Types of Measurements in the Behavioral Sciences

We'll move on now to talk about the types of measurements that we use in the behavioral sciences. The first major type of measurement is a self-report measure, which means that you ask the participants to report their own assessment of the variable you're attempting to measure. There are two main types here: the fixed-format response and the free-format response. In a free-format response, the participant provides any thoughts or feelings that they have about a given question. These responses give you a great deal of data, with rich amounts of detail and a lot of information. But you can imagine that if you asked a free-format question of 100 different people, and each answered it in a slightly different way, it would be quite difficult to do any kind of numerical analysis of their responses. You would have to find some way to code all of these responses into the same numeric scale. Say that you had a free-format question to ask participants to say what they see when shown an inkblot (much like the Rorschach test). Of course, you could get a great deal of information from the participants, but how would you do any kind of statistical analysis on such responses?

Fixed-format responses, on the other hand, ask participants to choose from a set of alternatives when giving feedback, much like a multiple-choice test. You don't get as much detail from a fixed-format response, but it is much easier to code into numbers for analysis. A common way of asking fixed-format responses is to use what is called a Likert scale. Likert scales ask a participant something similar to, "On a scale from 1 to 7, 1 being strongly disagree and 7 being strongly agree, rate your agreement with each statement." Likert scales may have any number of response choices, but you need to consider a couple of major things when deciding how many to offer to participants. If you have too many choices, participants may become confused. But if there are too few, you may not get as nuanced a view of what participants are thinking. Another consideration is whether you want to have a middle point (such as 4 on a 1-to-7 scale). People like to have the option of a neutral response, but neutral responses don't really tell you much about the participants' thoughts and behaviors. If you give the option of a neutral response, people will choose it a good deal of the time, generally speaking. One more thing to consider about fixed-format response types is the notion of acquiescent responding, or the possible tendency for participants to simply agree with every question. For this reason, it helps to write half of your questions in one direction of agreement, and the other half in the opposite direction. For example, if you were asking about job satisfaction, for half of the questions, a "strongly agree" response would indicate high job satisfaction; while for the other half of the questions, a "strongly agree" response would indicate low job satisfaction. At the very least, this format ensures that your data won't come out completely neutral or skewed because of acquiescent responders.

Another type of fixed-format responses is the semantic differential, which puts two extremes at either end of a line, and participants are asked to put a dot or a marker on the line to indicate what they think. See the following example:

How do you feel about Bill and Ted's journey?

Bogus ------X-------------------------------------------------------------------- Excellent

How do you feel about Bill and Ted's adventure?

Bogus----------------------------------------------------------------------X-------Excellent

When people know they're being assessed and measured, they may behave differently; as such, any time you ask people about their feelings and thoughts in a behavioral science context, you run the risk of getting answers that are not entirely honest or truthful. Participants may want to make themselves seem like better people than they truly are, or they may want to simply mess with your data. These are examples of reactivity, or participants responding differently because they know they are being studied. A great example of reactivity is the Hawthorne effect. Social desirability, on the other hand, is a special case of reactivity in which people want to seem like good people; they will give answers that don't necessarily reflect what they actually believe.

 

Behavioral Measures

One way of avoiding the problems associated with self-report measures is to use what is called a behavioral measure. This type of measure is one in which you actually measure something about the participant's actual behavior in response to a variety of different stimuli. Suppose you have a hypothesis about how much eye contact employees make with their bosses and the subsequent achievement of those employees. You could measure a variety of different things:

 

 

These main categories of behavioral measures can be applied to any number of different behaviors. But again, the difficulty is choosing which is the best measure for your specific research question and hypothesis.

Psychophysiological measures are another type of behavioral measure that allow us to look at the participant's physiology in response to different stimuli. If you want to know some specifics about psychophysiology, take a look at this psychophysiology Wikipedia page. In short, psychophysiological measures use electrical and mechanical devices to transform the participant's heart rate, respiration, and even brain activity into numerical signals that can then be analyzed. These techniques are challenging, complicated, and sometimes rather ambiguous in terms of what they tell us; a number of researchers have devoted their entire careers to developing these methods. While we can learn interesting things about behavior from psychophysiological measures, you must be cautious when jumping into this realm of study, as it can be quite complicated.

Choosing the right measure for your research question can be quite tricky, but the best guide for you can often be prior research. If you're investigating job satisfaction, and you want to know what's the best way to measure it, look into how other people have investigated job satisfaction. Often, there are a lot of very clever and educated people studying the same thing you want to study, and they have found the best way to measure that thing through extensive investigation. This is an area where you can truly stand on the shoulders of giants, and perhaps you may see farther than others because of it.