# commres.net

### Site Tools

types_of_variables

# Identifying variables

Please focus on dependent and independent variable first. You should be able to distinguish them. This is a good material to see if you understand the textbook alright.

# Dependent

A variable assumed to be dependent or be affected or caused by another (called the independent variable). If A is the result of the function of B, A is dependent variable. We can think A is a dependent variable because A is caused by the B('s function); A depends upon B('s function); A is affected by B('s function); A is effected by B('s function).

Example:
A research story: A researcher wanted to know what happens when a car blocks another car on green light. He had a hunch – frustration causes aggressive behavior. He conducted a field research. And here is what he did basically.

He took a spot (a light signal) where traffic is not so crowd in LA local area; He asked the driver (an experimenter) to stop on the light signal, to wait until another car approached behind his car, and not to move even when the light changed (to green). To make the research interesting, he used two cars instead of just one (a very big and expensive car – like '01 Lincoln Town Car and a small, cheap, and old car – '76 Chevy Nova) Each blocking was done at a time. He wanted to see whether blocking a car will cause the driver (behind the researcher's car) to use a horn. In short, the researcher thought: (1) that blocking cars will give the drivers “frustration” (2) that frustration will cause the drivers' aggressive behavior (honking). So, in shorter, aggressive response is caused by frustration.

The aggressive response (hone honking response) is the dependent variable, here. But, by using two types of cars, the researcher went further. He assumed that the degree of aggressive responses are affected by the status of the source of the frustration . That is, the researcher assumes that more aggressive responses will be observed when the frustrator's status is low (the old, cheap Nova?) and less aggressive responses will be observed when the frustrator's status is high (the big, expensive Town Car?). This notion is related to the concepts of moderator or intervening variables. I will talk about them later.

# Independent

– An independent variable is presumed to cause or determine a dependent variable. B is an independent variable, if B causes A, B affects A, B has an impact on A, B has effects on A. In the above case, the frustration causes aggressive responses. Therefore, frustration is the independent variable.

There is another opinion though! As mentioned, the researcher tweaked the situation by using two cars. He assumed that the status of the frustration sources will affect the aggressive responses – if the status of the frustration source is high, less aggressive responses will occur, and vice versa. So, in this case, the independent variable is the status of frustrator. Why is that so (having two kinds of independent variables)? It has something to do with the control variable, which is explained below.

# Control

– A variable that is held constant in an attempt to clarify further the relationship between two other (dependent and independent) variables. Let's think of the above example. The researcher might have thought that there are differences in aggressive responses among the local drivers and out-of-state drivers. The latter might come from Logan, Utah, where people hardly use the horn in traffic. Or he might be a cab driver from New York enjoying a vacation in LA. Or she might come from London in which … we don't know how they usually behave in such a situation. The researchers might think the responses of people outside of LA area are not valid since LA is not their everyday life environment and decide to exclude them. Now, we see the variable is controlled to see a specific relationship.

Another example. What if the researcher wants to choose only male subjects to see how they responded to the independent variable (Note that the researcher wanted to see only male)? In this case, the variable gender is a controlled (being held constant) variable. But, what if the researcher wanted to see the difference between male and female regardless the (high and low) status? In this case, the variable gender becomes a moderator (intervening) variable which will be explained the below.

# Moderator (intervening)

– A moderator variable is a variable that modifies the relationship between the independent and dependent variable. Let's think of the above example. We can draw a picture of how the variables work as below.

Figure 1.
_____________                            _____________
|             |                          |             |
| The status  |                          | Aggressive  |
| of the      |    ------------------>   | Response    |
| frustrator  |                          |             |
|_____________|                          |_____________|
(independent)                             (dependent)

Now, the researcher wanted to tweak the situation and see how male and female responded differently. The variable, gender, changes (moderates) the major relationship between the independent and dependent variable. So the picture is as follows.

Figure 2.
_____________                            _____________
|             |                          |             |
| The status  |                          | Aggressive  |
| of the      |    ------------------>   | Response    |
| frustrator  |            |             |             |
|_____________|           /              |_____________|
(independent)           /                 (dependent)
______|______
|             |
|   Gender    |
|_____________|
(moderator)

In fact, this is how the researcher's study was done. He said, “In both conditions, men tended to honk faster than women.” The picture identifies each variable and the relationships among them. The independent variable (the status of the frustrator) affects dependent variable (aggressive response) and the effect on dependent variable will be changed according to the attributes (male and female) of the moderator variable (gender).

But, some of you may think that the independent variable is “frustration,” not “the status of frustrator.” It is also a valid (even correct) answer, if we look at the research problem at a different angle. Do you know the phrase, “theory should be parsimonious”? In short, parsimony means that the shorter (or simpler) one is better, if two theories explain the same thing. This principle has an advantage. If the theoretical statement is true and simple, it may be used in a different study as a (axiomatic) basis. So, we may assume (I do not know if the study was designed this way, though because the author did not mention this) that there is a simple axiomatic (law-like) assumption of which diagram looks like the below.

Figure 3.
_____________                            _____________
|             |                          |             |
| Frustration |    ------------------>   | Aggressive  |
|             |                          | response    |
|_____________|                          |_____________|
(independent)                             (dependent)

This is very simple and plausible. Suppose that this is a valid result of a study so that the researcher implicitly used this model in his study. What he did, then, is that he made another variable to see if it influences the dependent variable (Aggressive response). The variable is “status.” The variable changes (moderates) the relationship between the independent and dependent variable. Guess what it is called! – moderator (intervening) variable! In this case, the independent variable (frustration) is held in constant (it is controlled! So, it becomes the control variable, too!) So the picture describing the study should look as follow.

Please note that both control and moderator variables are considered as independent variables. It is up to the researchers to determine which are the independent, moderator, and control variable (Sproull, 1995). In the below picture, we can simply think that the variable, frustration, can be out of the picture because it is held in constant. In other words, there is only one attribute in the frustration variable – blocking a car. This condition is going to be held constant no matter what in this study. As a result, if we omit the frustration variable, the picture will be the same as we saw in the figure 1. And if it is the case, the moderate variable here (status of the frustrator) becomes the independent variable.

Figure 4.
_____________                            _____________
|             |                          |             |
| Frustration |   ------------------>    | Aggressive  |
|             |           |              | response    |
|_____________|           |              |_____________|
(independent)      ______|______          (dependent)
(control)        |             |
| Status of   |
| frustrator  |
|_____________|
(Moderator)

The same things can happen with gender variable. I told that gender is a moderating (intervening) variable in Figure 2. Then, what if the researcher wants to hold the independent variable constant (no status difference). Then he is controlling the status variable, use the gender variable as an independent variable.

You shouldn't be confused. As a researcher, the dependent and independent variable should be clarified. Then, he or she can add a moderating variable to see how the relation between the original independent and dependent variable change. If the researcher wants to control either independent or moderating variable, the controlled variable becomes an control variable.

For tests… You should be aware of (1) that the independent variable does something to the dependent variable; (2) the independent variable does happens earlier than the dependent variable. So, in the statement (hypothesis) man are stronger than women, you should be able to figure out that the statement implicitly suggests: (1) that being a man may have caused more strength; (2) that gender (being a man) happened earlier than having more strength.

<Reference>

Sproull, N.. (1995). Handbook of research methods : a guide for practitioners and students in the social sciences, (2nd ed). Metuchen, NJ: Scarecrow Press.