All things study designs
Posttest only
differences in groups only become threats to internal validity if they are what?
systematic
what are some perks to within group designs?
- need fewer participants
- don't have to worry about differences between groups
what allows quasi-experiments to have high construct validity for the manipulated variable?
they use real-world manipulations or experiences
What type of research design measures the dependent variable before and after the manipulation?
pretest-posttest
Define practice effect
A type of order effect where the participant improves because they are experienced with the dependent measure
keeping as much as possible the same between groups is good for eliminating what?
confounds
weak manipulation refers to what?
when the manipulation of the IV does noes not effectively create distinct groups (what is used to assess for weak manipulations?)
what is needed to make something a TRUE experiment?
random assignment (without its a quasi experiment)
what is it called when the measuring instrument changes over time?
instrumentation, like when a coder gets lenient over time
How might you control for order effects?
counterbalancing
when everyone scores toward the top of the DV range regardless of condition, what should you be concerned about?
Ceiling effect
what type of study is conducted before the main study?
A pilot study
What does attrition refer to?
when participants drop out of the study
what might you do to eliminate testing threats?
use different measures of the same construct at pre and post
what is a perk of a small n design
they take advantage of unique situations that we may not have otherwise been able to study
What type of experimental design exposes participants to all levels of the IV at the same time?
concurrent design (look at the diagram in the book to learn these designs)
what threat occurs when a participant scores at an extreme value at the pretest?
regression to the mean
If you are concerned about a potential confound, what might you do in addition to random assignment to ensure that the variable is distributed evenly among groups?
use a matched groups design
do small N designs normally use statistical analyses?