III. Activities Which Solve Problems and Produce Better Knowledge
Once you have a general understanding of why to perform experiments and a topic you want to understand you can usually go forward. But sometimes it's hard to decide exactly what activities to engage in, or how to conceptualize the kinds of things you might do in that research area. Often we say we do research in order to discover new information or discover the cause of something or even simply to answer why. Unfortunately these goals don't give beginners much of a handle on what to actually do. If you have difficulty knowing where to start, your research question may be too global for a beginning. Start with a very specific subset, one which will help you understand the big picture but may not solve it all at once. At its most specific a research question is simply a specific research activity. But generally a research problem is more general and could be solved by any of a number of activities. A general approach to help you get started, therefore, may be to not start with “how can I solve the problems of the world?” Rather, start with how can I help my roommate study more? Then think of the more general issue of what factors influence studying? Then, why does behavior work the way it does? The answer to this final, most general statement of the specific question is pretty close to “how can I solve the problems of the world?”
This activity is perfectly acceptable research. Simply do what seems like an interesting manipulation. Your background lead you to wonder what would happen if you did it, other people may have the same question. In fact many important scientific discoveries were made because the researcher did exactly that, "I wonder what would happen if I do ......" The risk in this type of activity is that your manipulation may have already been studied or no one else may be interested in your question.
Example: Building on the social attention example, you might want to know what would happen if subjects were exposed to several people interested in their course material.
Some findings have only been demonstrated once. The original study could have been flawed. It is important that our knowledge base be on better grounds than a single demonstration.
Example: Do an experiment over. Do everything exactly the same as the original.
Change the subject, apparatus, or procedure in order to obtain a slightly different perspective. In this case, you maintain the same process thought to be the causal factor, but you freely change whatever is thought to be “irrelevant” factors (systematic replication).
Many pioneering research studies did not use what have become the best measures for the independent and dependent variable. Measure the behavior directly rather than indirectly. Measure it more accurately, or move up from a nominal, ordinal, or interval, to a ratio scale.
Example: You refine the "interest" measure into those which simply show interest, those which ask for a fact, and those which ask for some integration of information. You then compare the effectiveness of each.
Occasionally research is published with an unobserved confound which you may notice. The study can be redone without the confound.
Example: You read an article which manipulated a variable intended to increase exploration in rats (it failed). You realize that it was conducted during the day when rats normally sleep. You redo the experiment and run it during the dark part of the rat's light cycle.
General functional relationships may or may not apply to specific situations. Measure what is proposed to be a general effect in a particular group.
Example: The effectiveness of social attention on studying behavior was examined in students from lower socioeconomic status families.
Functional relationships obtained in a specific situation may or may not generalize to other situations or populations. Measure behavior in other groups or other situations.
Example: The effectiveness of social attention on reading the instruction manuals which accompany VCRs was tested with a variety of people to see if social attention affects behaviors of noncollege students.
A technological advance in one discipline may have an enormous impact on the production of knowledge in another discipline. New tools may reveal the answer to long standing questions. If you have a particular research problem already chosen and you can bring new technology to bear on it - do it.
Example: Using a TV camera with a telescopic lens visitors at a zoo were monitored while near instruction signs. The effectiveness of other people asking questions on increasing reading was assessed.
Sometimes the information we need is the synthesis of two or more previous functional relationships. The combination of two different research designs can provide a simple and elegant demonstration of a complex functional relationship.
Example: A 3x3 factorial design was used to manipulate educational level and previous experience with social attention.
You may be interested in demonstrating some new functional relationship. In this case you would change your procedures around until you have a clear demonstration.
Example: You believe that the effectiveness of social attention is determined by specific experiences in life before the age of two. You find groups which react differently to social attention and examine their infancy. You also sort infants into two groups according to your criterion experiences and follow up their susceptibility to social attention over the next thirty years.
The validity of a theory is assessed by generating a prediction based on the theory and then testing to determine if the theory was correct. This is classical deductive research.
Example: Your theory predicts that this early determinant of susceptibility to social attention also affects the person's verbal skills. You compare GRE verbal scores of the two groups known to differ with respect to early experiences.
This type of research is forced when an unexpected behavioral relationship is uncovered. The existing theoretical formulations do not handle the result and, therefore, a functional context for it must be established.
Example: You find that once someone begins to obtain reinforcers from one person regularly, other people do not function as reinforcers as well.
The identification of the “underlying cause” of a psychological phenomenon can be the goal of research. You may know of a particular functional relationship and be interested in discovering exactly why it works the way it does. This is classical inductive research.
Example: You find five students who do not respond to social attention at all. Your initial test is to determine if money will function as a reinforcer.
Send comments/criticisms/speculations to
Date Last Reviewed: November 17, 2002