SOCECOL 10 Lecture Notes - Lecture 10: Standard Deviation, Internal Validity, Pocket Cube
Document Summary
Random sampling: the selection of participants from the large population: increases our external validity. Random assignment: assign to groups: best way to distribute potential confounds, controls for known and unknown confounds, differences are evenly spread out. Individual differences: most powerful tool, when you assign people who are already chosen to participate into random group, important for internal validity. Random assignment is the best method: between groups variance. Mean of group 1 mean of group 2. How different are the two means: within group variance. A general class of designs in which different research participants are used in each group. Topics to be covered: post-test only control group design. Simple randomized design/simple random sample design/completely randomized design. Randomly assign people to group a or group b, do something to one group and not the other, measure results: multilevel completely randomized design. Can have more than 2 levels of the independent variable.