UIC logo

Psychology 242
Research in Psychology
Dr. David J. McKirnan


Week 6.
Sampling

Lecture notes

This week we will finish with basic experimental design, and discuss sampling.

RDS image

Key issues in gathering a research sample:

Who do you want to generalize to? Is a Probability or random sample possible? Types of Non-probability or “convenience” sampling

Depression ScreenLecture notes are here.

Readings

Text:
Control and the Experimental Method and The Ecology of the Experiment.

Readings:
Two short articles from the New York Times on whether children of working mothers have lower test scores. Do simplistic analyses really address this question? There is so much else going on in the lives of children and their parents.

The image to the right shows the extensive screening that takes place for a drug trial of anti-depressants. Click to ask whether participants enrolled in such studies actually represent depressed patients generally. I will be using this in lecture this week.

Discussion group Assignment

Analyze a research design

You will analyze a research stud this week.  Last week you found journal articles for your paper. Take one of the empirical papers you will use for your paper - i.e., a paper that ran an experiment - and analyze it.

You will address each of these areas for your research paper, so get familiar with them!

As a first step go to the Guide to Reading Journal Articles. Then, answer the questions in the box in one typed page or so.

If you cannot find an appropriate article for your paper, you can use: sunscreen use at the beach.

Participant Selection

Participant Assignment

Experimental Procedures

Experimental Treatment or Manipulation

Interpretation of Results

Sample arrow

Group A arrow

Procedure A arrow

Treatment arrow
(experimental group)

Outcome

Group B arrow

Procedure A arrow

No treatment arrow
(control group)

Outcome

How did the researchers ensure that the sample was representative of the larger population?
If they did not, how should they have?

How were participants assigned to groups?...random assignment? matching?

Did they ensure the groups started out exactly the same? Was there any self-selection or other bias?  

If there was bias, how should the groups have been assigned?

How did they ensure equality of procedures across groups -- what experimental controls were present?
What should have been present?

What was the independent variable?  Dependent variable?

Were there any confounds of the independent variable?
How did they (or should they have...) avoided confounds, so that the treatment was the only variable that differed between groups?

Did the groups "really" differ - was the difference statistically significant?
This is beyond what we have covered so far, but see if you can make some sense of the statistics.

What would it have meant if the groups were different, but: a) only by a very small amount, or; b) there was a confound?