PSYCH 100A Lecture Notes - Lecture 6: Sampling Distribution, Null Hypothesis, Effect Size
Document Summary
Participating in one condition affects participation in another condition. Don"t have the same number of people in both groups. Sometimes not relevant to study before & after. Eg. want to study something that affects men & women, cannot switch people from one condition to another. Eg. schizophrenics vs non-clinical sample pre-post conditions do not work, cannot switch them between conditions. Do not know what to match them on. Standard error of difference between 2 means sx1-x2. Typically larger than standard error of means of each sample. Because the potential population of the differences between two means encompass the entire sampling distribution of means of both populations. Calculate standard error of difference between 2 means. For standard error calculation, denominator is sample size not degrees of freedom. Effect size sign (-) (+) does not matter. Effect size always the same sample statistic divided by standard deviation appropriate for the test.