Threats to Internal Validity

1
Threats to Internal Validity

• Internal Validity establishment of cause and
  effect relationship

2
Selection

• Groups differ from each other prior to
  administration of treatment. (sources ex)
• Traps
• Non random assignment (self-selection).
• Assigning to groups (without randomization).
• Arbitrary assignment rules.
• Matching (good but does not achieve outcome)

3
Selection by Maturation

• Groups begin at the same level (pretest) but
  change as the result of maturation, rather than
  treatment.
• Boys and girls example (weight lifting)

4
Regression Effect

• Random errors tend to balance out. Extreme scores
  will often be less extreme during a second round
  of testing.
• Penny example

5
Mortality

• Participants drop out of study.
• Reasons for dropping out rather than treatment
  could be the issue.
• Example Treatment of addicts in either a
  discussion group or family therapy group.
  Treatment effect may not be type of therapy but
  likelihood of remaining in treatment.

6
Maturation

• People change! They get older and may change as
  a function of getting older, not the treatment.
• Head Start

7
History

• Changes in the outside world may be the cause of
  the change in a person. Environment changed
  between pretest and posttest.
• World Trade Center

8
Testing

• Pretest influences responses on posttest.
• People may respond differently because they had
  the pretest.

9
Instrumentation

• Different pretest and posttest instruments are
  used therefore it is difficult to ascertain
  actual changes as the result of a treatment.