Contact RJ

Statistical Test:

What it Determines:
(With DV & IV Tips)

Explanations:

ANOVA/Factorial

Interaction i.e. are differences similar over time?
(1 DV, ³2 IV)

·          If no significant interaction, then look at “main effects”

·          Perform Post Hoc to determine where significant differences occur

ANOVA/One-Way

Compares ³2 means i.e. differences among group means and F value
(1 DV, 1 IV ³2 levels)

·         Do Post Hoc after ANOVA

·         Could also use a T-Test

·         Pre-test value wasn’t equivalent so need to adjust to make equivalent with post-test

·         Means random assigned & not related & unpaired

ANOVA/ Repeated Measures

Compares ³3 means

·           

Correlation

“Relationship” between two variables

·         Look for key word “relationship” in description for a clue

·         Tells “group” mean

·         p value tells if relationship is significant

·         r value tells strength of relationship

Descriptive Statistics

Mean, SD, Variance, etc.

·          Check off appropriate options that are relevant

Multiple Regression

Predicts Y from several X values
(1 DV, >1 IV)

·           

Paired T-Test

Also measures before & after

·          If not sure about direction, use two-tail

·          If you choose one-tail, must justify.

Post Hoc

Where the means occur and how significant they are

·          Use after the ANOVA

Regression Analysis

“Predicts” one variable from another
(1 DV, 1 IV & zero levels)

·          Tip: If a tests says it wants to “predict” then look at this & Multiple Regression.

·          Get a Y intercept i.e. y=mx+b

·          To use regression equation, correlation must be significant!  If X & Y not related, can’t predict one from the other.

Repeated T-Test

·          Must use Bonferroni Adjustment to compensate for multiple tests

Scattergram

Y value if you have X

·          Graph plotting X against Y

·          Find X, then match with Y, repeat for each X value

·          Estimate + line of best fit but still have error

T-Test
(General Definition)

Difference between two means

·          Measures two groups

·          Gives positive value

·          p value tells level of significance

Unpaired T-Test
(» Independent Samples T-Test)

Compares two different populations that don’t have correlated data.
(1 DV, 1 IV with 2 levels)

·          Example is two different schools