Exam 3 Advanced Statistics

Independent variable

Fully manipulated and randomly distributed across participants

Quasi-Independent

Variable is not manipulated

Subject variables

Measured characteristics of a participant…pre-existing characteristic

ANOVA tests

Looks for differences “between” groups

F is

the test statistic for ANOVA

In ANOVA, we reject the null when

The obtained F is greater than or equal to critical F and/or if p is less than or equal to 0.05

As test statistic (F) gets larger,

p gets smaller

Conceptually F must be ___ than 1 to be significant

Greater

X… with a line

Grand mean

i

row

j

Column

A

group

R

location in group

Assumptions of ANOVA

1. Homogeneity of variance

2. Normality

3. Independence of observation

Independence of observation

scores are independent of each other, every observation is unique and no score will effect another score (very important)

Levene’s Test

Does ANOVA on variances, is p = or <.05 reject null

Box’s Test

Corrects for violation, adjusts degrees of freedom down

Trimmed sample

Removing outliers (commonly 10% from each tail)

Winsorized sample

Replacing outliers with closest, less-extreme values

Fixed model ANOVA

Control treatment conditions involved

Random model ANOVA

No control of treatment conditions involved

Bonferonni’s Test

Follow up test/multiple comparison procedure for a significant F. Alpha/# of tests

A priori (planned)

rewards you for knowing which comparisons to run

Post hoc (unplanned)

no plan; assumes all comparisons, take a penalty for that (Tukey’s is most common)

Games-Howel

A post hoc test done to adjust for lack of homogeneity of variance

Conceptual Framework

1. Identify problem

2. Formulate hypothesis

3. Identify conceptual variables

4. Operationalize

Between-Groups

Examine differences in behavior between groups of participants

Within-Groups

AKA Repeated measures, Examine differences in each participant’s performance across each variable