47-512 Applied Research Methods

Detailed Notes for Assignment II

I expect that there will be significant individual differences in the caliber of your assignments. This is not unusual for this class, for which some of you have had extensive preparation and others have had minimal. For some of you, I will expect that you can push the boundaries of what is already familiar to you. For others, I may ask you to re-do assignments as we go along. Our notebook method will permit me to appreciate progress individually. Here are some points that everyone should take to heart.

• Please double space all papers submitted. It makes it easier to write comments that are legible.

• Plagiarism is not allowed.

• Anytime you use material without providing proper citations, that is plagiarism.
• Anytime you copy word for word without proper quotations and citations, that is plagiarism.
• All students are absolutely responsible for avoiding even the most inadvertent plagiarism.
• Please consult the APA Publication manual for citation guidelines and the UML catalogue for university sanctions.

• Professional journals do not include newspapers or popular magazines. They are generally peer reviewed publications. When in doubt, ask.

• Types of studies can be confusing and Leavitt (p. 81, Table 6.1) draws distinctions between categories that, in fact, have considerable overlap (Lecture 2/14). The major distinction is between experimental (random assignment, manipulation, and control) and non-experimental (studying what exists) which can be correlational (associations or connections between variables) or descriptive (levels of variables in a population).

• Examples
• How many hours per week of violent TV do 7 year old boys watch? Descriptive
• Is there a relation between the amount of violent TV watched at home and the amount of aggressive behavior observed on the playground among 7 year old boys? Correlational
• If 7 year old boys were randomly assigned to watch a video featuring either (a) a child playing with a BOBO doll by repeatedly punching it, or (b) a child playing with a toy drum by repeatedly striking it, would there be a difference between the amount of hitting observed in the boys in (a) compared with the boys in (b) when they were left in a room containing both BOBO and the drum in an array of other toys? Experimental

• Within each of these many types of methods (survey, interview, etc.) can be used.
• Correlational studies need not imply that correlational statistics (r or r-squared) will be used to analyze the data. That is one source of confusion. A correlational study may examine the association between gender and math scores by comparing the average scores of girls to boys, not computing a correlation (r). As we will see, correlations are inherent in these comparisons, even though the statistics more commonly referenced in the analyses will be comparisons of means (t or F).

• Research questions are not absolutes but have some context, including the population in which you are interested.

• Samples must be described well enough to know the population to which the results of the study can be generalized. You should also say how many participants were in the study for issues of power (we will talk more about this).

• Variables are key to understanding what the results mean. How are the psychological constructs (happiness, health, self-esteem, etc.) assessed? Using self-report may provide different data than physiology or observation. Make sure to operationalize your reading of variables.

• Independent and dependent variables are terms that apply to experiments only.

• Independent variables are manipulated independent of all other variables. In correlational studies you cannot be sure of that independence. For example, studying two different types of curriculum in increasing AIDS prevention awareness can be studied two ways: (1) randomly assign students to curricula, or (2) study students in classes that employ different curricula. In the first experimental approach, assignment is independent of any student preferences for one approach over another. In the second, correlational approach, student enrollment may well depend on the type of approach that was most interesting or had a particular teacher or any number of other variables with which it is confounded.
• Dependent variables are those that you expect to vary depending on the level of the independent variable applied. If you are observing things in nature, you cannot assume that causal relation.

• In correlational studies, the terms antecedent and outcome variables are sometimes used. Be careful not to use terms that connote causality. This is a common mistake even in professional journals—be wary.

• Results and conclusions are not the same. “The average performance of the group that received Intervention B exceed the performance of all other groups” is a result. “Intervention B appears to be the treatment of choice for these types of youth” is a conclusion drawn from the result.

• Psychological studies do not prove hypotheses, they support or fail to support them.

• Use these comments to double check all papers before turning them in.