Cause and effect in social science research becomes far more complex than an experiment in the physical sciences. In the physical sciences it is easier to directly correlate a result to the cause and to decrease other variables. However, when working with human subjects many variables and extraneous situations come into play.
I do not believe that developing causal relationships should be the primary goal of research. Although it may be appropriate in certain situations, I think there are many other research designs that are just as, if not more, valuable. Causal relationships are often difficult to prove when dealing with social sciences because of the nature of human subjects and the environments provided.
One of the main problems that arise when trying to develop cause is the idea that just because there is a relationship does not mean it is a causal one. The relationship could be caused by many other factors. The researcher tries to eliminate all factors except for the independent variable, such as the conditions and procedures. However, in order to truly conclude that an intervention was causally related to the outcome, all extraneous variables must be controlled and this is often very difficult to do. Some examples of factors that may influence a study include instrumentation, pretesting, experimenter effects and subject attrition. For example, a longitudinal study concerning motivation may be swayed by students who moved or dropped out of school. In another case, a researcher may assist students in receiving higher test scores or sway their answers on an interview. A pretest may affect results if it changes attitudes or perceptions towards a topic. Finally, a difference in results could be related to changes or unreliability in instruments or raters. Cause and effect research can be valid, but direct control of the intervention and maximum control of extraneous variables should be closely monitored. This is often difficult to regulate in an educational setting.
I definitely believe there is value to research that does not lead to causal inferences. For example, even if a specific intervention was not conducted, simply observing naturally occurring phenomena can help build the knowledge base on teaching methods, student behavior or achievement. This information could also be extremely valuable to pre-service teachers in learning classroom management techniques, strategies and teacher-student relationships. Other types of research do not require such control and therefore may show more realistic and reliable findings. You do not necessarily need to be looking for a desired outcome based on a cause to find valuable information and data. For example, observing the book selection of elementary school boys in a library could help a teacher determine common interests which could lead to stronger reading achievement and motivation.
Great. You make a strong case for non-causal studies. However, think about where this line of reasoning is leading you. This research is still leading us to make decisions that we hope will have effects. We hope the research will affect better classroom management skills in pre-service teachers. We hope it will affect more reading among elementary school boys. We are still looking for causal understanding at a certain level.
ReplyDeleteYou also make a solid critique of studies that claim causation. There are always the extraneous variables. And not only is it very difficult to account for them all, but I would say it is fair to say that is impossible to account for them all.