Common Statistical Test in Research Flashcards
(21 cards)
Compare means of two groups
T-Test
Compare means of more than two
ANOVA
Test independence between categorical variable
Chi-Square test
Measure linear association between two continuous variables
Person Correlation
Predict a continuous dependent variable based on one or more independent variables.
Regression Analysis
Compare distributions of two independent groups.
Mann-Whitney U Test
Compare distributions of more than two independent group.
Kruskal-Wallis Test
Compare distributions of more than two related groups
Wilcoxon Signed-Rank Test
Compare paired proportions or frequencies in a 2x2 contingency table
McNemar’s Test
Compare proportions in a 2x2 contingency table
Fisher’s Exact Test
Assumptions of T-Test
Normally distributed data, equal variances.
(Ex. Compare test scores of two teaching methods).
Assumptions of ANOVA
Normally distributed data, equal variances
(Ex. Compare test scores among three teaching methods.)
Assumptions of Chi-Square Test
Random sample, Large enough sample size. (Ex. Test if there’s an association between gender and preference for tea or coffee.)
Assuptions of Pearson Correlation
Linearity, Homoscedasticity
(Ex. Examine the relationship between age and income.)
Assumptions of Regression Analysis
Linearity. Homoscedasticity, Independence of errors.
(Ex. Predicting house prices based on square footage, number of bedrooms, etc.
Assumptions of Mann-Whitney U Test
Independence of observations (Ex. Compare test scores between two different schools.)
Assumptions of Kruskal-Wallis Test
Independence of observation, Similar distributions. (Ex. Compare performance of different teaching methods across multiple schools
Assumptions of Wilcoxon Signed-Rank Test
Dependent variables should be continuous or ordinal. (Ex. Compare pre-test and post-test scores within a group.)
Assumptions of McNemar’s Test
Binary Data, Dependent Groups
(Ex. Compare the effectiveness of two treatments on a binary outcome.
Assumptions of Fisher’s Exact test
Low sample size (<20), Independence of observations.
(Ex. Assess the association between gender and smoking status in a small sample).
