Now that you have a research question, you will need to develop a method to test it. One way to develop operational definitions of the variables of interest is to read literature in your area (i.e. the literature review on your topic may be useful here as it may describe how other researcher have measured or manipulated the variables of interest in the past). However, you may also be interested in testing a research idea in different. Additionally, you may be able to modify methods that have been used before. You will soon be required to develop a research design for your research question. In this document I offer additional assistance in operationalizing your variables.
Survey Measures: Defining some variables is very easy. For example, when students have predictive research questions they can usually use survey methods, and for many constructs or variables there exists reliable and validated measurement scales. A group of students in the past were interested in early childhood parental interactions and preference for short-term mates. They discovered that early childhood experiences are usually measured via the adult attachment scale (Collins, 1996). Further, they found a scale that measured preference for short-term mates. There are entire websites devoted to scales, and some researchers have developed and published so many measures that they have links to their scales from their personal websites. The following resources may help you. There are a large number of scales listed on these sites.
http://www.muhlenberg.edu/depts/psychology/measures.html
http://www.webpages.ttu.edu/areifman/qic.htm
http://ipip.ori.org/ipip/
If you can’t find what you need from above, try putting the construct of interest in a search engine, in a library database such as PsycInfo, or use Google scholar. In some cases, students work really hard to find a scale to measure something really easy. For example, students have looked and could not find scales that measured 1) how many sexual partners a person has had, 2) how much net income a person makes annually, 3) how much time a person spends getting ready (hygiene and grooming behaviors), or 4) how bad a person feels at the moment. In each case, I suggested that the student just ask this question in a survey. If you are really only interested in the answer to a single question, just ask the question. For most other situations of course, it is best to find a validated and reliable measure. Often you can get ideas about what scales are used in your area of interest by reading reports in your area of interest.
Observational Measures: Sometimes defining your variables is trickier. For observational measures, you should think about being in that situation. What kind of behaviors are likely to be elicited? Which behaviors are most clearly associated with the construct I am trying to measure? What is the best way to code the behaviors of interest? Will I be observing one individual or several individuals? How many raters will be used to code each sequence of events? Oftentimes, it is easiest to read papers and see how other researchers have handled the issue. Sometimes, it helps to go observe the behavior, if possible, with a coding schema in mind. This can serve as a practice run. These strategies will help you define your construct and develop a coding schema that makes sense.
Experimental Measures: Manipulated constructs may be the easiest or the hardest variables to operationalize. When you manipulate a construct there are two things to keep in mind. One is whether the manipulation is ethical. Is there a way to manipulate the construct of interest with minimal social, psychological, or other risk? Second, when considering control conditions, you should make sure that the conditions you create are exactly the same except for the one thing you are manipulating. Again, it is helpful here to go to the literature and see how other researchers have manipulated the variable of interest. It is also worthwhile to ask your instructor or advisor for help if you get stuck but would like to manipulate a variable both cleanly (no confounds) and ethically (minimal risk).
Quasi-Experimental Measures: Keep in mind that some independent variables are quasi-experimental. That is, you might be interested in a variable that cannot be manipulated (sometimes called an individual difference variable). These variables include drug use, educational attainment, age, gender, race/ethnicity, health status, attachment style, personality, etc.
Ethical and Practical Consideration: As you design your study keep ethical and practical issues in the back of your mind. This will help you develop a realistic study. Some things to consider:
? What have I done to minimize risks to participants?
? Are there benefits to participation?
? Will I use deception? Is it justified? Why?
? Will participation be anonymous (no names will be attached to data) or private (no one except the researcher will have access to the data)?
? Will I be able to ensure that consent is voluntary and informed? If not, is this justified? Why?
? Will I be able to fully debrief participants? If not, why? Is this justified?
? Where and how will I recruit participants? Will I have access to the sample that I need?
? Do I need any special equipment (e.g., galvanized skin conductor? fMRI? Reaction time software? Video camera?)?
Scales and Measures: Many constructs have correspondent scales or measures that have already been developed and tested for reliability and validity. This makes it much easier for researchers and students to measure the variables they are interested in without developing their own surveys. For example, imagine that you were interested in the effect of rejection on self esteem. One operational definition of self-esteem is ratings on a validated scale that measures self-esteem. In this case, finding a scale to measures a person’s self-esteem is fairly easy. In fact, there are a number of websites devoted to measurement scales in psychology. I will list a few of these below. You may also find scales by simply typing the name in a search engine such as Google or Google Scholar.
? Finding Measures and Manipulations: http://www.muhlenberg.edu/depts/psychology/Measures.html
? Questionnaire Instrument Compendium (QIC): http://www.webpages.ttu.edu/areifman/qic.htm
? International Personality Item Pool: http://ipip.ori.org/ipip/
o Note: To cite this page you should use the navigation buttons to the left side of the page. There is a section that describes the endpoints of the scales (e.g., very accurate). Use these links also to cite these scales. There is a “how to cite these scales” link.
Please note that when describing scales you should describe 1) what the scale measures, 2) provide an example item, 3) describe how many items are in the scale, and 4) describe the endpoints of the response scale. See the example below regarding a scale of shyness.
We also administered the Cheek and Melchior (1985) shyness scale. This scale includes 20 items that are rated on a five-point scale ranging from 1 (very uncharacteristic or untrue, strongly disagree) to 5 (very characteristic or true, strongly agree). Sample items include, “I feel tense when I’m with people I don’t know well” and “I am uncomfortable at parties and other social gatherings”.
Now imagine you wrote your own scale. You should describe this fully as well. Take the example below where I created an attribution scale and described it as follows.
Participants rated the extent to which the actor wanted to perform the action, performed the action intentionally, and how much praise the actor deserved. They also judged the extent to which the actor was moral, rational, trustworthy, and positively motivated. All items were worded positively and to maintain consistency praise rather than blame was measured. Ratings were made on 7-point scales ranging from -3 (not at all) to +3 (extremely).
Please ask your instructor for help if have difficulty locating or creating a measure for your variables.