Non-Experimental Research Flashcards
what do we do for things that are unethical to produce in lab
observational research
quasi-experimetns
often conducted, analyzed and interpreted just like “true” experiments. The critical difference in experimental design is that we do not manipulate IVs in quasi-experiments
- cannot claim a causal link between IV and DV, only a correlation between the two
quasi-experiments time series
- We do not manipulate the IVs but we take repeated measurements of the DV before and after the treatment.
- This grants us some confidence: we study the behaviour before the
- critical point and after, and detect the changes.
interrupted time series
several measurements before and after the event.
multiple time series
a control group is present, that is, a time series measurements in a group outside of even
non-equivalent group design
- What if we need to compare groups that are (known or suspected to be) not equivalent at the outset? Fast runners and slow runners.
- The difference between groups at the end of experiment is not surprising, so we can’t compare the groups on the post-test. You train both groups of runners, and the fast ones end up running faster than slow ones.
- Yet we can compare the amount of change in their behaviour. How much did fast runners gain? How much did slow runners gain? This allows to control for initial differences.
- Very common in educational and SLP research: comparing efficiency of educational programs for children with different condition
correlational research
- Still, the interpretation is “correlational”, not causational. There is always a chance that the correlation does not have the causal power, but rather is mediated by some other variable.
observation from the outside
naturalistic observation
need to be unobtrusive
goal is to override reactivity effect
observation from the inside
participant observation
observation under control
observation in the lab
