CPP 4402 Lecture Notes - Lecture 3: Chi-Squared Test, Contingency Table, Level Of Measurement
Document Summary
When data is from a sample that is normally distributed. Student"s t-test - to compare 2 independent groups. Paired student"s t-test to compare 2 related groups (for matched or paired data) Can not tell you which of the groups are different. Tells you that there is a statistically significant difference between 2 or more groups. When data is from a sample that deviates from normal. Data must be ordinal, nominal or skewed continuous. Can not be used for paired data. When chi square test can"t be used. Aka when number of individuals in one cell of the contingency table is <5. Mcnemar"s test - for nominal data with 2 samples. Mantel-haenszel test - not on the table so don"t worry about it. Wilcoxon test - nonp version of paired t. Kolmogrorov-smirnov - not on the table so don"t worry about it. When 3+ groups need to be compared.