Main Content Area
Loosey-Goosey Measurement of Purchase Intention
Two of the most measured constructs in studies of consumers, at least in scholarly research, are Attitude Toward the Brand and Purchase Intention. Those are two very distinct constructs. (A “construct” to those of us in the social sciences is a theoretical variable that can not be directly observed.) A person can have a good attitude about a brand and yet have little or no plan to buy it. In-between those two constructs, however, are several other constructs researchers have said they are measuring but which are not as clearly distinct: Attitude Toward the Act, Willingness to Purchase, and Likelihood of Purchasing. These are used as synonyms for purchase intention by too many researchers but I do not believe they are necessarily the same. My concern is that if you do not know exactly WHAT you are trying to measure then you will not know very well HOW to choose the proper measurement scale!
First, let me briefly describe the constructs without using too much scholarly verbiage and theoretical discourse. The first couple of constructs below are definitely different from the ones discussed afterwards having to do with the purchase itself but I will start with them because they are such critical precursors. By the way, these descriptions are drawn from my years of studying, teaching, and researching consumer behavior as well as my experience reviewing scales for nearly 30 years. Having said that, I am not going to be dogmatic about the issues because I, too, am seeking some clarity.
Attitude Toward the Product (Ap): This refers to a person’s beliefs and feelings about a product in general without reference to a particular brand, model, or version. A simple example statement that reflects the construct is Generally, I like beer.
Attitude Toward the Brand (Ab): More specific than Ap, Ab focuses on a person’s beliefs and feelings with respect to a particular brand of a product. While we would expect a person’s Ab to be strongly related to his/her Ap, they are not the same. That is reflected in a statement such as I generally like beer but not Brand X or I do not like most beers but Brand X is okay.
Attitude Toward the Act (Aact): In consumer behavior, this usually has to do with a person's attitude toward an activity involving the product. The most popular activities studied by consumer scientists have been purchasing a product and adopting innovations but many other uses have been measured. Measuring this construct makes sense when the act being studied is hypothetical. For example, the product may be real but not yet available to consumers. Other times, the product has not been produced yet and a goal of the study is to estimate how much demand there is for it.
Willingness to Purchase (WtP): This has to do with how open-minded a consumer is to buying the product but without a clear plan to do it. In other words, the person may be willing to purchase the product but is not expressing an clear intent to do so. I view this construct as a special case of Aact. Also, as with Aact, this construct is appropriate to measure when the purchase is hypothetical. That is because consumers may know if they are willing to buy a proposed product even if it is not on the market and even if they have not developed a clear PI yet.
Likelihood of Purchasing (LoP): More like PI and less like WtP, this has to do with a consumer’s expression of the probability that he/she will buy a particular product. While it might be used with respect to hypothetical products, it appears to be better suited to those situations where respondents are already familiar with the product. In fact, the more information the consumer has, the more the measure is like a PI scale. For example, if respondents are asked what is the likelihood that you will purchase Brand X in the next month, I would consider it very close to if not the same as a PI measure because it identifies a real product that could be purchased in the near-term. On the other hand, if the question is vague, the product is hypothetical, and the time-frame is unstated or far-off then the construct being measured is more like WtP.
Purchase Intention (PI): This is the construct that most researchers prefer to measure if possible because it is the degree to which a consumer has a clear plan to buy a particular product. What makes this construct so valuable is that, when measured correctly, it can be the best predictor of the actual purchase. That is why the construct is included in so many predictive models of purchase behavior. Of course, for a measure of the construct to have high predictive validity with regard to a purchase, it needs to be precise about the product and the time period. The more specific the product (brand, model, version) and the more proximal the time period consumers are asked about, the more predictive the measure. In other words, consumers can “predict” with greater accuracy what they plan to buy tomorrow compared to what they will buy a year from now. A one-question example of measuring PI would be asking consumers to agree/disagree with I intend to buy Brand X in the next week.
I hope it is clear that some of these constructs are quite distinct from the others while some are very similar. My strongest suggestions to researchers are to take care in naming and referring to these related constructs, really think about which construct they want to measure, and carefully select the best measure of the construct. If at all possible, borrow measures that have good track records in past studies. Further, do not rely on the names that previous users have given scales; measures can be called one thing but really measure something else. With multi-item scales, scrutinize the items as well as the question and/or instructions used with them. It is only then that you can determine what construct is being measured and how well it fits what you need for your research situation.