Tests of Differences

Question 1

How do we choose our tests of differences (4 questions)?

Answer 1

1. Is the level of outcome measurement continuous or categorical
2. Is the data normally distributed with equal variances across groups?
3. How many groups are being compared?
4. Is the collected data independent or matched?

Question 2

If the analysis mimics a normal distribution curve, what two tests of normality might we run?

Answer 2

Shapiro Wilkes Test
Kolmogorov - Smirnov Test

Question 3

What test might we run if there is positive or negative skewness?

Answer 3

Fisher’s Kurtosis
Positive skewness - leptokurtic
negative skewness - platykurtic
normal distribution - mesokurtic

Question 4

Error variance is:

Answer 4

the variability in data points unexplained by the trx/ manipulation;

Question 5

The larger the error variance, the ________ (more/less) impacting the scores

Answer 5

more

Question 6

What data can be included in parametric data?

Answer 6

Interval and ratio (continuous)
normal distribution of data in both groups
roughly equal variance of both groups.

Question 7

What data is included in non-parametric data?

Answer 7

continuous data with non-normal distribution or variance
ordinal data
rank data (allows for the evaluation of hypotheses for entire population distributions).

Question 8

What test evaluates whether 2 means collected from 2 independent samples differ significantly?

Answer 8

Independent T-test

Question 9

What test evaluates whether 2 means collected from the same sample differ significantly?

Answer 9

Paired T-test

Question 10

What test do we use to see if 3 or more group means differ?

Answer 10

ANOVA

Question 11

What test d do we use if data has been measured repeatedly on the same subjects?

Answer 11

RM ANOVA

Question 12

What type of ANOVA allows researchers to explore the influence of 2 or more independent variables on a single dependent variable? (ex: 2x2 ANOVA)

Answer 12

a factorial ANOVA

Question 12

What type of ANOVA allows researchers to explore the influence of 2 or more independent variables on a single dependent variable? (ex: 2x2 ANOVA)

Answer 12

a factorial ANOVA

Question 13

What type of ANOVA is used when a continuous outcome has been measured w multiple groups repeatedly over time? (Ex: 2 points in time, 2 groups)

Answer 13

Mixed ANOVA

Question 14

What type of ANOVA is used when a continuous outcome has been measured with a confounding variable for the purpose of adjusting for the confounder?

Answer 14

Analysis of Covariance (ANCOVA)

Question 15

ANCOVAs are commonly used:

Answer 15

in experimental designs to control for factors that cannot be randomized but which can be measured on an interval scale
and
in observational designs, to remove the effects of the variables that modify the relationship of the categorical independents to the interval dependent.

Question 16

What are the 4 limitations to non - parametric tests?

Answer 16

1. Hold less statistical power than parametric tests on parametric data
2. May not be directly testing the null hypothesis of interest
3. Limited choices available for testing
4. Cannot be related to the confidence interval

Question 17

Non - Parametric substitution for Independent T - Test

Answer 17

Mann Whitney U Test

Question 18

Non - Parametric substitution for Paired T - Test

Answer 18

Wilcoxon Test

Question 19

Non - Parametric substitution for ANOVA

Answer 19

Kruskall Wallace Test

Question 20

Non - Parametric substitution for RM ANOVA

Answer 20

Freidman Test