PSYCO105 Lecture : Jan 16 - Methods Continued.docx
1 May 2012
School
Department
Course
Professor
Document Summary
Inevitable, because we can"t precisely control all variables that can influence behaviour (ex. Individual differences among the research subjects; imperfections in the measure of behaviour) Its consequences tend to disappear when averages are calculated. Non-random (directed) effects caused by some factor or factors extraneous to the research hypothesis. Can"t measure bias and can"t be controlled. Can"t be identified or corrected for statistically. Problems caused for researchers: error: reduces the chance of finding statistically significant results, bias: can lead to the false conclusion that a hypothesis has been supported when, in fact, some irrelevant factor has caused the results. If the members of a particular group are initially different, from: those of another group, Differences between groups: larger population: with random assignment, individual differences are a source of error, without random assignment, differences can also be a source of bias. Literary digest"s poll of u. s. votes in 1936. A good measure is reliable and valid.
