Tuesday, May 14, 2013

Psychology IS a Science

My daughter is working on a science fair experiment that happens to coincide with some psychological principles. She is studying the effects of video game play on memory. It is an elementary experiment that was intended to teach the kids the fundamentals of scientific research. Unfortunately, it appears that they have not been given the truth about the basics of the scientific method.

Before you turn your "ears" off because "psychology is not a science" or some other related prejudicial thought enters your mind, hear me out first. Psychology is, indeed, a science built upon empiricist ideals. We use the full scientific method to gather the data we collect and analyze to better our understanding of the world. Yes, early psychological "facts" did arise from somewhat questionable techniques--insight, introspection, philosophical debate. However, today's psychologist is trained in the full scientific method and no psychological research paper, journal article, book is professionally published without a peer review to ensure that the science is sound. I will acknowledge that there are a lot of "pop psychology" books filled with nonsense (too many self-help books fall into this category) and advice, none of them based on solid scientific research. This is why it is so difficult for professionals to educate the public. It is an uphill battle because we use data to back up our writing, not empty platitudes that do not stand up to logical analysis.

Butt getting back to my point, the scientific method involves 5 basic steps. 1) Ask a question. 2) Propose a hypothesis that provides an answer to that question. 3) Test your hypothesis with a valid research method. 4) Analyze your results. 5) Report your results (and your methods) so that others may be able to corroborate your findings and grow the body of knowledge in this area.

1) Asking the question is easy. We have all asked these questions before. Why are girls more sociable than boys? When do babies recognize their names? At what age is a child more influenced by what they see on television? How can I improve my memory for names and faces? I'm sure the average person wonders about something in the realm of human behavior at least once a week.

2) Hypothesis: an "educated guess" that is a) directional, b) testable, and c) measurable. A good hypothesis will identify the independent and dependent variables, as well as the different groups being studied. For example: Children between the ages of 2 years old and 4 years old are more likely to imitate the actions of similar-aged characters they see on television than children older than 5 years of age. Let me define some more terms before I explain.

Variable: anything that can change and/or has different levels or conditions under the same categorical label. Gender, for example, has two different conditions--male and female. While you cannot directly manipulate a person's gender--you cannot expect a subject to simply switch their gender instantaneously--it is sometimes considered an independent variable if no other manipulations are made. Gender, age, ethnicity, and other demographic variables are subject variables, sometimes called quasi-variables, because they cannot be manipulated, though they may be used to separate subjects and categorize data.

Independent variable: the condition (variable) that is manipulated or changed at the beginning of a study. Usually in an experiment there are at least 2 different levels of an independent variable, it is there or it is absent, such as in the case of a placebo/medication experiment. The independent variable would be the pill the subjects receive. Some receive the placebo, some receive the medication. In other experiments, the independent variable can have several levels. If you were studying the effects of sleep deprivation, you might have several experimental groups, each with different amounts of deprivation assigned to them, and the control group that receives no deprivation.

Dependent variable: the condition (variable) that is measured after the manipulations have occurred. The dependent variable is what you really care about in a scientific study. With a standard hypothesis you hope to find some kind of difference in this variable between the groups.

Experimental group: This is the group of subjects who are treated to some form of manipulation in regards to the independent variable. If your independent variable has multiple levels, then you will have multiple experimental groups.

Control group: this is the baseline group of subjects. Aside from participating in a research study, no changes are made to this group of subjects so that any differences in the dependent variable found between the experimental and control groups can be attributed with greater certainty to the manipulation made for the experimental group. You will always have only one control group.

Let's get back to my hypothesis example. Children between the ages of 2 years old and 4 years old are more likely to imitate the actions of similar-aged characters they see on television than children older than 5 years of age. This is directional because it includes the phrase "more likely than." Comparison phrases identify the expected differences you hope to find. It is testable. We can show several children of various age groups television shows that include characters of their age own age groups. We can even complicate our experiment by showing children characters of varying age groups, since we did mention "similar-aged characters" in the hypothesis. We could then allow the children to play for a specified amount of time and record the number of behaviors that were similar to the ones seen in the shows. This indicates the measurable nature of the statement, as well.

3) Test your hypothesis with a valid research method. Many scientists, psychologists included, strive for experimental research designs. The controlled experiment is really the best way we have to identify causation, the holy grail of research. Sometimes an experimental manipulation is not practical or ethical, however. Surveys, naturalistic observation, laboratory observation, and case studies all offer valuable information, but it is primarily correlational in nature. This post is getting lengthy even by my standards, so I will have to address the idea of correlation in another entry.

4) Analyze your results. Most psychologists plug their data into statistical software packages to run the numbers for them, allowing them to quantify behaviors in a way that is presentable. In some cases, such as with non-quantifiable information, a "trend analysis" may be used to summarize the general impressions discovered during the research project.

5) Report your results. For the academic and research professional this is accomplished through publishing in various journals or, in the case of a large body of work collected over a career, in books. For others, especially students, smaller papers or articles, or even presentations are the means through which your work is shared with others.

Some words of advice: Keep your individual studies simple! The more independent, dependent, and subject variables you try to measure and manipulate in a single study, the more complicated your statistics are going to have to be. It will also make it more difficult to sift through the meaning of your data. Manipulating only one variable at a time will help you better control for possible confounds (external factors or unaccounted variables that may have an impact on your dependent variables, even though you may not have controlled for it). If you feel that other factors may affect your dependent variable, then you run additional studies with these factors as the new independent variables. This is how the science is grown without muddying up the waters. This is also how others can study the same topics but still provide valuable information that allows us to paint a much more complete picture.