Lecture 8 - Comparing 2 Groups: t-tests and non-parametric equivalent tests

A collection of vocabulary flashcards based on concepts related to t-tests and non-parametric equivalent tests as discussed in Lecture 8.

t-tests

Statistical tests used to compare the means of two groups.

Independent samples t-test

• Compares the means of two groups that are independent of each other (between-group design). groups do no influence each other’s scores on dependent variable

• variances of dependent variable in two groups are equal

• independent variable is dichotomous and levels are paired or matched

• assumes normal distributions

• no need to worry about equal variances

Paired samples t-test

• Compares means from the same group at different times (within group design).

• either the same group is tested at two different times, or there are matched cases (related, like if it was spouses) → two different groups that are identical or not independent

• level of measurement of DV is nominal/dichotomous

Mann-Whitney U test

• A non-parametric test that compares the distributions of two independent groups (between).

• used when DV level of measurement is ordinal

Wilcoxon signed-rank test

• non-parametric equivalent of paired t-test

• A non-parametric test used to compare medians (or mean ranks) of two related samples (within).

  • when DV has ordinal-level of measurement

• null hypothesis is that medians of two related groups are equal 

• can also be used in scale situations where assumption of normality is violated 

Chi-square test

• A statistical test used to determine if there is a significant association between categorical variables. non-parametric

• non-parametric equivalents to independent samples t-test when measure of DV is nominal

McNemar test

• A non-parametric test for paired nominal data. (within) - non-parametric equivalent of paired t-test for categorical (two-times repeated) data

• compared proportions of two related groups

• used when comparing two related groups on DV w dichotomous level of measurement

• tests will show significance of .000 → this is NOT zero. this is actually <0.001

Levene's test

• Statistical test for homogeneity of variances.

• null hypothesis for this is that there is no difference in variance

• if Levene’s (F) test is NOT significant (sig > 0.05), then use the equal variances assumed line for the t-test

• if the test IS significant (sig < 0.05), then null hypothesis is rejected, and you use the equal variances not assumed line for the t-test

Confidence Interval (CI)

A range of values that is likely to contain the population parameter.

sampling distribution

• set of different scores (or statistics) that results from repeated replications of same study using different samples of same size from same population

T-score

• like z-scores

• tell us how many standard deviations the difference between sample means is from the mean of sampling distribution, and in which direction