COMM 1 Lecture Notes - Lecture 10: Internal Validity, Random Assignment, Experiment
Document Summary
Variations: more groups, several different treatment just a manipulation and some observation. Have an observation before any observations timing a questionnaire does not count as pretesting that"s just piloting. Possible problem: interaction of manipulation with pretest. But to check on random assignment to get info on change (between time 1 and time 2) Bad idea if treatment/pretest interaction is likely. Threats to internal validity internal validity means you can be sure that x caused y. (no other explanation for results) if not a true experiment or if do experiment improperly, then. Alternative explanations become possible (i. e. , threaten internal. Pre-experimental designs (manipulation of iv, no ra, thus many threats to validity) internal validity) Showed everyone an ad, after, they were really anti-smoking. Want to be able to say that the ad made them that way, but they could have already been.