- A study has internal validity when observed differences on the dependent variable are directly related to the independent variable, and not due to some other unintended variable.
| Threat | description | how to fix | | —- | —- | —- | | Selection Bias | selection of people, subject characteristics | make selections random | | Mortality | leaving the study | make the case that those who | | Location | different locations for different subjects | Keep every subject’s location similar | | data collector threat | person collecting data could be intimidating/change outcome | make the person nice :) | | testing | pretest posttest | different tests at the end | | History threat | something unexpected happening | make it consistent | | maturation | people change and grow over time | pre-post data and comparison group | | Hawthorne Effect/Attitude of Subjects | act of being studied makes people act differently | | | Regression | Outliers come back to a normal result, take the test again, they’ll move back to the center | | | Implementation | subject group are treated differently | double blind | | | | | | | | |
how to counteract threats to internal validity
- Standardize conditions
- Obtain more information on the subjects
- more information on the details of the study
- choose and appropriate design