If the road is not yet clear to you,

we will show you the way how to get through.

Variables are the outcomes of measurements. The way a variable is measured is called the operationalising. Even small changes in measuring may have a large impact.

The way things are measured influences the type of analysis that may be performed. An example can clarify a lot. Suppose you want to know the age of the respondent. That is possible with a simple question is 'What is your age?' The respondent can state his age on a dotted line (pen and paper questionnaire) or in a text field (online questionnaire). An alternative is to present the respondent with a list of age groups of which he can tick one. Although the same question is asked, the different types of answers have an enormous impact. The first – writing a number – provides a variable measured at ratio level. The second – a radio button to tick an answer – provides a variable measured at ordinal level. This has major consequences. Variables measured at a ratio level can be involved in much more types of analysis, while variables measured at the ordinal level are limited in their possibilities. Moreover, data measured at interval or ratio level usually require fewer cases or respondents to find statistical differences. All in all, it is much more lucrative and satisfying to have data measured at the interval or ratio level.

Instead of measuring an aspect with a single question a number of questions can be used. For example, if you want to know if a respondent is satisfied with the house in which he lives, one question can be asked: Give a mark to your home about how satisfied you are with it. This is not a very sophisticated variable. Maybe the person is satisfied with the living room, but not with the bedroom and somewhere halfway with the kitchen. And maybe the garden is important or the state of maintenance. All in all a more sophisticated variable for ‘satisfied with the house you live in’ can be accomplished by asking many more questions. This raises the question of how this variable should be constructed and which content should be involved. You can read more about these topics on our content validity and construct validity pages. Factor analysis and Cronbach’s alfa will help you to create these variables. Every way to operationalize the variables influences what will be known.

Measuring difficult things should be operationalised. In the sixties of the last century there were many discussions about how to measure intelligence. There were too many tests to measure the same thing, but all of these tests did not get the same results. Even nowadays it is hard to tell what intelligence is and how it should be measured. Some tests survived all commotion and it looks like these tests are becoming the standard for measuring intelligence.

To prevent discussion about gathering valid and reliable data, try to make use of standard tests and questionnaires. It will assure you of getting data about which the validity and the reliability needs hardly to be discussed.

Related topics to Operationalization:

• Measuring
• Variable
  • Nominal level
  • Ordinal level
  • Interval level
  • Ratio level
• Validity
  • Construct validity
  • Content validity
• Factor analysis
• Cronbach’s alpha

Having reliable and valid instruments in research is a necessity. Read how to make such instruments:

• Testing instruments with factor analysis and Cronbach's alpha

Return to the index of this dictionary