Variables

Variables are things that vary.  That means that they will appear in different values when measured.  If your study is only on women, then gender is not a variable, right?  It's a constant because it does not vary, meaning it appears in only one value.  If you were to study both males and females, then gender (or actually sex) becomes a variable.

When testing a hypothesis, if you are making a nomothetic causal explanation (one that you hope to generalize to others), you will have an independent (cause) and a dependent (effect) variable.  What you are saying is that this variable, the IV, has created the change we have noted in the DV.  Sounds simple, right?  We measure those two variables and we have our answer!  So why do surveys have more than just two questions on them?

Questions asked on a quantitative survey are typically assessing a variable.  I mentioned gender, but other variables could be income, age, occupation, ethnicity, and so on.  None of those may be your IV or your DV.  So why bother measuring?  The reason that you may be asking about those items is that you want to make sure that those variables, which we call extraneous variables, aren't the ones that are causing the change that you are measuring.  By measuring them, you can hold them constant (control for) when conducting your statistical analysis.

Let's break this down further.  You want to know if adolescent females have more issues with body image than male adolescents.  You measure gender and then you measure attitudes towards his or her own body (you would probably do that using a scale, which is a series of questions that are tallied to provide a score.  I'll save those details for a later blog).  Using an independent t-test, you can tell if the scores on body image for females differ significantly from the males' scores.  However, is there anything else that could possibly cause this difference?  Could weight be a factor?  Age?  How about activities the person is involved with?  If you think that any of those could be causing the difference, then you ask those on your survey as well.  Then, we can hold those variables constant in our analysis to make sure that the difference still exists.

So, when planning a survey, consider not only those variables that comprise the hypothesis statement, but also add any extraneous variables that may be interfering.