PSY 3100 Ch. 11: Threats to Internal Validity

one-group pretest-posttest design

measure bhvr, present treatment/internvetion, measure bhvr

maturation

 change in bhvr resulting from passage of time (in participants) 

To identify: consider how much time passed, consider developmental changes

To prevent: minimize time btwn measurements, include comparison

history

change in bhvr cuz of external events

To identify: consider time passage, what happened between measurements

To detect: include comparison thats also affected and look for changes

regression to the mean

 bad things lead to better things “natural balancing act”

To identify: was initial value of DV extreme, anything unusual abt same or conditions?

To detect: include comp group, examine pre-test for equivalence

attrition

participants drop out of study b4 the second measurement.

To identify: is the sample for posttest smaller than pretest?

To correct: remove the dropouts data

testing

taking pretest affects scores on post-test – something learned from pre

To identify: not obvious; be considered w/ pre-post designs

To prevent: use diff version, dont pre-test, include comp group

instrumentation

 change in value of DV on post-test bc of change in measure/thing using

To identify: were diff measures/versions used for pre/post?

To prevent: Use equivalent forms of M

design confounds

age, gender, MUST vary systematically w/ ind variable

demand characteristics

aspects of an experiment that provide cues to lead participants to guess hypothesis

Address: BLIND (singleblind/masked; double blind)

observer bias

researchers who are observing participant behavior may bias coding in favor of hypotheses SOLUTION: don’t tell observers hypotheses or condition

placebo effect

ppl dont receive active treatment but change because they BELIEVE they are receving it (reactivity) SOLUTION: double-blind placebo control study