Null Results

Null Results

Why did you get null results?

Confounding variables are obscuring your effect

Not enough variability between groups

Too much variability within groups

There really was no difference between groups

Produce results

Obscure results

Weak manipulations

Insensitive measures

Floor/ceiling effects

Too much overlap on DV between-groups

Stats for within-groups

(REVERSE CONFOUNDS)

SOURCES OF ERROR VARIANCE

Measurement Error

Valid & reliable measures

Precise measures

Large N (increase power)

Individual Differences

Within-groups

Matched pairs

Situation Noise

Really obvious IVs

Highly sensitive DVs

Large N (increase power)