Price As a 'Give' Component in an Exchange Theoretic Multicomponent Model

Olli T. Ahtola, University of Denver
ABSTRACT - This paper proposes a model that is similar to multiattribute attitude and decision making models, except that the purchase decision is modelled as a decision to exchange ('give') something in order to 'get' something. Usually, the monetary price that needs to be paid for the product or service is the main, or only, attribute on the 'give' side. Negative product attributes are not considered part of the 'give' side but are part of the 'get' package. Furthermore, hierarchical and interacting nature of attributes is explicitly dealt with. Special problems associated with overt behavior predictions are discussed and son solutions are suggested.
[ to cite ]:
Olli T. Ahtola (1984) ,"Price As a 'Give' Component in an Exchange Theoretic Multicomponent Model", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 623-626.

Advances in Consumer Research Volume 11, 1984      Pages 623-626

PRICE AS A 'GIVE' COMPONENT IN AN EXCHANGE THEORETIC MULTICOMPONENT MODEL

Olli T. Ahtola, University of Denver

ABSTRACT -

This paper proposes a model that is similar to multiattribute attitude and decision making models, except that the purchase decision is modelled as a decision to exchange ('give') something in order to 'get' something. Usually, the monetary price that needs to be paid for the product or service is the main, or only, attribute on the 'give' side. Negative product attributes are not considered part of the 'give' side but are part of the 'get' package. Furthermore, hierarchical and interacting nature of attributes is explicitly dealt with. Special problems associated with overt behavior predictions are discussed and son solutions are suggested.

THE NATURE OF PRICE AND ITS SIGNIFICANCE TO CONSUMERS

The price of a product or brand seems to have multiple significance to consumers.

Price Is What You 'Give'

The obvious significance is that the price normally tells what the consumer must 'give' in order to get the product. This aspect is often left unmentioned in behavioral studies about price effects on consumers, except among those who apply economic theories (e.g., Ratchford 1975, Agarwall and Ratchford 1980). In most cases, when buying the product, its (total) monetary price is available to the consumer before the purchase decision is made. Furthermore, in some cases it is possible for the consumer to know what s/he has to 'give' in order to get a given characteristic of a product or brand. For example, s/he may know the extra price s/he has to pay in order to have her/his car with cruise control or air conditioning. However, in most cases the price cannot be broken down to the level of attributes. Or, perhaps, it can be done only for some attributes, and only more or less unreliably. For example a consumer may find two products that seem to differ only in terms of one characteristic and deduct t!lat their price difference reflects the price for that characteristic.

Marketers often Justify higher prices by not only claiming that the product is better in overall sense, but that the price difference reflects a difference in some product characteristic. For example, they may claim that their ham has less fat or water than other brands, and that is why it is more expensive. This also may cause consumers t 'know' something about the prices of certain product characteristics.

In summary, the price is what the consumer has to give in order to get the product, sometimes broken down to the prices for different components or characteristics of the product.

Price-Quality Relationship

Another significance of the price to consumers is that it is used frequently as a cue to infer values of different product attributes. Large amount of research on 'price-quality relationship' (for a review, see Olson 1977) indicates that, especially when not much is known about the brand, the price is used to infer its overall quality. It is quite possible that the price is also used to infer SOD more specific product attributes as well. For example, if we only know the typical prices charged by a restaurant, we most likely make inferences with different levels of certainty whether it is self-service or not, accepts credit cards or not, etc.

Price as Purchase vs. Consumption Attribute

Still another significance of the price is that it is an aspect of instrumental purchasing behavior and may have no, or at least has very different, significance in consummatory consumption behavior. We have given something to get something. It is what we have that we consume, not what we have given away. Of course, the price that was paid can keep operating as information and rationalization cue, and may be a very important determinant of purchase and consumption satisfaction. However, if the consumer knows that s/he got a product s/he is quite familiar with at a discount or special price, there is no reason why s/he should perceive that a somehow different product was purchased and consumed. Or, if a consumer buys Exxon super unleaded at $1.48 in a small town and then the same brand at $1.29 in a bigger town, s/he most likely feels that the latter was a better deal, but that the same product was bought and consumed, i.e., its intrinsic characteristics are the same.

Operating Characteristics of Price as an Attribute

As to the operating characteristics of the price, as compared to intrinsic product attributes, there are many similarities. The price certainly is often used in conjunctive fashion. Product exceeding a certain price is totally excluded from consideration based solely on that information. This most likely reflects the 'give' aspect of the price. That is, too much would have to be given. There is evidence that this kind of conjunctive limit exists also at some low end of the price continuum (e.g., Monroe 19/9). This most likely reflects the cue properties of the price. That is, a product that is very cheap must have something wrong with it (cf. Hansen 1972). Given this reasoning for the lower conjunctive limit of the price, we could hypothesize that this limit disappears when products are well known. Why wouldn't a consumer accept a product s, he knows to be the best if it is offered real cheap?

It is also highly likely that, at least within some range, the price is utilized in a compensatory fashion. Also, the price may be for some consumers, in some circumstances, a disjunctive criterion. Bargain hunters seem to use the price in this fashion.

Given the above reasoning and empirical evidence, how should consumer purchase decisions be modelled to validly reflect the influence of the price to final outcomes? The purpose of the rest of the paper is to point to some directions, and raise several crucial issues that need to be resolved.

TOWARD A MODEL

The basic orientation when developing the forthcoming model is not to develop a model of consumer reactions to prices but a more general product purchase model. It is, however, believed that this rather general model more adequately reflects the true impact of prices than the existing similar models.

TABLE 1

PARTIAL DIAGRAMMATIC DESCRIPTION OF THE MODEL : AN EXAMPLE

In the existing, psychology based, multiattribute models, be they compensatory, conjunctive, disjunctive, lexicographics or some other variation, it is typical that the price is included as a product attribute and its impact is considered similar to any other attribute. In fact, the compensatory model does not in any way categorize the salient attributes. The above mentioned noncompensatory models do point out that different attributes may be utilized in different fashions. However, they do not, a priori, help to hypothesize which attributes are used by consumers in which fashion. It is hypothesized here that the price is similar to many intrinsic product attributes in that it can operate as a compensatory, disjunctive, conjunctive, and/or lexicographic criterion. How to predict, a priori, which role a given attribute takes in a given purchase situation remains unsolved.

However, it is proposed here that we need to identify, in separate categories, on one hand those attributes that a consumer feels are acquired from the ownership and consumption of the product, and, on the other hand, those attributes that are involved in obtaining the product. The first set of attributes will be called 'get' attributes, and the latter set 'give' attributes.

It has earlier been suggested that negatively valued attributes should be separated from positively valued ones because the former may need be weighted more (or otherwise differently)(references not found). This is not what is being proposed here. It is quite possible, and often likely, that some 'get' attributes are negatively valued. Similarly, it is possible that some 'give' attributes are positively valued, i.e., we prefer giving them away from retaining them. This can be possible, for example, in some barter situations, when what we 'give' has a negative 'have' value for us, or when our purchase helps a worthwhile cause of the seller, or when the purchase act has some other valued or pleasant attributes, such as excitement and enjoyment of bidding in an auction.

It is hypothesized that for the purchase (exchange) to take place, 'give' plus 'get' must be positive. The price, of course, is normally the major component on the 'give' side in economic transactions. And, for most of us, giving an amount of money away is negatively valued. What has been proposed above, does not, from a mathematical point, require that the 'get' and 'give' components are 'on different sides of equation'. However, from decision making process and measurement points of view this division is important.

Decision Making Process

From the process modelling point of view, it is important to make the 'give' versus the 'get' categorization because it is hypothesized that customers contrast in their decision process 'what they give' against 'what they get'. This comparison may take place at different 'levels'.

It is argued here that attributes (whether intrinsic product characteristics or not) can, and in many cases do, exist at different and hierarchical levels. For example, a consumer may start with descriptive, defining characteristics, and possibly evaluate them one by one. Using a car purchase as an example, s/he may start by observing that a certain car's engine has eight cylinders, 5.7 liters, four throat carburator, and that it is a gas engine. These four characteristics may, or may not, have separate evaluative meaning to her/him. However, based on this information, s/he probably derives a cognition 'powerful engine'. This intermediate summary cognition most likely has evaluative meaning to her/him.

It is hypothesized that if the evaluative meaning exists for lower level attributes and if these lower level attributes to not interact (in factorial ANOVA sense) the evaluation of the higher level attribute is similarly determined than the overall evaluation (attitude) toward the product. That is, it is proposed to be a linear compensatory function of the values of its attributes, or perhaps some noncompensatory function of them. (It is believed, however, that evaluation processes are likely to be compensatory, while the decision processes may or may not be.) It is also hypothesized that if the lower level attributes are known to interact, they are not evaluated separately, but the evaluation is postponed to the level of attributes that are considered additive.

In case of new experiences,it is hypothesized that a consumer tentatively assumes the attributes to be additive, but may learn that it is not the case, and learns to postpone the evaluation until the more complete picture (i.e., a higher level attribute) is available.

These intermediate attributes may exist at different levels. Using the same car example, 'powerful engine' together with some other attributes may lead to an attribute 'fast car', the 'fast car' together with some other attributes may lead to an attribute 'sporty car', etc. (cf. Myers and Shocker 1981).

It is also possible that an attribute may elicit other attributes that contribute to its evaluation. For example, 'sporty car' may elicit 'would impress my girl friend'. This will affect the evaluation of 'sporty car' in addition to those attributes that elicited it in the first place.

If an attribute is used to infer another attribute and evaluate it, there seems to be no reason why these inferred attributes would operate any differently from the observed attributes. Consequently, the much investigated 'price-quality relationship' need not to be a problem from the modelling point of view.

If the price of the attribute is available to the consumer, s /he most likely makes some value judgements (in good/bat deal sense) at had level with respect to that characteristic (cf. Park, Lessig and Merrill 1981). If only the overall price is known, the good/bad deal judgement has to be postponed until the overall value of 'get' characteristics has been derived. If some attribute specific good/bat teal judgements have been made, they will strongly influence the overall good/bad deal judgement. However, it is very seldom, if ever, when the consumer is in the position to put the separate cost to all salient attributes that are the immediate determinants of the overall 'get' package.

In the above discussion, 'price' has been used as a proxy for the total 'give' package. It should be kept in mind that other things besides monetary price may also be part of the 'give' component, as was discussed earlier.

Measurement Issues

When we measure values of attributes, it is customary to use identical scales to measure them all. It is also typical that we have identified several attributes that are 'get' attributes, but normally only one -price- that is a 'give' attribute. It is very likely that a typical respondent reacts to our scales in a relativistic fashion, i.e., s/he anchors the scales separately for every attribute, sets average attribute level in the middle, most desired reasonable level on one extreme and the most disliked, but still conceivable, level on the other extreme of the scale. This works reasonably well if all the attributes are of similar nature. However, if the value of the price attribute is what the sum of the values of the other attributes is compared against, it is very likely that it will be grossly underweighed using the customary measurement techniques. If we conceptualize the 'give' and 'get' attributes as the "two sites of the coin", we are more likely to develop a measurement method that better reflects their relative contributions.

COMPARISONS TO PREVIOUS WORK

At this point it is useful to compare the above conceptualization to some similar models and research practices.

The Concept of 'Quality'

A commonly seen concept in pricing research is 'quality'. Often the concept 'quality' is operationalized as whatever it means to the respondent. Sometimes 'quality' is defined as "the ability of the product to provide satisfaction" (Monroe 1979, p.38). According to dictionaries, 'quality' is the term used to describe grade of excellence, often in contrast with quantity. It is quite possible that the ability to provide satisfaction is as much, or more, a function of quantity than quality, or perhaps sometimes a function of neither ('looks good', 'impresses neighbors', etc.), and, consequently, the 'quality' operationalizations to not measure what they are intended to measure. Assuming that we have measured 'perceived ability to provide satisfaction', we still have the problem to prove that that is what is contrasted with 'price' to make the value judgement about the goodness of the deal. The suggestion here is that the satisfaction concept may not truly parallel the meaning of the 'get' component (ant the quality concept even less so). This is because satisfaction most likely is the function of the deal, not only function of the 'get' component. Consequently, the measurement problem for the overall 'get' component has not been solved.

It d seem that the value of the 'get' component has to be measured relative to the state of 'not getting or having', ceteris paribus, i.e., under the circumstance when nothing needs to be 'given'. People are willing to pay excessive, dissatisfactory prices on gasoline on the lone gas station in the middle of nowhere when the alternative of running out of gas is a much worse alternative. It is suggested here that the state of being without the product should be used as the baseline against which the value of product alternative(s) should be measured.

The lack of meaningful baselines may explain why many attitude models often predict purchase behavior so poorly. If the conceptualization proposed here is correct, it is quite possible to have situations where very positive attitude exists toward a product or brand, but no purchase takes place, even when resources to get it are available and when it is considered a good deal. That can be when being without it is no worse than what has to be 'given'. Also, as the above gas example demonstrates, it is quite possible for a purchase to take place when the deal is considered bad.

Attitude Toward an Act

Even though the Fishbein's attitude toward- an act model (e.g., Fishbein and Ajzen 1975) is more likely to include various attributes related to acquiring the product than most other attitude models, it still has all the other problems mentioned above. Furthermore, the extended model with the normative component does not address any of the above issues. As to the normative influences, in the proposed model they are considered attributes that need not be separated from other attributes. It is not an attitude model.

Other Problems

Another problem with previous research is that the hierarchical nature of attributes is not recognized. This leads to double and triple counting of the impact of some attributes. It is not at all uncommon to see attributes like 'high quality', 'good value', 'reasonable price' in the same attribute list. Techniques to elicit and organize attributes need to be revised. Also, the interactive nature of many attributes goes unrecognized or ignored.

PURCHASE BEHAVIOR

The whole issue of the nature of "overt behavior" has received little attention in consumer behavior literature. Usually some single ratio scale or dummyable measure is taken to fit requirements of a familiar statistical technique. In many cases these behavior measures may not be the most meaningful. Instead of using a common ratio scaled magnitude measure, it should be more useful to use what we might call "amplitude" measure. In the amplitude measure the first stage is a dichotomy in terms of whether the behavior took place or not. The second stage, which is relevant if the behavior took place, measures its magnitude. The difference between this magnitude measure and the traditional magnitude measure is that the latter includes magnitude zero while the former takes only strictly positive values. This qualitative distinction between no purchase and some purchase is consistent with the above conceptualization.

The same reasoning appears relevant if we measure behavior in terms of rate, i.e., how much in a given time period, or in terms of allocation, i.e., how much of each among the choice set.

The proposed model was conceptualized in the 'will buy will not buy' dichotomous framework, but should generalize to other behavior types. What the exact functional relationship between, say, the magnitude of 'give' plus 'get' and behavior magnitude, rate, or allocation is, becomes an interesting research issue if empirical work supports the validity of the model.

REFERENCES

Agarwall, Manoj R. and Brian T. Ratchford (1980), "Estimating Demand Functions for Product Characteristics: The Case of Automobiles, " Journal of Consumer Re- search, 7 (December), 249-262.

Fishbein, Martin and Icek Ajzen (1975), Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research, Reading, MA: Addison-Wesley.

Hansen, Flemming (1972), Consumer Choice Behavior: A Cognitive Theory, New York: The Free Press.

Monroe, Kent 3. (1979), Pricing: Making Profitable Decisions, New York: McGraw-Hill.

Myers, James H. and Allan D. Shocker (1981), "The Nature of Product Related Attributes," Research in Marketing, 5, 211-236.

Olson, Jerry C. (1977) "Price as an Information Cue: Effects on Product Evaluations." In A.G. Woodside, et. al., eds., Consumer and Industrial Buying Behavior, New York: Elsevier North-Holland.

Park, C. Whan, V. Parker Lessig and James R. Merrill (1981). "The Elusive Role of Price in Brand Choice Behavior," Advances in Consumer Research, 9, 195-200.

Ratchford, Brian T. (1975), "The New Theory of Consumer Behavior: An Interpretative Essay," Journal of Consumer Research, 2 (September), 65-75.

----------------------------------------