniques used in psychological research when data does not meet the assumptions necessary for parametric tests
Normal Distribution of Data
Parametric tests assume that data follows a normal distribution
Homogeneity of Variance
Parametric tests assume that the variance of data is equal across groups
Independence of Observations
Parametric tests assume that observations are independent from each other
Non-parametric tests are ideal for analyzing categorical data, handling outliers, and managing small sample sizes
Non-parametric tests use a ranking system for data points instead of the actual data values
Reference Value
Data points greater than the reference value are marked with a '+' while those less than the reference value receive a '-'
Impact of Outliers
The ranking method diminishes the impact of outliers in the data
Facilitation of Analysis
The ranking method facilitates the analysis of data that does not fit the specific distributional criteria required for parametric tests
The Wilcoxon signed-rank test is used to compare two related samples
The Mann-Whitney U test is used to compare two independent samples
The Spearman rank-order correlation assesses the strength and direction of association between two ranked variables
The Kruskal-Wallis H test is used to compare the distributions of more than two independent groups
The Friedman test is used to compare more than two related groups
Non-parametric tests typically do not provide estimates of parameters such as effect sizes or confidence intervals
Non-parametric tests may have a higher risk of Type I errors due to their focus on medians rather than means