Hypertext Tutorial on Social Science Experimental and Correlational Study Design Issues
Don R. Osborn, Ph.D. Professor of Psychology Bellarmine University
(If you review the Methodology chapter in your textbook before working on this tutorial, you will be able to skip some of the hyperlinks since you will already know some of the concepts. Since any user may skip some pages, repetition is built in for maximum clarity. There is also repetition in some of the explanations and in some of the hyperlinks because users may take different links in developing their understanding. While most of these examples are based on actual studies, the statements about them may be altered for instructional purposes. You should always consult the original research before making conclusions about what it means. A hint --using your Browser's back arrow might be the easiest way to navigate.)
Psychologists use an empirical way of knowing . (If you click the highlighted term you can go to a page that explains more about it.) The research study is the basis for empirical knowledge so this tutorial is designed to teach you how to recognize the basic elements that make up a psychological research study. This tutorial is interactive since you will be reading statements and within those statements identifying the basic elements of a psychological study with the hyperlinks giving you specific feedback. Your first goal from working through this tutorial is to be able to identify the independent and dependent variables within a study. You will also be asked to identify the participants, the setting, and the hypothesis since those are critical elements you need to recognize to correctly pick out the independent and dependent variables and to understand what the study actually shows.
Then you will learn to recognize the difference between a true experiment, where the experimenter caused a difference between participants in the different groups, and a correlational study, where the correlational studier used some preexisting quality of the participants to place them into groups. I am calling the experimenter in a correlational study the studier to emphasize it is NOT an experiment. This is an important distinction since you can only reach unequivocal causal judgments with true experiments.
The true experiment has strong internal validity. This means the true experiment's causal conclusion on what the data shows is accurate and unambiguous. In contrast, correlational study conclusions are uncertain and speculative because of questionable internal validity. We will explore the meaning of these abstract ideas with concrete examples. Finally, if a study does have internal validity, can you generalize from the study's findings to other situations? This is the external validity question. This tutorial is primarily designed to teach you how to accurately think about internal validity. Internal validity is the most important because without internal validity it is impossible to have external validity.
Since the most basic elements to
identify are the independent variable = the cause, and the
dependent variable = the effect, read through the statement A below based on the Garrity & Degelman (1990) tipping study and see if you can click on the effect, also known as the dependent variable. Then find the cause, also known as the independent variable. (If you cannot pick the right answers within the statement you can click on the terms above and the hyperlink will take you to the answer.)
A. Patrons in a buffet restaurant tip more when the server is assigned to introduce herself by name to a table compared to when she is assigned to not introduce herself to a table. The reader says, "This shows server name introductions leads to higher tips."
Now, see if you can pick the participants, the setting, and the hypothesis out of this statement. (Again, if you can't pick them out of statement A you can click on the terms.) Once you understand these concepts you can understand any psychological study and develop your own proposed studies. There are six examples provided to learn from.
Now consider the statement below based on the McCall & Belmont (1996) tipping study. See if you can click on the effect, also known as the Dependent Variable (DV). Then find the cause, the Independent Variable (IV).
B) When patrons were given either the randomly assigned tip tray with a credit card emblem or the tip tray with no insignia the patrons who received the tip tray with the credit card emblem tipped more at two different restaurants. The reader says, "This shows that credit cues cause higher tips."
Now pick the hypothesis prediction, participants, and setting in the McCall & Belmont (1996) study.
For the LaTour, Snipes, & Bliss (1996) study below see if you can pick out the dependent variable:
C. Women who were assigned to watch a strong fear inducing stun gun ad on TV are more likely to say they would buy a stun gun than those who were assigned to watch a low fear inducing stun gun ad on TV. The reader says, " This shows the effectiveness of fear in making effective ads."
Now pick the IVs, participants, hypothesis prediction, and setting in the above statement.
Studies A, B and C are all true experiments since the experimenter can manipulate the level of the IV. In our example A the experimenters can cause any diners to either be exposed to an introducing or non-introducing waitress; in B the experimenters can cause any diner to either be exposed to a tip with a credit card logo or without one; in C the experimenter can arrange for any woman to see a strong fear ad or a weak fear ad. You can use true experiments to make unequivocal causal statements like the "reader says ..." statements in the above examples.
Let's work with a different kind of study, a correlational study, where the experimenter cannot assign the independent variable to the participants. In this imaginary study D identify the dependent variable.
Now identify the independent variable.
Now identify the hypothesis prediction
D. Men will tip more than women will.
Notice the independent variable cannot be manipulated by the researcher in this proposed study. This type of IV is called a subject variable or sometimes a pseudo-independent variable (PIV). The studies that use these PIVs are called correlational studies to distinguish them from true experiments. In contrast to the manipulated variables of true experiments, studies with subject variables are not causally interpretable because they have confounds.
Here is another study (Freedson et al., 1983) to practice your analytical skills on. Read through problem E. and first identify the independent variable, then the dependent variable and finally the hypothesis.
E. Female body builders were statistically significantly less anxious, neurotic, depressed, angry, and confused than the general population. The reader says, "This shows body building leads to better mental health."
What would you say to the reader?
Here are some choices for you to think about. After thinking click the link for feedback
1e) The reader is right. This study shows body building is an effective way to improve mental health.
2e) The reader is wrong. There is so much individual variability in body builders' mental health that this study could not find that the mental health of the body builders was better than the general population. Some people are just mentally healthy and others aren't.
3e) The reader is wrong. It could be that body building leads to better mental health but there is at least one alternative explanation I can think of for these results.
4e) The reader is wrong. It could be that body building leads to better mental health but there are at least two alternative explanations I can think of for these results.
Here is the last study to practice on (from Straus, 1994). Identify the independent variable, the dependent variable, the hypothesis.
F) Older parents spank less frequently than younger parents. The reader says, "This shows younger parents are stricter than older parents."
Here are some evaluations of the reader's conclusion for you to think about. After you have thought through your own ideas click on the hyperlink.
1f) The reader's statement is supported by the data so I think it is a correct inference from the data.
2f) Assuming spanking is a measure of strictness, then I think the reader's statement is supported by the data and it is a correct inference from the data.
3f) I think the reader's statement has a problem with reverse causation since it is based on a correlational study.
4f) Even assuming spanking is a measure of strictness, the reader's statement has a problem because I see a plausible rival alternative hypothesis for this data. That is, I see another reason why older parents might spank less frequently besides the reader's hypothesis that younger parents are more strict.
You have now worked your way through this tutorial and, if you chose the correct matches and/or learned from your incorrect choices, you can be confident you have the ability to identify true experiments versus correlational studies; independent variables, dependent variables, hypothesis predictions, and other components of psychological studies. You see why, compared to a correlational study, a true experiment can lead to unequivocal causal conclusions. You are ready to move ahead and study psychology, sociology, political science, or public policy issues with a deeper understanding of how facts and theories work together to increase our understanding of human nature.