A Discussion of Theoretical Development and Extensions in Consumer Behavior

Moshe Handelsman, University of Santa Clara
ABSTRACT - Three papers, concerning Brunswik's lens model, a composite MCI model, and variety seeking are discussed. It is shown that the Lens Model differs from the MCI model with respect to the need for accuracy of consumer's judgment of the importance of attributes. The Variety Seeking research is still in an early stage, thus, the need for a clear definition of the variety seeking construct and the difference between variety-seeking and innovation-seeking are emphasized.
[ to cite ]:
Moshe Handelsman (1984) ,"A Discussion of Theoretical Development and Extensions in Consumer Behavior", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 120-121.

Advances in Consumer Research Volume 11, 1984      Pages 120-121


Moshe Handelsman, University of Santa Clara


Three papers, concerning Brunswik's lens model, a composite MCI model, and variety seeking are discussed. It is shown that the Lens Model differs from the MCI model with respect to the need for accuracy of consumer's judgment of the importance of attributes. The Variety Seeking research is still in an early stage, thus, the need for a clear definition of the variety seeking construct and the difference between variety-seeking and innovation-seeking are emphasized.


The discussion of the three papers by Tapp, Cooper and Finkbeiner and Hoyer and Ridgway will be conducted along two dimensions of consumer behavior research. The first dimension is the degree of theoretical and methodological advancement. The second dimension is the consumer purchase decision process: that is, what stage or stages does each model address? The discussion will conclude with data requirements common to all three models presented in the papers.


The three models are not at parallel levels of development. The MCI model is in the most advanced stage compared to the Lens model, while the Variety Seeking research is the "youngest" one.


We will begin with the MCI model: Cooper and Finkbeiner t S paper is an extension of the basic MCI model, which is in turn, a special case of the Luce choice axiom.

Major contributions to this model in the past are:

1. a model developed by Huff to determine selection of shopping centers (1962).

2. the formulation of the MCI model by Nakanishi and Cooper (1974).

3. Nakanishi and Cooper's method of using dummy variables to reduce the computational burden of the estimation (1982).

And now we have a refinement aimed at overcoming the problem of using the model in situations that allow few degrees of freedom. The solution: employ subjective reports of the importance of explanatory variables instead of estimating them.

My question: are we trading off accuracy for ease of estimation? In other words, how much accuracy do we lose (if any) when we use the suggested model? I suggest that a data set will be used with the suggested MCI model and with the prior version (where B's are estimated) and results will be compared.

Another point, and this time it is from a practitioner perspective: when I am testing a new product and have to make a "Go-No Go" decision, the MCI can be a very helpful tool. In many cases the expected market share is relatively small (depending on the product category) and in such situations the MCI has not proven to be accurate. It seems that further research directed at improving prediction of small market shares will be useful.

Lens Model

The Lens model was originated by Egon Brunswick and was later employed by cognitive psychologists to study the process of clinical judgment. Tapp adapts this model to consumer research.

The question is: can we overcome the limitations of this model when applied to consumer research?

At present, the major limitation is that the model requires a true measurement of the correlation between the criterion and the attribute. As long as we define the criterion as the post purchase satisfaction - by definition - it is a subjective/perceptual criterion.

Is it possible to replace the criterion variable with a more objective and measurable one? This may serve as a topic for further research. Let us assume that the problem of objective measurement has been solved. In that case, the Lens model can serve as a tool to screen attributes and product categories before we submit them to other models, such as MCI. I will discuss this application of the Lens model later.

Variety Seeking

The Variety Seeking model is the least developed of the three in terms of theoretical and methodological refinement. In this descriptive stage, we are formulating framework; it is therefore important to have a clear definition of the notions under discussion. In other words, what do we mean by variety? Let me quote four descriptions of variety seeking as they appear in this paper:

1. "Purchase exploration through brand switching or innovating"

2. "The desire for new and novel brands/products"

3. "Variety drive (i.e., optimal stimulation level)"

4. "The selection of a new brand or a new product"

The concept of variety seeking takes on two meanings, depending on the limitations we attach to the notion of "new/novel."

If new/novel is restricted to stimuli that the consumer has never before used, then we can place the above definitions into two different categories:

1. Satisfaction of variety drive through brand switching: the variety drive is assumed to result from the change alone; so brands purchased in the past will satisfy the variety need as well as new/novel brands.

2. Satisfaction of variety drive through innovation: here new/novel stimuli as defined above are the only stimuli that will satisfy the consumer's need for variety.

If we accept the first definition, variety seekers may satisfy their need for variety through innovation. However, under the second definition, variety seekers cannot satisfy their needs with brands that are not new/novel.

On the other hand, we can allow the definition of new/novel to take on the meaning of a stimulus that is "relatively new" to the consumer (as modelled by Berlyne (1950) and Jeuland (1978)). This "relatively new" notion is based on the consumer's experience with a brand. As a consumer accumulates experience with a brand through repeat purchases, his preference for this brand declines. If the brand is not purchased over a period of time, the consumer's experience will fade. Thus the brand regains its novelty and its ability to satisfy the consumer's need for variety. In this case, where new/novel is a relative issue, both variety categories - variety seeking through switching and innovating - converge.

In sum, the definition of variety seeking should be clarified. If new/novel is taken to signify absolute newness, it is suggested that the first definition (i.e., variety seeking through switching) will be adopted, because the second (innovation) is nested within the first.

I have two other comments on hypotheses suggested in the variety seeking paper, before moving on to the place of the models in the Purchase decision process.

One hypothesis suggests that the lower the interpurchase frequency is, the less variety seeking is expected, such as in product classes like watches or automobiles It is suggested that we also be alert to interconsumption frequencies. For example: a watch is "consumed" every time we look at it, which may lead to boredom/satiation with that particular brand, thus leading to variety-seeking when repurchasing a watch.

Some of the hypotheses in the paper have already been supported by prior research (Handelsman 1982); more varied seeking occurs when: (a) the perceived difference between brands is low, (b) the perceived risk of product failure is low, (c) the assortment in a product class is relatively wide.


The consumer purchase decision process consists of five stages:

1. Need Arousal

2. Information Search

3. Evaluation of Alternatives

4. Purchase Decision

5. Postpurchase Feelings

The models under discussion address different stages in the process, as illustrated in figure one.



We see that all three models overlap with respect to the consumer's evaluation of attributes. Here we have important differences between the MCI and Lens models.

The MCI model assumes that consumers are able to quantify their perceptions of the level of each attribute, as well as being able to give an accurate estimate of the importance of each attribute's influence on their purchase decision. The Lens model, on the other hand, questions the consumer's ability to produce objective judgments.

Therefore, when the previously described limitations on the Lens model are overcome, we may employ the Lens model to screen attributes or products. This screening will identify those attributes or products for which consumers cannot report objective judgments, thus their information cannot be submitted to further analysis using MCI (or similar) techniques.

The Variety Seeking model is concerned with stages prior to the purchase decision, in part because the model is in an early stage of development. It is predicted that the model will incorporate the purchase decision stage in the future, as the model is further developed to predict when, and to which brand, consumers will switch when motivated by the variety drive.


The three papers discussed above are on the state-of-the-art edges in different areas of consumer research. However, these models share one easily-observable problem: the data requirement problem. The composite MCI model was motivated by a lack of degrees of freedom; the Lens model lacks data to prove its usefulness; and the variety-seeking model will require longitudinal data containing purchase histories and continuous measurements of attitudes, environmental factors and personal traits corresponding to purchase histories. This kind of data is costly and complicated to obtain. The Variety Seeking model will particularly be difficult to validate.


Berlyne, D.E. (1950), "Novelty and Curiosity as Determinants of Exploratory Behavior," The British Journal of Psychology, 41, 68-80.

Handelsman, Moshe (1982), "Varied Buying Behavior: Measurement Based on Perceptual and Purchase Diary Data," unpublished Ph.D. dissertation, Krannert Graduate School of Management, Purdue University.

Huff, David L. (1962), "Determination of Intraurban Retail Trade Areas," Real Estate Research Program, University of California, Los Angeles.

Jeuland, Abel P. (1978), "Brand Preference Over Time: A Partially Deterministic Operationalization of the Notion of Variety Seeking," in the AMA Research Frontiers in Marketing, Dialogue and Directions, 1978 Educators' Proceedings, Series No. 43, Subhashe Jain, ed.

Nakanishi, Masao and Cooper, Lee G. (1974), "Parameter Estimation for a Multiplicative Competitive Interaction Model - Least Squares Approach," Journal of Marketing Research, 11, 303-311.

Nakanishi, Masao and Cooper, Lee G. (1982), "Simplified Estimation Procedures for MCI Models," Marketing Science, 1, 314-322.