Exam 3 Tuesday 4 15 in class The exam will consist of 25 multiple choice questions 2 pts each 50 pts 15 points short answer 10 points short essays Total 75 points Pay close attention to recommendations regarding research methodology or the pros and cons of various approaches discussed in class This is what you need to know for the exam 3 25 14 Causation and Experimental Design Know what a causal effect is Change in one variable leads to change in another variable all other thing being equal Know and understand the three key criteria for establishing causality association time order and non spuriousness be able to read a study abstract and describe how well the study meets these criteria Association between the two variables o Most published studies will have association Time order the independent varies must come first and be established o Cross sectional studies are not good at having time order If you only get a snapshot of a family you don t know what happens before or after Non spuriousness a third variable or nuisance variable cannot be explaining the association Understand what makes a study a true experiment random assignment Random assignment to experimental and control groups is key Only thing that differs between the two groups is experimental o Establishes non spuriousness manipulation o Establishes time order Researcher controls how and when the manipulation is administered Know the difference between random assignment and random sampling Random assignment how individuals are assigned to groups Random sampling every aspect is based on random chance Know the problems with experiments Self fulfilling prophecy positive staff expectations create positive outcomes Placebo effect the treatment only working because participants believe the treatment can work Hawthorne effect the experimental group feels special Know what process analysis is and why it is used Periodic measures taken to determine whether a treatment is being delivered as planned Fidelity checks o Usually in a field experiment To prevent the self fulfilling prophecy placebo effect Hawthorne effect Know the characteristics of non equivalent control group and before and after designs quasi experiments Non equivalent control group o No random assignment o Matching can help create a similar comparison group Before and after designs o No comparison group o Multiple assessments before and after the treatment intervention Know the threats to causal validity in experiments and quasi experiments and in which designs these problems are most likely to exist selection bias mortality testing maturation and regression Threats to casual validity o Noncomparable groups Selection bias more likely casual validity Mortality differential attrition more likely casual validity o Endogenous change Testing taking a pretest can help posttest quasi experiment before Maturation developmental change more likely quasi experiment and after before and after Regression extreme scores become less so more likely quasi experiment before and after for all endogenous change not a problem if there is a control group o History Influence of external events Particular concern for before and after designs also for treatments done at multiple sites o Contamination When control and experimental groups affect each other