Internal Validity in Research

Overview

This lecture reviews the concept of internal validity in research, focusing on common threats and strategies to address them.

Internal validity is how well an experiment rules out alternative explanations for its results.
High internal validity means fewer confounding variables and greater confidence in the causal relationship between variables.

Attrition Threat: Participants dropping out can bias results, especially if loss differs between groups.
Treatment Diffusion: When control group learns about or mimics treatment, group differences are minimized.
Testing Threat: Repeated testing can increase familiarity and performance, unrelated to intervention (pretest sensitization).
History Threat: External events between observations may change participant outcomes, not due to the intervention.
Instrumentation Threat: Changes in measurement tools or procedures over time can affect results.
Selection Threat: Pre-existing differences between groups confound results; comparing non-equivalent groups is problematic.
Maturation Threat: Natural changes in participants (development, fatigue) over time can influence outcomes.
Inequitable Treatments: Researchers' or participants' expectations produce unequal treatment, influencing performance.
Special Treatment Threat: Groups receive different treatment due to their status in the study, not the intervention.
Statistical Regression Threat: Extreme pretest scores tend to move toward the mean upon retesting, independent of treatment.
Interaction with Selection: Multiple threats can combine, compounding effects on validity.
Ambiguous Directionality: Unclear causality when temporal order between variables isn't established.

Random assignment to treatment and control groups helps mitigate most threats.
Isolate groups to reduce diffusion and inequitable treatment.
Consistent measurement tools and procedures avoid instrumentation threats.
Blind and double-blind studies can prevent inequitable and special treatment threats.
Shorten time between pre- and post-tests to lessen maturation and history threats.

Internal Validity — Degree to which a study rules out alternative explanations for its findings.
Confounding Variable — An uncontrolled variable that may affect the dependent variable.
Attrition — Loss of participants from a study over time.
Random Assignment — Randomly placing participants into treatment or control groups to reduce bias.
Single-Blind Study — Participants do not know if they are in the treatment or control group.
Double-Blind Study — Neither participants nor researchers know group assignments.
Regression to the Mean — Tendency for extreme scores to move closer to the average on retesting.

Review the definitions and examples of each threat to internal validity.
Write down and memorize the differences between single-blind and double-blind studies.
Prepare to identify threats and solutions in sample research scenarios for next class.