A Path Analytic Model of External Search For Information For New Automobiles

Narasishan Srinivasan, State University of New York at Buffalo
ABSTRACT - Decision making is dependent on the information inventory that consumers build up, and the study of information gathering becomes critical in understanding final choice behavior. This paper presents a meaningful conceptualization of a model of external search for information. The cost-benefit framework is enriched with the addition of prior beliefs and perceived risk. Individual differences, such as knowledge, experience, involvement and goal-orientation have been included in the explanatory model. The model is validated empirically, using path analysis.
[ to cite ]:
Narasishan Srinivasan (1987) ,"A Path Analytic Model of External Search For Information For New Automobiles", in NA - Advances in Consumer Research Volume 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT : Association for Consumer Research, Pages: 319-322.

Advances in Consumer Research Volume 14, 1987      Pages 319-322


Narasishan Srinivasan, State University of New York at Buffalo


Decision making is dependent on the information inventory that consumers build up, and the study of information gathering becomes critical in understanding final choice behavior. This paper presents a meaningful conceptualization of a model of external search for information. The cost-benefit framework is enriched with the addition of prior beliefs and perceived risk. Individual differences, such as knowledge, experience, involvement and goal-orientation have been included in the explanatory model. The model is validated empirically, using path analysis.


Every individual is constantly being exposed to information stimuli from a variety of sources, including personal and non-personal sources. Due to processing capacity limitations (Miller 1956; Bettman 1979) and selective perception (Engel, Kollat and Blackwell 1982), only a part of the information one is exposed to is actually digested and stored in memory. The information inventory that an individual builds up is critical because any choice is dependent on this data-base. The assumption of full information made in studies using the multi-attribute model or multi-dimensional analysis is not strictly true.

For non-durables, it may be sufficient for a consumer to recall information that has been previously stored in memory (internal search); in the case of durables, a consumer would more likely find internal search to be inadequate and hence may engage in external search. The study of external search, which is easier to observe, provides a researcher with an opportunity to understand the building up or revision of the information inventory that precedes the final decision. The objective of this investigation is to posit a model of external search behavior and empirically validate it.

Automobiles have been chosen as the product class for this study because (i) it is an important purchase, accounting a sizeable portion of an individual s income (ii) it is so ubiquitous and there are a variety of sources for information, and (iii) a number of earlier studies have shown a large variance in the search behavior for information, with a sizeable proportion not engaging in extensive search (Newman 1977).

Previous Relevant Literature

The economic and psychological perspectives found in the marketing literature are reviewed briefly.

The economic approach to search behavior, concentrating on the cost-benefit framework provides a parsimonious and meaningful framework to investigate search behavior. The optimum amount of search can be arrived at by equating expected marginal returns of search to the expected marginal costs of search. Being ignorant has a penalty and reducing ignorance comes only at a cost. Taking price alone, Stigler (1961) algebraically showed that the expected diminishing returns of search accounts for search to be quite limited. Taking real prices of refrigerators, a monetary value of the "welfare loss" (difference between the utilities of the best alternative and the chosen alternative) points to the same conclusion - a limited amount of search is sufficient to obtain the lowest price in the market - about four alternatives (Ratchford 1980). The main drawback of the economic approach is its exclusive concentration on purely monetary costs and monetary returns, with the psychological costs and benefits being neglected.

The cost-benefit framework has been expanded to include risk (Meyer 1982) but a restricting assumption of search occurring only across brands distracts from its usefulness. Considering a Bagesian approach, the Expected Value of Sample Information (EVSI) model concentrates on successive information bits available for processing (Hagerty and Asker 1984). Besides being myopic, the EVSI model works best only when there is homogeneity in attribute perceptions and attribute weights.

Turning to the traditional psychological approach to search behavior, the motivational approach is well established (Howard and Sheth 1969; EKB 1982). Motivation is viewed as providing the drive for any activity. One aspect of motivation is involvement. The higher the involvement, the higher would be the propensity to engage in activities related to the concerned product class. Another psychological dimension that accounts for heterogeneity amongst consumers is the differential goal-orientation. Consumers faced with the same market environment would search differentially not only because the costs of search are different, but also the strategy chosen for making a choice mag vary - whether the goal is to satisfice or optimize (Wright 1975). A satisficing objective would require a less thorough search than an optimizing objective.

The information processing approach (Bettman 1979; Sternthal and Craig 1982) lays stress on the processing capacity limitations and also deals with internal search more extensively. Familiarity and experience are constructs that relate to information that has been previously acquired and stored. The relationship of familiarity to search has been ambiguous - positive, negative, no influence, or an inverted U (Katona and Mueller 1955; Newman & Staelin 1972; Bettman & Park 1980; Moore & Lehmann 1980; Punj & Staelin 1983; Johnson & Russo 1984; Brucks 1985). The positive effect can be understood to be due to the facilitating influence of prior knowledge on the understanding of new information; the negative effect can be understood to be due to the efficiency influence on the processing of new information i.e. knowledgeable consumers selectively gather only relevant information, like a chess expert knowing the right moves. The effect of positive experience on search has been fount to be negative when the previous purchase has been satisfactory leading to the repeat purchase behavior or loyalty (Swan 1969). Empirical studies have typically dealt with univariate or bivariate analysis or typologies of search segments (Katona and Mueller 1955; Dommermuth 1965; Udell 1966; Bennett and Mandell 1969; May 1969; Claxton Fry and Fortis 1974; Westbrook and Fornell 1979; Riel and Layton 1981; Duncan and Olshavsky 1982; Furse, Punj and Stewart 1984). In multivariate analysis, the relationships diminish or disappear (Newman and Staelin 1971; 1972).

The Punj-Staelin Study

A major gap that the literature presents is the absence of any explanatory motel to explain the differences in the amount of external search. A previous attempt (Punj and Staelin 1983) to build an explanatory motel leaves much room for improvement. That particular study provides a reliable measure of the amount of external search, which has been arbitrary in earlier studies. The authors themselves state that "(the findings) indicate that the model of consumer information search process, although capturing some important aspects of the process, still ignores numerous factors concerning search behavior. Thus the problem seems to be one of developing better model specification ..." . This study seeks to develop an improved explanatory model of external search behavior by reexamining the constructs and linkages and by positing some important missing constructs.

* Cost Savings has been used as a benefit accruing from search: A better construct would be the use of perceived benefits as an antecedent to search itself. What happens after the search would be less useful than the effect that the expectation of cost savings has on the search process, when the objective is to explain search behavior. Perceived benefits (which are expected) is an antecedent to search and fits very well in the cost-benefit framework.

* Size of the fesible set was used. Not every available alternative is considered for evaluation even if it feasible. Typically, the size of the considered alternatives is small (Newman 1977); the size of the evoked set would be a better construct, because this reflects the self-imposed restrictions of the individual.

* Perceived risk theory (Bauer 1960) deals with the uncertainty in choice situations, which is characteristic of the pre-purchase process, especially for durables. Risk reduction is a benefit resulting from search and hence risk is an antecedent of the perceived benefits. The greater the perceived risk, the greater would be the perceived benefits expected by searching for information.

* Prior beliefs about the marketplace: If there is no perceived heterogeneity in the product offerings or the environment is not changing much, then the amount of search need not be high. However, if the market is perceived as volatile or changing, the internal search would be less useful and external search would be enhanced. Alternative limiting prior beliefs have a significant impact on the amount of external search (Duncan and Olshavsky 1982).



The exogenous constructs are costs of search, degree of involvement, goal-orientation, experience, prior knowledge and prior beliefs. The endogenous constructs are the perceived risk, perceived benefits of search, size of the evoked set and the amount of external search.

Essentially, it is an enrichment of the cost-benefit framework. Costs will affect search negatively and perceived benefits will influence it positively. The size of the evoked set, the goal-orientation (measured as the need to do a thorough job), and involvement affect search positively whereas alternative limiting beliefs influences it negatively. Perceived risk would be diminished by greater familiarity and enhanced by the goal orientation and involvement. Similarly, perceived benefits would be positively affected by perceived risk, involvement and goal orientation and negatively by familiarity.

Research Design


Questionnaires were mailed out to participants in a panel study. The main decision-maker in the household was requested to complete the questionnaire if he/she had purchased a car within the previous twelve months. Otherwise, they were requested to pass the questionnaire along to a friend who had purchased a car within the specified time frame. The subject was asked to consider himself/herself in a new car purchasing situation and responded to Likert-type (seven point scale) statements. The total number of usable responses was 129. The survey was conducted in Dec. 1985/Jan. 1986.


Multiple items were used to measure each construct, except the size of the evoked set . The reliability of the scales is shown in Table 1.



Perceived risk had only two items; The correlation between the two was 0.41. The size of the evolved set was 8 single item measure: the mean was 4.0 and std. dev. was 2.5.




A path analysis was run using cost, experience, knowledge, involvement, and prior beliefs as the exogenous variables. The order of entry of the endogenous variables was as follows: perceived risk; perceived benefits; size of the evoked set. The dependent variable vas the amount of external search. The validated path motel is shown in Figure 8; the computing schedule is shown in Table 2 and the correlational analysis in Table 3.





Discussion of the Results:

The measures used for the various constructs have moderate to high reliabilities, ranging from 0.60 to 0.95. These scales are an improvement over measures used in earlier studies. In path analysis, the assumption has to be mate that the variables are measured without error, in addition to the other usual assumptions of regression analysis. The reliabilities obtained show that measurement error, if any, would not vitiate the results. Also, using 129 subjects and 10 variables, (ratio of 13:1), lends stability to the regression co-efficients..

The path analysis results support the cost-benefit framework. The perceived benefits of search has a positive influence and costs exerts a negative influence on the amount of external search conducted. Involvement and goal orientation have a positive impact on perceived risk, perceived benefits and the amount of external search. These were all expected, except goal orientation affecting perceived risk. Perceived risk vas not related to experience as vas hypothesized and knowledge affected risk positively instead of negatively. The reason for this may be any of the following:

i) the measure of risk is inadequate

ii) subjects are over-stating the riskiness to corroborate or enhance their own self-image to be congruent with knowledge

iii) this is a tricky situation where greater knowledge increases the perceived risk. An analogy is the complexity perceived by consumer researchers in any decision making process, whereas in actual situations, an individual goes through many purchase decisions with relative ease.

iv) the facilitating explanation may be holding true

Maybe each of the above is contributing to the observed positive relationship between knowledge and risk. This must remain a speculation until a follow-up study using a better measure of risk is completed.

The size of the evoked set positively affects the amount of search and positive experience with the previous automobile reduces the size of the evoked set. In addition, positive experience directly affects the amount of search negatively, thereby confirming the importance of experiential knowledge. Prior beliefs were posited to negatively affect the amount of search and this vas also supported.

Examining Table 3, it can be seen that the total causal relationship is direct in the case of perceived benefits, size and belief and slightly less so in the case of cost, experience and goal orientation. Risk performed very poorly; and the measure remains a drawback.


The cost-benefit framework is useful for examining the external search behavior. In addition to prior beliefs, the goal orientation, involvement, the self-limitation of size of the evoked set and experience have a significant impact.

Limitations of the Study

The sample vas not comprised just of new car buyers and the amount of forgetting is not certain. The only known fact is-that every respondent had purchased a car in the previous year. Also, the sample vas not a strictly random one.

One cannot confirm true causality through a cross-sectional analysis but the selection of the variables and the order of entry of the endogenous variables is meaningful and hence the path analysis appears to be reasonably valid. This is a preliminary study and a follow-up investigation using the validated scales is planned. Also a larger sample size and less restrictive analysis would throw more light on this important phenomenon.


Bauer, A. Raymond, (1960), "Consumer Behavior as Risk Taking" in 'Dynamic Marketing for a Changing World'. et. by R.S. Hancock, Proceedings of the 63rd Conference of the American Marketing Association, 389-398.

Bennett, P.D. and Mandell, R.M. (1969), "Prepurchase Information Seeking Behavior of new Car Purchasers-The Learning Hypothesis," Journal of Marketing Research, VI (November), 430-433.

Bettman, R. James (1979), An Information Processing Theory of Consumer Choice, Reading, Mass.: Addison-Wesley Publishing Company.

Bettman, R. James and C. Whan Park (1980), "Effects of Prior Knowledge & Experience & Phase of the Choice Process on Consumer Decision Process," Journal of Consumer Research, 7 (December), 234-248.

Brucks, Merrie (1985), "The Effects of Product Class Knowledge on Information Search Behavior," Journal of Consumer Research, 12 (June), 1-16.

Claxton, J.D. and Fry, J.N. and Fortis, B. (1974), "A Taxonomy of Prepurchase Information Gathering Patterns," Journal of Consumer Research, I (December), 35-62.

Dommermuth, P. William (1965), "The Shopping Matrix and Marketing Strategy," Journal of Marketing Research, 2 (Hay), 128-132.

Duncan, C.P. and Olshavsky, B.W. (1982), "External Search: The Bole of Consumer Beliefs," Journal of Marketing Research, XIX (February), 32-43.

Engel, F. James and Blackwell, D. Roger (1982), Consumer Behavior, IV Edition, The Dryden Press.

Furse, D.E. and Punj, G.N. and Stewart, D.w. (1984), "A Typology of Individual Search Strategies Among Purchasers of new Automobiles," Journal of Consumer Research, 10 (March), 617-631.

Hagerty, H.R. and Aaker, D.A. (1984), "A Normative Model of Consumer Information Processing," Marketing Science, 3 (Summer), 227-246.

Howard, A. John (1977), Consumer Behavior: Application of Theory new York: McGraw-Hill Book Company.

Howard, A John and Sheth, N. Jagdish (1969), The Theory of Buyer Behavior, new York: John Wiley & Sons, Inc.

Johnson, Eric and J. Edward Russo (1984), "Product Familiarity & Learning new Information," Journal of Consumer Research, 11 (June), 542-550.

Katona, George and Mueller, Eva (1955), "A Study of Purchase Decisions," in Consumer Behavior: The Dynamics of Consumer Reaction, ed. Lincoln E. Clark, new York : new York University Press, 30-87.

Kiel, G.C. and Layton, B.A. (1981), "Dimensions of Consumer Information Seeking Behavior," Journal of Marketing Research, XVIII (May), 233-239.

May, F.E. (1969), "Adaptive Behavior in Automobile Brand Choices," Journal of Marketing Research, VI (February), 62-65

Meyer, R.J. (1982), "A Descriptive Model of Consumer Information Search Behavior," Marketing Science, I (Winter), 93-121.

Miller, G.A. (1956), "The Magical Number Seven, Plus or Minus Two" Some Limits on Our Capacity for Processing Information," The Psychological Review, 63 (March), 81-97.

Moore, William L. and Donald R. Lehmann (1980), "Individual Differences in Search Behavior for a Nondurable," Journal of Consumer Research, 7 (December), 296-307.

Newman, W. Joseph (1977), "Consumer External Search: Amount and Determinants," in Consumer and Industrial Buying Behavior (ed.) Jagdish N. Sheth and Peter D. Bennett. new York: Elseiver North-Holland Inc., 79-94.

Newman, W. and Staelin, R. (1971), "Multivariate Analysis of Differences in Buyer Decision Time," Journal of Marketing Research, VIII (Hay), 192-198.

Newman, W., and Staelin, R. (1972) "Prepurchase Information Seeking for new Cars and Major Household Appliances," Journal of Marketing Research, IX (August), 269-257.

Punj, N. Girish and Staelin, R. (1983), "A Model of Consumer Information Search Behavior for new Automobiles," Journal of Consumer Research, 9 (March), 366-380.

Ratchford, B.T. (1980), "The Value of Information for Selected Appliances," Journal of Marketing Research, XVII (February), 14-25.

Sternthal, Brian and Craig, C. Samuel (1982), "Consumer Behavior - An Information Processing Perspective", new Jersey: Prentice-Hall Inc.

Stigler, G.J. (1961), "The Economics of Information," The Journal of Political Economy, XIX (June), 213-225.

Swan, J.E. (1969), "Experimental Analysis of Predecision Information Seeking", Journal of Marketing Research, VI (Hay), 192-197.

Udell, J.G. (1966), "Prepurchase Behavior of Buyers of Small Electrical Appliances," Journal of Marketing, 30 (October), 50-52.

Westbrook, R A and Fornell, C. (1979), "Patterns of Information Source Usage Among Durable Goods Buyers", Journal of Marketing Research. XVI (August), 303-312.

Wright, Peter (1975), "Consumer Choice Strategies: Simplifying vs. Optimizing," Journal of Marketing Research, XII (February), 60-67.