(most said that they were HIV positive when they weren’t) October 18, 2017 Experiments Is AA effective? o Design 1 Interview 1000 people. Measure alcoholism and ask if they ever attended AA. Finding: AA increases alcoholism Problem: spurious variable= past alcoholism Cross sectional study o Design 2 Interview 100 alcoholics who never attended AA. Measure their alcoholism. Ask a year later if they attended AA in mean time. Measure alcoholism again. Finding: Alcoholism decreased among AA attendants and did not decrease among non-attendants
Panel study; made the survey more valid but there are still spurious variables like maybe family and friends are the reason why you quit drinking and decided to go to AA The classic experiment has… o Experimental subjects o Experiment leader o Stimulus/treatment/experimental manipulation (thing that people are exposed to, react to; takes on the form of the independent variable) o Experimental (gets the stimulus)/control group (does not get the stimulus) o Random assignment-process to which you decide who goes into the experimental group and who goes into the control group (not same as random selection! -how you choose subjects for your research in general) Randomization- everyone within the experiment has an equal chance of being in the experimental or control group Matching-like stratified sampling; if there is a characteristic that you want to ensure that is evenly distributed throughout Ex: you take people who have been an alcoholic for a long time and random put them in experimental and control, then you do the same with those who have not been alcoholics for a long time) o Pretest/posttest Measure dependent variable before and after the experiment allows us to assess whether any change has occurred o Double blind (preferred) Neither the experimental leader nor the subjects know which group (experimental/control) the subjects will be in Random assignment vs. random selection o What’s the difference? Random selection describes the process by which you collect your subjects to begin with; important for survey research, not the most important thing for experimental research Random assignment is how you decide who is in the experimental and control group o Why random assignment more important for experimental research designs? But experiments still face the problem of external validity (results of experiments can’t be used to make a generalized assumption about a group of people because its specific to that group of people at that point in time) o When do you use randomization vs. matching? Randomization might not appropriately represent the population when the Neighborhood effects o Large body of sociological research suggests neighborhoods affect health Is the relationship spurious?
You've reached the end of your free preview.
Want to read all 27 pages?
- Fall '10
- Sociology, ex, Simple random sample