MCRS Statistical tests

All None Speichern

• Level of measurement one-sample t-test Test value is hypothetical mean. Test variable is continuous
• What it does: One sample T-test Compares sample mean to hypothetical population mean
• Why one-sample t-test Make sure the sample mean is a good fit for hypothetical population
• Assumption one sample t-test Normal distribution
• Run One-sample t-test SPSS Analyze > Compare means > One-sample t-test (Assign test variable, insert hypothetical mean in test value)
• Measure of strength one-sample t-test/Paired/Independent Cohen's d
• Level of measurement paired samples T Test variable 1 & 2 Continuous
• What paired sample T does Compares the means of two scales belonging to same respondent to check if population means are significantly different
• Why paired sample T Compare mean differences within groups in the pop
• Assumption paired sample T Normal distribution, N>100
• Paired samples SPSS Assign variables 1 & 2 antichronologically
• Level of measurement independent ST Grouping X categorical, Test Y continuous
• What independent S-T does Compare means of two groups determined by a grouping variable to see if the population means are significantly different
• Why independent S-T Make sure that there is a significant difference between the means of two groups in the population
• Independent ST Assumption Normal distribution, N>100, Variances in both groups are roughly equal
• Levene's test significant (<.05) Check second row (equal variances not assumed)
• Levene's test not significant Equal variances assumed - check first row
• Independent S-T SPSS Assign test variable(DV) Assign grouping variable(IV)
• OWA Level of measurement Grouping variables (2IV) Categorical. Dependent continuous
• TWA/OWA/Regression Relationship direction Assymetrical
• What OWA does Compares means of more than two groups, F-test sees if there's variance between at least 2 groups. Posthoc looks at differences between groups
• Why OWA? Compare means of two or more groups while testing effects
• Assumptions TWA/OWA Normal distribution, Variances in both groups roughly equal (Homogeniety of variances)
• Observation OWA Only (X) can test model (F-test) and compare means across more than two groups
• Effect size OWA/TWA EtaSquared
• TWA Levels of measurement Grouping variables 1 & 2 categorical, Dependent is continuous
• Why TWA? Adds a categorical moderator for OWA
• Chi-square level of measurements Nominal/nominal, Nominal/Ordinal, Ordinal
• Chi: Symmetric/Nominal Cramer's V (Phi for 2x2 tables)
• Chi: Asymmetric/Nominal Goodman & Kruskals tau
• Chi: Symmetric/Ordinal Gamma
• Chi: Asymmetric/Ordinal Somer's D
• What Chi does Checks if there's any association between two variables
• Why Chi? We have 2 categorical variables
• Chi Assumptions Expected counts of at least 1, Max 20% with expected counts >5, Each item contributes to only one cell (mutually exclusive)
• Regression level of measurement Interval or ratio
• Measure of effect Regression Pearson's r
• What regression does Predicts the value of a variable based on the values of other variables
• Why regression Check if IV can predict a single DV, and to check whether each of these IV's are significant predictors of the DV or not
• Assumption regression Linearity
• Observation regression Slope, R2 as percentage

All None Speichern