Threats to Internal Validity

Threats to Internal Validity

Maturation

History

Regression to the mean

Attrition

Testing effects

Instrumentation effects

Design confounds

COMPARISON GROUPS

PROPER METHODOLOGY

AVOID BY USING

Selection effects

Order effects

Demand characteristics

Experimenter expectancy effects

Placebo effects

Spontaneous change/recovery?

External/historical event? (same for everyone)

Random factors? (can be same/different)

Simple or Selective?

Progressive order effect?

Omit pre-test

Use alternate forms

Use comparison groups

Instrument decay?

Equivalent alternate forms

Counterbalanced versions across pre-test & post-test

Suspicion probes

Psyhological realism

Pilot testing

Unobtrusive measures

Between-groups designs

Statistics!

Manipulate awareness of hypothesis (+ stats)

Cover stories

Rigorous training

Automated procedures & instructions

Keep experimenters blind to hypothesis/condition

Double-blind placebo control experiment

Coders must be consistent

SOLUTIONS:

Use a comparison group

Design an excellent experiment

Keep participants/researchers blind to hypothesis.condition