Concepts and Constructs
The first step in the measurement process is to define the concepts we are studying. Researchers generate concepts by generalizing from particular facts. Concepts are based on our experiences. Concepts can be based on real phenomena and are a generalized idea of something of meaning. Examples of concepts include common demographic measures: Income, Age, Eduction Level, Number of SIblings.
We can measure concepts through direct and indirect observations:
- Direct Observation: We can measure someone's weight or height. And, we can record the color of their hair or eyes.
- Indirect Observation: We can use a questionnaire in which respondents provide answers to our questions about gender, income, age, attitudes, and behaviors.
Constructs: Constructs are measured with multiple variables. Constructs exist at a higher level of abstraction than concepts. Justice, Beauty, Happiness, and Health are all constructs. Constructs are considered latent variable because they cannot be directly observable or measured. Typical constructs in marketing research include Brand Loyalty, Purchase Intent, and Customer Satisfaction. Constructs are the basis of working hypotheses.
Brand loyalty is a construct that marketing researchers study often. Brand loyalty can be measured using a variety of measures:
- Number of items purchased in the past
- Monetary value of past purchases
- Frequency of past purchase occasions
- The likelihood of future purchases
- The likelihood of recommending the brand to a friend or family member
- The likelihood of switching to a competitive brand
An attribute is a single feature or dimension of a construct.
Measurement: Measurement is the assignment of numbers or symbols to phenomena. Measurement requires a scale. A scale provides a range of values—a yardstick—that corresponds to the presence of the properties of the concept under investigation. A scale provides the rules that associate values on the scale to the concept we are studying.
We can classify the values derived from measurement into two broad categories: 1) Variables and 2) Constants.
Variables: Variables are measurements that are free to vary. Variable can be divided into Independent Variables or Dependent Variables. A dependent variable changes in response to changes in the independent variable or variables.
Constants: Constants, on the other hand, do not vary.
In statistics and survey research, responses are typically described as random variables. The value of a random variable varies by chance or in a hit-or-miss or haphazard manner. This means the respondents' responses to a survey cannot be predicted with absolute certainty. For example, when people are asked whether they intend to purchase a new product, or whether they approve or disapprove of a particular public policy there is uncertainty about what the responses will be.
In physics, on the other hand, the speed of light in a vacuum—186,000 miles per second—is a constant. It does not vary.
A variable can be transformed into a constant when the researcher decides to control the variable by reducing its expression to a single value. Suppose a researcher is conducting a test of consumers' taste preference for three brands of frozen pizza. There are a number of variables in this test: 1) Respondents' ratings of the taste of each brand of pizza, 2) The manner in which is each pizza is presented, the type of the plates and table cloths used, and 3) The manner in which each brand is prepared. To get an accurate measure of the first variable—respondents' ratings of the taste of the three pizza brands—the researcher will hold the second and third variables constant. By serving all three pizzas on the same kind of plates with the table dressed in the same manner, preparing the pizzas in identical ways, and serving them at identical temperatures, the research controls for these variables. In doing so, the researcher has removed, or controlled for the affect of the second and third variables on respondents' taste preferences. Researchers call variable 2 and 3 control variables.