Direct and Indirect Effects of Confidence on Purchase Intention

ABSTRACT - This study investigates two very different processes through which confidence toward a brand could affect intention to buy the same brand. The direct process describes the direct effect of confidence on intention. The indirect process represents the influence of confidence on brand attitude, which in turn affects intention. The empirical results from a structural equation analysis show that intention to buy a specific brand is both directly and indirectly affected by confidence toward that brand.

Citation:

Michel Laroche, Chankon Kim, and Lianxi Zhou (1995) ,"Direct and Indirect Effects of Confidence on Purchase Intention", in NA - Advances in Consumer Research Volume 22, eds. Frank R. Kardes and Mita Sujan, Provo, UT : Association for Consumer Research, Pages: 333-339.

Advances in Consumer Research Volume 22, 1995      Pages 333-339

DIRECT AND INDIRECT EFFECTS OF CONFIDENCE ON PURCHASE INTENTION

Michel Laroche, Concordia University

Chankon Kim, Concordia University

Lianxi Zhou, Concordia University

ABSTRACT -

This study investigates two very different processes through which confidence toward a brand could affect intention to buy the same brand. The direct process describes the direct effect of confidence on intention. The indirect process represents the influence of confidence on brand attitude, which in turn affects intention. The empirical results from a structural equation analysis show that intention to buy a specific brand is both directly and indirectly affected by confidence toward that brand.

INTRODUCTION

The confidence construct was first proposed by Howard and Sheth (1969) as one of the determinants of purchase intentions. They postulate that confidence is positively related to intention. Similarly, Bennett and Harrell (1975) suggest that confidence plays a major role in predicting brand attitudes and intentions to buy. Despite some evidence of a positive relationship between confidence and intention (e.g., Bennett and Harrell 1975; Howard 1977; Laroche and Sadokierski 1994), the fundamental question as to how confidence affects intention has not been satisfactorily resolved. This is an issue of process which warrants more theoretical and empirical inquiry.

The existing literature provides two alternative explanations. In the first, confidence in brand evaluation is considered as one of the determinants of intention to buy. Thus, a direct relationship between confidence and intention is hypothesized. The second explanation posits that the influence of confidence on intention is through its influence on attitude toward the brand. In this view, an indirect relationship between confidence and intention is hypothesized, with attitude as an intermediate link.

To date, it is still not clear which explanation is more valid. One may also suspect a combination of direct and indirect processes. Therefore, the central question of interest in the present study is: Is the influence of confidence on purchase intention explained by a single process (i.e., a direct or an indirect effect) or by a multiple process (i.e., a combination of direct and indirect effects)? This question is examined in this paper under a multiple brand context, as it will be explained later.

The Confidence Construct

According to Howard (1989), confidence is the buyer's subjective certainty - his/her state of feeling sure - in making a judgment on the quality of a particular brand. In other words, confidence is the degree of certainty that one's evaluative judgment of the brand is correct (Howard 1989, p. 34). This definition has two theoretically different meanings. It may refer to the buyer's overall confidence in the brands, or alternatively, to the buyer's confidence in his ability to judge or evaluate attributes of the brands (Bennett and Harrell 1975). The latter meaning of the confidence construct is largely associated with the perception of risk underlying decision making. As Howard and Sheth (1969) postulated, confidence is the buyer's belief that s/he can estimate the payoff of purchasing a particular brand. In fact, Ostlund (1973) operationalized confidence as the degree of certainty regarding product performance risk and psychological risk. This postulation is also related to the term 'specific self-confidence' defined by Cox and Bauer (1964) as "confidence in performing a specific task or in solving a specific problem."

In the psychological literature, the overall confidence in brand evaluations is often considered as one of the dimensions of the attitude construct. For example, Sample and Warland (1973) indicate that the certainty (or confidence) with which a person makes attitudinal judgments reflects the extent to which the person has actually formed an attitude toward a focal object. Fazio and Zanna (1981) further propose that confidence with which one's attitude is held represents one of the four qualitative dimensions of attitude - clarity, confidence, reliability, and accessibility. Moreover, based on a literature review, Raden (1985) summarizes several strength-related attitude properties in which the certainty or confidence toward an attitude object is also included.

However, Howard and Sheth (1969) and Howard (1989) do not agree with such a conceptualization. They argue that confidence is not a dimension of the attitude construct. Indeed, Day (1970) indicates that the confidence level may reflect uncertainty about the correctness of the brand judgment, or ambiguity as to the meaning of the attitude object. Accordingly, previous experience and familiarity with a particular brand are some of the determinants of one's confidence about that brand. Bennett and Harrell (1975) further note that the multiattribute attitude doesn't fully reflect the property of confidence. In fact, an individual's attitude toward an object, according to the multiattribute attitude model, is determined by the multiplicative summation of his/her subjective probabilities (i.e., beliefs) that the object possesses the considered attributes and his/her evaluations of these attributes. For the traditional attitude model, the certainties or probabilities apply to beliefs about the attitude object rather than to the evaluative attitude per se (Raden 1985).

Notwithstanding the probabilistic nature of beliefs in forming attitude, the overall certainty about the object has not been integrated into the attitude model (Bennett and Harrell 1975). Consequently, one tends to accept that overall confidence toward a brand is separate from the attitude toward the same brand.

This research has been designed to examine both these direct and indirect effects of confidence in brand judgments on purchase intention.

The Two Alternative Ways by which Confidence Affects Intention

The direct positive relationship between confidence and intention to buy was first proposed by Howard and Sheth (1969). Bennett and Harrell (1975) provide empirical evidence to support this argument. In a more recent study, Laroche and Sadokierski (1994) further show that intention to select an investment firm is a direct function of confidence in one's evaluations on the firm. They report that the effect of confidence on intention is much stronger than that of attitude toward the same brand. They also found that, in 14 out of 18 cases, attitude toward brand i is significantly related to intention toward the same brand in addition to the effect of confidence.

On the other hand, an argument may be made that confidence has a direct effect on attitude toward the brand, which in turn influences intention to buy, i.e., intention may be an indirect function of confidence. There is some evidence to support this argument. For example, Warland and Sample (1973) show that attitude toward student government is a better predictor of self-reported voting behavior in student elections for individuals with high certainty in their evaluative responses. Fazio and Zanna (1978a; 1978b) obtained similar results. Their studies demonstrate that certainty in judgments about the attitude object significantly enhances the prediction of behavior from attitude toward the object.

Given the above findings, it appears that the overall confidence toward a target brand can influence behavioral intention in two ways: it can have a direct effect on intention or it can indirectly affect intention through its effect on attitude toward the brand. Which explanation makes more sense, however, remains largely unclear in the existing literature. Furthermore, there exists yet another possibility; confidence may affect intention simultaneously in both direct and indirect ways.

Intuitively, the direct effect may be explained as follows: individuals who have higher certainty about their brand evaluations will be closer to some motivational equilibrium with respect to additional information search (Howard 1989). When individuals approach this level of equilibrium, they will need less product information, curtail their information search and thus, be more likely to be ready to act.

An explanation for the indirect effect, on the other hand, may come from the perspective of the consumer information search behavior. When consumers are not quite sure about their subjective brand judgments, they tend to search for more product-related information so as to reduce uncertainty (Urbany et al. 1989). Given limited information, consumers with higher levels of uncertainty are likely to be frustrated. As a result, the perceived attractiveness of the judged brand tends to be affected as a "compensation" for the uncertainty in brand evaluations, which, in turn, influences behavioral intention.

If both of these explanations are correct, there is finally the possibility that behavioral intention may be affected simultaneously by confidence both directly and indirectly.

The Competitive View of Intention Formation

Consumers are often faced with several brands or alternatives from which they have to choose. Most researchers acknowledge the influence of competing brands in explaining intentions and behavior. In their multi-brand model of intentions, Laroche and Brisoux (1989) formally postulate that consumers' intention to buy a specific brand is determined not only by the attitude toward the same brand but also by the attitudes toward other competing brands in the choice set (i.e., the distribution of attitudes toward all the brands in the choice set).

Similarly, Tyebjee (1979) suggests that an attitude should be considered in both absolute and relative senses. In the prediction of voting behavior, Jaccard (1981) shows that more variance is explained when attitudes toward other alternatives are included in the analysis. This conceptual position is also supported by several other studies (e.g., Malhotra 1986; Nantel 1986).

Using data on consumers' attitudes toward beers in a choice set, Laroche and Brisoux (1989) tested two different types of effects: a direct effect and a competitive effect. The direct effect suggests that intention to buy a brand is determined only by the attitude toward that brand. The competitive effect implies that attitudes toward other competing brands may affect intention toward a focal brand. According to the authors, the competitive effect, if it happens, tends to negatively influence intention to buy that specific brand. The expected direction of competition reflects the similarity effect (Tversky 1972), which suggests that choice alternatives that are similar tend to draw shares from each other and are more competitive/substitutable than dissimilar alternatives (Kamakura and Srivastava 1984). Their findings show that the model specifying the multiple effects (combination of a direct effect and a competitive effect) is superior to the single effect model. More recently, Laroche and Sadokierski (1994) provided further evidence to support the competitive view of intention formation.

Alternative Models and Hypotheses

To incorporate the different views of the effects of confidence on intention into the multi-brand model of intentions (i.e., Laroche and Brisoux 1989), the present paper compared and tested three alternative models (see Figure 1).

The first model represents the direct link between confidence and intention. It is hypothesized that:

H1: Intention to buy a specific brand is directly affected by one's overall confidence toward that brand, independently of the effects of attitude toward the same brand.

The second model postulates the indirect relationship be-tween confidence and intention. The hypothesis is formulated as follows:

H2: Intention to buy a specific brand is indirectly affected by one's overall confidence toward that brand through attitude toward the same brand.

The third model combines both the direct and indirect effects of confidence on intention. It is expected that:

H3: Intention to buy a specific brand is not only directly affected by one's overall confidence toward that brand but also indirectly affected through attitude toward the same brand.

METHODOLOGY

Sample

The data used in this study came from a survey on the selection of cough/cold syrup medications. The four most popular brands (i.e., Dimetapp, Robitussin, Benylin, and Triaminic) were selected to test the alternative models. Many researchers (e.g., Lussier and Olshavsky 1979) have indicated that consumers tend to limit their final comparison to an average 4 to 7 brands since people have a tendency to reduce their information load and the cognitive complexity of the brand choice process. Therefore, a heavy competition among these four selected brands is expected. The respondents who had these four brands in their evoked sets were selected for analysis. Due to listwise deletions on some missing values, the sample sizes were respectively 107, 93, 118 and 95 for the four brands.

The population for this survey consisted in residents of the Ottawa and Hull census metropolitan areas. Area sampling was used due to convenience, and the lack of discriminating characteristics of cough/cold syrup consumers. Streets were randomly selected from a local street directory. An equal number of streets were chosen from each municipality. Streets were partitioned in two and delivery alternated between the mid-point and end or beginning of each street. Efforts were made to contact as many households on the selected streets as possible. The self-administered questionnaire was left with the consenting individuals to be picked up at a later time or mailed to the researcher (a stamped self-addressed envelope was provided). In total, 324 completed responses were obtained.

FIGURE 1

THE THREE ALTERNATIVE MODELS FOR FOCAL BRAND A

The sample's average age is between 30 and 39, the median level of education is junior college, and the median family income is between $60,000 and $69,000. Of those respondents, 28.7% are male and 71.3% are female. Overall, the sample seems to be skewed toward somewhat affluent families.

Measures

Confidence. The overall confidence toward the brand was measured by two 9-point semantic differential scales (Not confident at all/Very confident; Very uncertain/Very certain). The respondents were required to indicate how confident they were about their evaluation of each brand; and the extent to which they were certain about each brand. The reliabilities of the two measures were .84, .89, .84, and .81, respectively, for the concerned brands (i.e. Dimetapp, Robitussin, Benylin, and Triaminic).

Attitude. Four 9-point semantic differential scales (Very unfavourable/Very favourable, Dislike very much/Like very much, A very bad brand/A very good brand; Very unsatisfactory/Very satisfactory) were used in measuring attitudes toward the selected brands. The mean of these four scales was taken as the indicator of the attitude construct. The reliabilities of the four scales were .95, .93, .94, and .94, respectively, for the four brands.

Intention. Intention was measured by asking the respondent to indicate how many times s/he would buy a certain brand in the next ten purchase occasions. The measure was then divided by ten, thereby varying between 0 and 1. This measure has been used in the literature by Juster (1966) and Howard and Ostlund (1973).

ANALYSES AND RESULTS

In the present study, three different models were tested and compared in order to see the potential effects of confidence toward a brand on intention to buy the same brand. Model 1 represents the direct effect of confidence on purchase intention. Model 2 addresses the indirect effect, in which confidence in one's brand judgment directly affects attitude toward that brand which in turn influences purchase intention. Model 3 combines both the direct and indirect effects of confidence.

The structural equation analysis was used to test the alternative models. For each model, one of the four selected brands was taken as the focal brand and the other brands were considered as the competitors. Since Model 3 represents both direct and indirect effects, only the LISREL specifications for this model are given in Figure 2.

As summarized in Table 1, when Dimetapp was considered the focal brand and the other three were the competitors, Model 1 (hypothesizing a direct effect only) yielded a Chi-square value of 25.65 with 14 degrees of freedom, an adjusted goodness-of-fit index (AGFI) of .883, and a root mean square (RMSR) of .118. The probability level of the Chi-square value (p=.029) showed that the fit of the model was poor. Similar results were also obtained when any of the other three brands (Robitussin, Benylin, and Triaminic) was considered as the focal brand. Hence, none of these findings provides support for the direct effect model.

FIGURE 2

A LISREL SPECIFICATION OF THE CONBINED EFFECTS MODEL (MODEL 3)

On the other hand, Model 2 (hypothesizing an indirect effect only), in all four situations involving a different focal brand, yielded a satisfactory goodness-of-fit indicators (i.e., Chi-square value, AGFI, and RMSR). Thus, hypothesis 2 was not rejected, i.e., intention to buy a specific brand was indirectly affected by one's confidence toward that brand through attitude toward the same brand.

The best fit indicators, however, were obtained for Model 3 in all four situations involving a different focal brand. These findings suggest that the influences of confidence on purchase intention are both direct and indirect. The Chi-square difference test (Table 2) showed that Model 3 was significantly better than both Model 1 and Model 2. Consequently, Model 3 was the best among the three alternative models. The squared multiple correlations for structural equations ranged from .12 to .44, and the total coefficient of determination for structural equations ranged from .23 to .39 for the best model (i.e., Model 3) in all four situations involving a different focal brand.

TABLE 1

GOODNESS-OF-FIT INDICATORS FOR THE THREE MODELS

TABLE 2

CHI-SQUARE DIFFERENCE TESTS FOR THE THREE MODELS

Table 3 presents the estimates of the parameters and their corresponding t-values of the model 3. Table 4 provides the same results in a matrix format to a better visualize the results on the various paths. With respect to the path between confidence and intention, a significant positive relationship was found no matter which one of the four brands was used as the focal brand (see g₂₄). Similarly, the confidence-attitude link was also positive and highly significant for any of the four alternatives as the focal brand (see g₁₄). Furthermore, a significant positive relationship between attitude toward the focal brand and intention to buy the same brand was found for each of the four brands (b₂₁). The significant negative influences of other competing brands on intention to buy the focal brand were observed in some cases (i.e., g₂₁ and g₂₃ for Dimetapp as the focal brand; g₂₁, g₂₂ for Robitussin as the focal brand; g₂₁ and g₂₃ for Benylin as the focal brand). The influence of other competitors on intention to buy a focal brand is consistent with the predictions of the multi-brand model of intentions (Laroche and Brisoux 1989), which suggests that intention to purchase the focal brand is negatively affected by the global attitude toward other competing brands.

TABLE 3

STANDARDIZED ESTIMATES FOR THE THREE MODELS (MODEL 3)

TABLE 4

PATH COEFFICIENTS FOR MODEL 3 AND EACH BRAND

In conclusion, of the three hypotheses, Hypothesis 3 was best supported by our data. Some support was shown for Hypothesis 2. Finally, Hypothesis 1 was rejected by the findings.

CONCLUSION AND DISCUSSION

This study has incorporated the confidence construct into the multi-brand model of intentions (Laroche and Brisoux 1989) and tested some alternative ways by which confidence toward a specific brand may affect intention to buy that brand. The findings show that the combined effect model (Model 3) is the best one when compared to the other two models, suggesting that intention to buy a brand is not only directly affected by confidence toward the brand but also indirectly affected through one's attitude toward the same brand. This study also shows that intention to buy a specific brand is a positive (negative) function of attitudes toward the same brand (other competing brands), which is consistent with the earlier findings reported by Laroche and Brisoux (1989) and Laroche and Sadokierski (1994).

Given the above findings, it appears that individuals with different degrees of confidence toward a specific brand tend to have different perceptions and attitudinal judgments of the same brand, and are likely to have different levels of motivational equilibrium. The magnitudes of the parameter estimates indicate that the influence of confidence on attitude (g₁₄) is significantly greater than the influence of confidence on intention (g₂₄₎ for each of the four competing brands.

One major implication of the present study is that, in order to increase a consumer's intention to buy a specific brand, a marketer may try to enhance his/her confidence toward the brand. This can be realized by providing the consumer with more product-related information, or direct experience.

Future research may focus on whether the findings here can be replicated for other product classes. It is believed that the more important the product category is (i.e., the higher the level of involvement), the stronger the effects of confidence on intention will be found. In addition, there is also a need to incorporate cognitions into the proposed model. It is believed that the inclusion of cognition constructs will provide a more complete understanding of the influences of confidence on brand attitudes and purchase intentions.

REFERENCES

Bennett, Peter D. and Gilbert D. Harrell (1975), "The Role of Confidence in Understanding and Predicting Buyers' Attitudes and Purchase Intentions," Journal of Consumer Research, 2 (September), 110-117.

Cox, Donald F. and Raymond A. Bauer (1964), "Self-Confidence and Persuasibility in Women," Public Opinion Quarterly, 28 (Fall), 453-466.

Day, George S. (1970), Buyer Attitudes and Brand Choice Behavior, New York: The Free Press.

Fazio, Russell H. and Mark P. Zanna (1978a), "On the Predictive Validity of Attitudes: The Roles of Direct Experience and Confidence," Journal of Personality, 46, 228-243.

Fazio, Russell H. and Mark P. Zanna (1978b), "Attitudinal Qualities Relating to The Strength of The Attitude-Behavior Relationship," Journal of Experimental Social Psychology, 14, 398-408.

Fazio, Russell H. and Mark P. Zanna (1981), "Direct Experience and Attitude-Behavior Consistency," in Advances in Experimental Social Psychology, 14, ed. L. Berkowitz, New York: Academic Press, 161-202.

Howard, John A. (1973), "Confidence as a Validated Construct," in Buyer Behavior: Theoretical and Empirical Foundations, eds. John A. Howard and Lyman E. Ostlund, New York: Alfred A. Knopf, 426-433.

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

Howard, John A. (1989), Consumer Behavior in Marketing Strategy, Englewood Cliffs, NJ: Prentice Hall.

Howard, John A. and Lyman E. Ostlund (1973), "The Model: Current Status of Buying Behavior Theory," in Buyer Behavior: Theoretical and Empirical Foundations, eds. John A. Howard and Lyman E. Ostlund, New York: Alfred A. Knopf, 3-32.

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

Jaccard, J. (1981), "Attitudes and Behavior: Implications of Attitudes Toward Behavioral Alternatives," Journal of Experimental Social Psychology, 17, 206-307.

Juster, F. T. (1966), "Consumer Buying Intentions and Purchase Probability: An Experiment in Survey Design," Journal of the American Statistical Association, 61, 659-696.

Kamakura, Wagner A. and Rajendra K. Srivastava (1984), "Predicting Choice Shares Under Conditions of Brand Interdependence," Journal of Marketing Research, 21 (November), 420-434.

Laroche, Michel and Jacques E. Brisoux (1989), "Incorporating Competition Into Consumer Behavior Models: The Case of the Attitude-Intention Relationship," Journal of Economic Psychology, 10 (September), 343-362.

Laroche, Michel and Robert Sadokierski (1994), "Role of Confidence in a Multi-Brand Model of Intentions for a High Involvement Service," Journal of Business Research, 29 (January), 1-12.

Lussier, Denis A. and Richard W. Olshavsky (1979), "Task Complexity and Contingent Processing in Brand Choice," Journal of Consumer Research, 6 (September), 154-165.

Malhotra, Naresh K. (1986), "An Approach to the Measurement of Consumer Preferences Using Limited Information," Journal of Marketing Research, 23 (February), 33-40.

Nantel, Jacques (1986), "Attitude-Behavior Consistency: Some Considerations Specific to Marketing Research," in Marketing, Vol. 7, ed. Thomas E. Muller, Montreal: Administrative Sciences Association of Canada, 271-279.

Ostlund, Lyman E. (1973), "Product Specific Self-Confidence Related to Buying Intentions," in Buyer Behavior: Theoretical and Empirical Foundations, eds. John A. Howard and Lyman E. Ostlund, New York: Alfred A. Knopf, 434-442.

Raden, D. (1985), "Strength-Related Attitude Dimensions," Social Psychology Quarterly, 48 (4), 312-330.

Sample, J. and R. H. Warland (1973), "Attitudes and The Prediction of Behavior," Social Forces, 51, 292-304.

Tversky, Amos (1972), "Elimination by Aspects: A Theory of Choice," Psychological Review, 79, 281-299.

Tyebjee, Tyzoon T. (1979), "Response Time, Conflict and Involvement in Brand Choice," Journal of Consumer Research, 6 (December), 295-303.

Urbany, Joel E., Peter R. Dickson, and William L. Wilkie (1989), "Buyer Uncertainty and Information Search," Journal of Consumer Research, 16 (September), 208-215.

Warland, R. H. and J. Sample (1973), "Response Certainty as a Moderator Variable in Attitude Measurement," Rural Sociology, 38, 174-186.

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