OBSERVATION METHODS

Observation involves the process of physically or mechanically record- ing some specific aspect of a consumer’s activity or behavior. Some re-

search designs call for this type of data. Some contend that observation is more objective than communication. However, observation, whether in a field setting or a laboratory, is not very versatile. Observation cannot an- swer questions concerning attitudes, opinions, motivations, and intentions of consumers. Observation works well when measuring specific behavioral variables but is less effective measuring attitudinal variables. Often a design that starts with unobtrusive observations followed by communication to ob- tain information about motives, attitudes, etc., behind the observed behav- ior, is an effective approach. Direct observation involves actually watching an individual’s behavior. Their purchase behavior is watched or they are viewed while using a specific product. Sometimes this behavior is moni- tored in a natural setting or real-life situation. Observations can also be made in an artificial setting that is developed by the researcher to simulate a real-life situation. Quite often a contrived store setting is established to sim- ulate consumers’ responses to products in an “actual” purchase situation. A major advantage of artificial or simulated observation is the greater degree of control offered researchers to alter specific variables.

An observation may be accomplished with or without the knowledge of the subject. Obviously, the advantage of disguised observation is that the subject has no motivation to alter his or her behavior. Mechanical observa-

tion is sometimes appropriate to meet the study objectives of the research

design. When this is the case various mechanical devices such as cameras and counting instruments are used to make observation. Eye motion, gal- vanic skin response, perspiration, pupil size, and various counting devices are examples of mechanical methods of measuring and recording activity. Eye tracking equipment can be used to measure where the eye goes in view- ing an advertisement, a product package, or promotional display. A specific application of this technology would be to determine the impact of color in advertising and copy on a newspaper page. The tracking equipment records not only where the eye goes on the page, but also how long it focuses on a specific area. People meters are designed to measure both the channel that is being watched as well as which person in the household is doing the watching.

SUMMARY

So far, we have discovered that the pursuit of answers to research objec- tives necessitates both a research design and data-collection methods that permit the execution of that design. “Good research” in any particular case could involve collection of information using one or more of the combina- tions of research design type and data-collection method. Table 4.8 illus- trates that all combinations are possible. Which cell(s) a researcher should be in is a function of the research objectives, research budget, time avail- able, and the decision maker’s preferred data form for making decisions.

TABLE 4.8. Examples of Data-Collection Methods Used for Generating Explor- atory, Descriptive, or Causal Research Information

Data-Collection Form Research Design Types

Exploratory Descriptive Causal

Secondary

Communication Previous industry studies help to define what “customer ser- vice” means to con- sumers.

Annual survey re- vealing the number of times during the week people wash dishes by hand.

A journal article that reports on testing the hypothesis that public service advertising changes peoples’ atti- tudes toward donating blood.

Observation AWall Street Journal article reporting that there are a growing number of magazines devoted to Internet usage.

Syndicated data in- dicating the market shares of various brands of coffee.

A report by a market- ing research firm on the use of test market- ing by different indus- tries in the past year.

Primary

Communication A focus group of dis- tributors that dis- cusses trends in package handling technologies.

A survey of sales- people to deter- mine the frequency with which competi- tors have given free goods to dealers.

Running two adver- tisements in matched markets and determin- ing which ad resulted in greater recall and positive attitudes to- ward the company. Observation Watching production-

line workers using a handheld grinder and gaining insights on product design changes.

Recording license plate numbers in store parking lot and getting R. L. Polk to generate a map showing trade area and distribu- tion of customers.

Doing a test market for a new product and de- termining if sales reached objectives.

Chapter 5

Measurement

Measurement

INTRODUCTION

In Chapter 1 we described marketing research as producing the informa- tion managers need to make marketing decisions. In that chapter we also listed numerous examples of how the results of marketing research can in- form marketing decisions such as in concept/product testing, market seg- mentation, competitive analysis, customer satisfaction studies, etc. These, and many other examples of marketing research studies, illustrate the need for measurement—“rules for assigning numbers to objects in such a way as to represent quantities of attributes.”1_{Several things we should note about} this definition:

1. Not all of what researchers do involves measurement. Researchers are interested in generating information, which leads to knowledge, which leads to better decisions. Sometimes that information is in the form of insights from exploratory research studies such as focus groups, in-depth inter- views, projective research, and similar methods. For these techniques we are generating information, but we are not “assigning numbers to objects,” so we are not “measuring.” As we have said before, information does not have to have numbers attached to it to have value, and we are dangerously oversimplifying our analysis when we favor information with numbers over that without, simply because it has the appearance of being “hard evidence.” 2. The “rules for assigning numbers” will be discussed in greater depth in this chapter, but we should note here that those rules exist so that we can be more scientific in our measures, and can place more confidence in the num- bers that those rules help generate. We want to make decisions that are grounded in information that we believe correctly represent reality. This means the assignment of numbers should map the empirical nature isomorphically (i.e., on a “one-to-one” basis).

For example, if we assign the number “5 lbs” to represent the weight of an object, we want to make sure that the weight is 5 lbs and not 8 lbs or 3 lbs. Using a carefully calibrated scale is how we ensure in this example that we

have correctly measured the item’s weight—assigned numbers to the object to accurately represent the quantity of its attribute of weight. The mundane quality of this example disappears when we find ourselves confronted with the need to measure variables of interest to marketers such as intentions, at- titudes, perceptions, etc. How can we be sure that the number “4” correctly captures the intensity with which a respondent holds an intention, for exam- ple? We will devote further discussion to the ways of ensuring good mea- sures in our research in this chapter.

3. The definition states that we attach numbers to the attributes of an ob- ject and not to the object itself. For example, we cannot measure the quan- tity of what you are now holding in your hand. There is no scale for measur- ing the amount of “bookness” in a book. We can, however, measure the attributes of a book—its weight, dimensions in inches, number of pages, and so forth. We can even measure qualities less obvious such as its stature as great literature or its educational value; but, as described in point number 2, the rules for assigning numbers to those attributes will involve different measuring devices than those used to measure its physical properties. This caveat also holds true for the measurement of variables of interest to mar- keters.

We measure a consumer’s attitudes, income, brand loyalty, etc., instead of measuring the consumer. In some cases, such as attitudes, we go a step further and measure the subcomponents of the variable. Attitudes, for ex- ample, are said to consist of cognitive, affective, and conative components that we would want to measure to ensure we have captured the essence of how strong one’s attitude was toward an object. For example, if we are to claim we have measured a parent’s attitude toward a new product concept for a child’s fruit drink, we need to measure beliefs and knowledge (the cog- nitive component of attitudes), how positive or negative he or she feels about the concept (the affective component), and the parent’s predisposition to behave toward the product (the conative component).

4. Scientists in the physical sciences such as physics, chemistry, and biol- ogy have something of an advantage over behavioral scientists because the things they are interested in measuring have a physical reality, and the de- vices used to measure these things can be physically calibrated. “Good mea- sures” are generated by carefully calibrating the measuring devices (e.g., micrometers, weight scales, etc.). Behavioral scientists, such as marketing researchers, cannot see or feel those things of interest to them (e.g., percep- tions, intentions, brand loyalty, attitudes, etc.), and so must find ways of de- termining if the process they use to attach numbers is trustworthy in order to know if the numbers resulting from that process are trustworthy. In other words, while a chemist can trust that the weight of a chemical is what a care- fully calibrated scale says it is, the marketing researcher can trust that he or she has obtained a good measure of intent to purchase only by having faith

in the measurement process used to attach numbers to that intention. There is no way of comparing the numbers on the intention scale to a standardized measure for intentions the way a chemist can check the measures of weight against a standardized scale for weight. We trust the numbers because we trust the process used to attach those numbers.