Experiments (& Review of Sampling) Tuesday, October 9, 2018 3:29 PM MIDTERM: Chapters 1-7 Q. Which type of sampling chooses every kth element on the list? A.Systematic sampling Q. You are doing research on hospital personnel, orderlies, nurses, docs. You want to be sure you draw a sample that has cases in each of the personnel categories. You want to use probability sampling. What would be an appropriate strategy for conducting the research? A.Stratified (want to make sure we have people from each category) . Using a non probability technique, the answer would be quota sampling WILL BE ON MIDTERM: Q. In a UBC lecture of 500 students, 140 were from Alberta, 300 are from BC and 60 are from Sask. From a sample of 100 students in the lecture, the researcher drew 15 Sask students. The weight that should be to Sask is .... A.(On the Activity 2 hand out) Experiments: · Any mode of observation that allows us to make causal explanation · We do research to describe, explore, explain We do experiments when we want to explain the relationship between something Unlike surveys- give us lots of info, rich in info · Experiments not best method for exploring or describing because they are so small, but great for explaining because you can isolate variables and look at the relationship . Some not done in lab, some naturally occurring Some topics appropriate for experiments · Limited and well defined . When we know exactly what we want · A hypothesis you want to test The classical experiment · We need ..... · Must be able to identify ID's and DV's · Pretesting and post testing . Experimental group and control group . why do we have control group? · We want to isolate other factors that could be responsible as well · Hypothesis: green tea makes you do well on exam Can give you tea before midterm, but don't know if it's the tea or other factors . Pretest- see how you do without the tea . One group that gets the green tea and one that doesn't (they are the control group)
· Stimulus is the tea · 100% the stimulus is the ID · All the variables in this chapter are dichotomous · Assume stimulus for ID is always dichotomous Measure our DV before stimulus is given We need a bench mark- a pretest Then apply stimulus Then post test to see if there's a difference between pre and post test · Problem: validity · Ex of green tea- obviously you'll do better on exam 2nd time around because you've seen exam before · Validity problem when post testing- people will know what we're looking for, can effect result · Another issue: social desirability · Experimental group: A group that gets the stimulus Control group: Does not get the stimulus MUST KNOW CLASSICAL EXPERIMENT (in midterm, might refer to "group 1", "group 2") · Has 2 groups- experimental and control group ttroup 1- experimental group Should