CSAD 285 Threats to Internal Validity

for midterm

regression

use of extreme groups

subject-selection assignment

bias in assignment to groups; existing differences in group characteristics affect study results

attrition

(mortality) participants drop out before completing the study for reasons that affect outcomes

expectancy bias

beliefs, opinions, or expectations (researchers or participants) skew the results

subject-selection regression

use of extreme groups/subjects that are not representative of the population for generalization

maturation effects

natural growth or changes in participant development over time that affects the results

history effects

external events occur between measurements that influence outcomes (ex. tutoring sessions between study sessions)

practice/testing effects

performing experimental tasks or taking tests more than once alter the subject’s performance

infidelity

study treatment or procedures were not implemented as planned

reactivity-hawthorne effect

subjects behave differently because they know they are in a study

pygmalion effect

higher expectations lead to an increase in performance

john henry effect

(compensatory rivalry) individuals in the control group work harder to compete with those in the experimental group

sequencing effects

order of experimental tasks influences the outcome (learning effects or fatigue influences the next assessment)

diffusion effects

components of an experimental treatment unintentionally ‘diffuse’ or ‘spillover’ to the control group

social desirability bias

participants tend to answer questions in a manner that will be viewed favorably by others