Threats to Internal and External Validity

1) Internal Validity

• Definition: The extent to which the observed changes in the dependent variable are actually due to the independent variable and not because of other (extraneous) factors.
• Sine qua non (essential condition) of research design is Internal Validity (Campbell & Stanley).

Threats to Internal Validity (as per Kerlinger, Campbell & Stanley):

• History – External events occurring during the experiment influence the results. e.g., A drought during a fertilizer adoption study affects yield.
• Maturation – Natural changes in participants (biological, psychological, aging, fatigue). e.g., Farmers improve farming skills naturally with experience.
• Testing Effect – Effect of taking a pre-test influences the post-test performance. e.g., Farmers scoring better in a second test just because they saw similar questions earlier.
• Instrumentation – Changes in measuring instruments or observers lead to bias. e.g., Using different extension workers with different rating methods.
• Statistical Regression – Extreme scores tend to regress toward the mean when retested.
• Selection Bias – Non-equivalence of experimental and control groups. e.g., Progressive farmers choosing to join training programs.
• Experimental Mortality (Attrition) – Dropout of participants during the experiment.
• Interaction Effects – Interaction of selection with history/maturation.

II) External Validity

• Definition: The extent to which the findings of the study can be generalized to other populations, settings, and times.
• Representativeness or Generalizability of research (Campbell & Stanley).

Threats to External Validity

• Interaction Effects of Selection Bias – Results apply only to the selected group, not to the larger population. e.g., Findings from only progressive farmers may not apply to all farmers.
• Reactive or Testing Effects (Pre-test effect) – Pre-testing makes participants unrepresentative of the population.
• Reactive Arrangements / Hawthorne Effect – Participants change behavior simply because they know they are being studied.
• Artificial Experimental Situation – Lab-like or artificial settings may not represent real life.
• Carry-over Effects – Effect of one treatment carries over to the next.
• Outside Interference – Uncontrolled external influences reduce generalizability.

Comparison Table

Exam Quick Notes

• Internal validity = did X cause Y?
• External validity = will it hold true elsewhere?
• Hawthorne effect → External validity threat
• History, Maturation, Testing → Internal validity threats
• Balance needed: too much control ↑ internal validity, but ↓ external validity.