CMN 102 Lecture Notes - Lecture 7: Random Assignment, Internal Validity, Demand Characteristics
Document Summary
Hypothesis testing where one variable affects another. Steps: assign ppl randomly to diff groups. All subjects have equal chance of being in either group (treatment/control: manipulation of the independent variable. In control variable, no treatment: measuring dependent variable. True experiments: non-spuriousness, correlation, time order, association, direction of influence, treat all the same, except treatment group (aka independent variable) Ability to generalize from the sample to the entire population. Subjects feel need to act in a way to form a positive eval. The way the researcher acts contribute to how the results play out. Unwanted change in the instrument used to measure or measuring procedure. Symbolism: x= treatment, o= observation, r= random assignment. One-shot case study design: treatment done to a group. Group observed to see treatment effects: cons: One-group pretest-posttest design: observe/measure subject group. There is some basis but not enough/sufficient. Static control group comparison design: there are 2 groups.