Role of Affect and Need For Interaction in On-Site Service Encounters

Pratibha A. Dabholkar, University of Tennessee
ABSTRACT - With the growing availability of computerized service delivery systems in on-site service encounters, customers are increasingly able to perform services for themselves. The issues for practitioners are whether customers will view such options favorably and what would determine their attitudes. A causal model drawing on information processing research and services literature is developed to address these issues. The model is tested and the data are analyzed using LISREL 7. A moderating effect of prior behavior is also hypothesized and is tested using a nested MANOVA and regression analysis. The findings support the hypothesized effects, contributing to theory development and providing implications for practitioners.
[ to cite ]:
Pratibha A. Dabholkar (1992) ,"Role of Affect and Need For Interaction in On-Site Service Encounters", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 563-569.

Advances in Consumer Research Volume 19, 1992      Pages 563-569

ROLE OF AFFECT AND NEED FOR INTERACTION IN ON-SITE SERVICE ENCOUNTERS

Pratibha A. Dabholkar, University of Tennessee

ABSTRACT -

With the growing availability of computerized service delivery systems in on-site service encounters, customers are increasingly able to perform services for themselves. The issues for practitioners are whether customers will view such options favorably and what would determine their attitudes. A causal model drawing on information processing research and services literature is developed to address these issues. The model is tested and the data are analyzed using LISREL 7. A moderating effect of prior behavior is also hypothesized and is tested using a nested MANOVA and regression analysis. The findings support the hypothesized effects, contributing to theory development and providing implications for practitioners.

INTRODUCTION

In recent years there has been much innovation in service delivery and many service firms today offer customers the option to perform a service for themselves with the use of a machine. These options include services customers can perform for themselves from their homes, using a computer and/or telephone (e.g., home shopping or banking), as well as services they can perform for themselves on site, using a computerized service delivery system.

An early example of an on-site computerized service delivery system is the automated teller machine. Recent applications include computers in stores which customers can use to place their orders directly, ticketing machines at airports, closed-circuit check-out in hotel rooms, and touch screens for ordering in fast food restaurants.

A major issue to marketers who consider installing such service designs is whether or not customers would view such options favorably. Clearly, if they did view them favorably, service firms could experience substantial savings in labor costs (Business Week 1990; Lovelock and Young 1979).

Another issue is whether customers' familiarity with using similar systems has a bearing on how they view these computerized service delivery options (The New York Times 1989). If so, familiarity with computerized products could be used as a base for segmentation in providing such self-service options to specific targets.

A third issue relates to whether customers would be concerned about de-humanizing the service situation (Breakwell et al. 1986; Zeithaml and Gilly 1987). If this is a major concern, then marketers would be well advised to provide full service options alongside the new, computerized, self-service options.

This study investigates these practical issues in a fast food setting where touch screens are currently being tested as direct ordering options for customers (Marketing News 1990; Nation's Restaurant News 1991). A model is developed by drawing on information processing research and services literature and is tested with a student sample, a particularly relevant sub-segment for fast food restaurants (The Wall Street Journal 1990). The data are analyzed using confirmatory factor analyses, nested MANOVA, and regression, and the results support the hypothesized model. The theoretical contribution of the study is discussed, as are implications for practitioners and for future research.

CONCEPTUAL FRAMEWORK

When faced with a new situation, people tend to draw on similar past experiences to make judgments about the new situation. This commonly observed phenomenon can be explained using the information processing paradigm (Johnson and Puto 1987; Bettman and Sujan 1987a), and the concept of category-based affect (Fiske 1982; Fiske and Pavelchak 1986; Sujan 1985).

Based on the information processing paradigm, individuals make and store cognitive representations (Lachman, Lachman, and Butterfield 1979; Johnson and Fornell 1987) about different situations, and also make judgments or overall cognitive evaluations about situations they encounter (Einhorn and Hogarth 1981). On confronting a new situation, they may recall earlier cognitive representations or judgments to evaluate the new situation. This is similar to stimulus generalization, an extension of Pavlov's classical conditioning theory. When an individual perceives a new stimulus, it is compared to similar stimuli in memory and is responded to on that basis.

Fiske (1982) and Sujan (1985) explain that prior behavior is associated with category-based affect. They propose that past behaviors are stored as an affective response and when a stimulus matches expectations, it triggers this category-based affect. The category-based affect is extended to the current behavior without a review of cognitive beliefs about the consequences of the current behavior (Bettman and Sujan 1987b). When the prior behavior is exactly the same as the current one, it can influence intentions or behavior directly; when the behavior is similar only in some respects to the current one, the affective response is likely to be triggered (Fiske 1982; Fiske and Pavelchak 1986; Sujan 1985). The new behavior will be viewed favorably or unfavorably depending on the nature of the category-based affect.

Prior behavior may be defined as the extent to which the individual has engaged in that behavior. For computerized service delivery systems, relevant prior behavior would include the use of products based on similar technologies (computers, ATMs, VCRs, etc.). People are more likely to have engaged in these activities than in using touch screens to order fast food, a relatively new option, with limited availability. Since these activities are similar to, but not the same as, using a touch screen to order fast food, prior behavior related to these activities is expected to influence some type of general category-based affect, such as attitudes towards using computerized products in general. The link between prior behavior and generalized attitude has been empirically supported as well (Dickerson and Gentry 1983). It is expected that a pre-conceived, positive, category-based attitude, in turn, would cause the individual to view a new, but similar, situation in a favorable way. Hence, the following hypotheses are proposed:

FIGURE 1

DETERMINANTS OF ATTITUDE TOWARDS NEW COMPUTERIZED SELF-SERVICE OPTIONS

H1: Prior behavior with respect to using computerized products will have a positive effect on attitudes towards using computerized products in general.

H2: Attitudes towards using computerized products in general will have a positive effect on attitudes towards using a new computerized self-service option.

It would be interesting to consider whether a category-based affect exists towards self-service in general that could also affect attitudes towards using a new computerized self-service option. Although there is some evidence that people have different propensities to participate in service delivery (Langeard et al. 1981), it appears that this is largely situation-related (Lovelock and Young 1979). People may have a positive attitude towards certain forms of self-service such as filling your own gas or banking, but a negative attitude towards self-service in restaurants or home repair.

Researchers in services marketing (Bateson 1985; Langeard et al. 1981; Lovelock and Young 1979; Silpakit and Fisk 1985) have suggested that determinants of customer participation in self-service should be investigated in future research. A construct relevant to self-service versus full-service is the need for interaction. This refers to the need that some individuals feel for interacting with the service employee in a service encounter. Langeard et al. (1981) and Bateson (1985) found that the need for human contact in a service encounter was very important to some customers. Some people feel strongly that the use of machines in a service encounter dehumanizes the interaction (Breakwell et al. 1986). It is possible that these people may have a positive attitude towards using computers in their homes, but may dislike the idea of using them in a service situation. Hence, it is proposed that:

H3: The need for interaction with a service employee will have a negative effect on attitudes towards using a new computerized self-service option.

These three hypotheses H1-H3 form a causal model indicating the antecedents of attitude towards new computerized self-service options. The model is presented in Figure 1 below.

Attitudinal researchers (Ajzen, Timko, and White 1982; Oliver and Bearden 1985) recommend that in addition to developing attitudinal models, it would be worthwhile to explore moderating effects on linkages within these models. Researchers in information processing (Johnson 1989; Park, Iyer, and Smith 1989) also encourage investigation of moderating effects on relationships in information processing models.

On examining the variables in this study for possible moderating effects, it is noted that prior behavior would have a moderating effect on the relationship between the need for interaction and attitude towards the new option. This is based on the observation that greater familiarity with computerized products appears to make people comfortable with using them in any situation. Consequently, for these people, the need for interaction would become less important as a determinant of attitude towards new computerized self-service options (see Figure 2). Thus, the following hypothesis is proposed:

H4: Prior use of computerized products will have an attenuating effect on the relationship between need for interaction and attitude towards using a new computerized self-service option.

The literature does not suggest any gender differences in attitudes towards new computerized self-service options. However, two empirical studies did find gender differences that may be relevant for this study. Langeard et al. 1981 found that males showed a stronger preference for self-service options than did females. The authors linked this with the finding that males also tended to be more impatient and disliked waiting more. Breakwell et al. 1987 found that males had more positive attitudes towards new technology in general and believed technology would have more benefits for society. Based on these studies, two exploratory hypotheses are proposed:

FIGURE 2

MODERATING EFFECT OF PRIOR USE OF COMPUTERIZED PRODUCTS

H5a: Males will have more positive attitudes than will females towards using computerized products in general.

H5b: Males will have more positive attitudes than will females towards using a new computerized self-service option.

METHODOLOGY

Research Instrument

The research instrument is based on the scenario and questionnaire approach. The validity of scenarios and the equality of results between laboratory experiments and role playing studies have been well documented (Bem 1967) and this method has been advocated and applied by several researchers in consumer behavior (Bateson 1985; Belk 1974; Surprenant and Solomon 1987). The touch screen ordering option is clearly described in the scenario. It is also explained that price and menu are unchanged for this option. This is then followed by a questionnaire to measure the relevant variables.

The endogenous variables are attitudes towards using computerized products in general and attitudes towards using a new, computerized, self-service option, specifically using a touch screen to order fast food. Four, seven-point, semantic differential items, using endpoints such as good-bad and pleasant-unpleasant, are developed for each construct, consistent with the guidelines suggested by Fishbein and Ajzen (1975) for measuring attitudes.

The exogenous variables are prior behavior with respect to using computerized products and the need for interaction with a service employee. Four, seven-point, frequency items are developed for measuring prior behavior with respect to using computerized products (e.g., home computers, automated teller machines). A four-item, seven-point, Likert scale for measuring the need for interaction with service employees is developed, after testing several versions for ease of understanding and face validity. Examples of items from this scale are "human contact in providing services makes the process enjoyable" and "I like interacting with the person who provides the service."

Analysis

The unidimensionality and reliability of constructs developed in this study are tested with confirmatory factor analysis, considered superior to conventional measure validation techniques (James, Mulaik, and Brett 1987). The analytical tool is LISREL 7 (Joreskog and Sorbom 1989). The model itself (comprising hypotheses H1, H2, and H3) is tested with a confirmatory design (using LISREL 7). Hypotheses H4, H5a, and H5b are tested using conventional multivariate analyses because the theoretical support for these hypotheses is not as strong as for the model, a condition necessary for meaningful application of confirmatory analysis and LISREL. Hypothesis H4 is tested using a combination of a nested MANOVA and regression analysis, while hypotheses H5a and H5b are tested using t-tests.

Sample

The respondents are students from a large urban university. Using a student sample increases homogeneity for theory development (Calder, Phillips, and Tybout 1981) and thus minimizes the problem of omitted variables and unforeseen interactions. In addition, college students are frequent consumers of fast food (The Wall Street Journal 1990) and hence this is a relevant sub-segment of the population of customers of fast food restaurants.

A total of 141 undergraduate students responded to the questionnaire. Sample size is linked with the method of analysis, and it may be noted that successful applications of LISREL have been conducted with about 100 respondents (Davis, Bagozzi, and Warshaw 1989, 1991) and an experimental design using LISREL was successfully executed with 60-70 respondents in each cell (Bagozzi and Moore 1991). The sample consisted of 72 males (age range of 21-45, mean age 25.79) and 69 females (age range of 21-44, mean age 25.65).

TABLE 1

CONFIRMATORY FACTOR ANALYSIS AND CONSTRUCT RELIABILITY N=141

TABLE 2

CAUSAL ANALYSIS RESULTS N=141

RESULTS

Study 1: Construct Development

The scales for the constructs in the study were developed as described and tested for face validity by three expert judges (faculty). Confirmatory factor analysis was conducted to check unidimensionality and compute construct reliability (see Table 1). It is seen from the chi-square values, probabilities, and goodness-of-fit indicators, that two of the scales (ATC and NFI) have excellent fits and one more scale (ATS) has an acceptable fit. The fourth scale (PRIOR) does not have as good a fit, but this is understandable because the scale measured frequencies of several behaviors (e.g., using a home computer, using a VCR, etc.) that need not all be high or low for a given individual. Construct reliabilities were computed from the results of the confirmatory analysis and it is seen that reliabilities for all four scales are much higher than the acceptable 0.5-0.6 for theory development (Nunnally 1978).

Study 2: Causal Analysis

The causal model (see Figure 1) was tested with LISREL 7 (Joreskog and Sorbom 1989). The results of the causal analysis based on covariances are presented in Table 2. It is seen from the Table that the model has a good fit in terms of all the indicators. Furthermore, an examination of the t-values shows that all three hypotheses tested are supported. H1 is supported at p <.01, while H2 and H3 are supported at p <.001.

Thus, as hypothesized, prior behavior with respect to using computerized products has a positive effect on attitudes towards using computerized products in general, which in turn has a positive effect on attitudes towards using new computerized self-service options. Also, as proposed, the need to interact with service employees has a negative effect on attitudes towards using new computerized self-service options.

TABLE 3A

NESTED MANOVA RESULTS FOR HYPOTHESIS H4 N=141

TABLE 3B

REGRESSION ANALYSIS RESULTS FOR HYPOTHESIS H4

Study 3: Moderating Effect

The moderating effect (see Figure 2) was tested using a nested MANOVA and regression analysis. A nested MANOVA compares the slopes of the regression equations for the groups under consideration, and it is seen from Table 3A that a moderating effect was present (F=18.41, p <.001). Thus, the effect of the need for interaction with a service employee, on attitudes towards using a new computerized self-service option, was found to be significantly different for low and high prior use of computerized products.

To see if the moderating effect was in the predicted direction, regression analysis was conducted for low and high prior use groups (see Table 3B). Based on a comparison of the parameter estimates, F-values, and adjusted R2 values, it is seen that hypothesis H4 is supported. Thus, as predicted, prior use of computerized products has an attenuating effect on the relationship between need for interaction with a service employee and attitude towards using a new, computerized, self-service option in on-site service encounters.

Study 4:

Gender differences (H5a and H5b) were tested using two t-tests. The results are presented in Table 4 and do not support either hypothesis. There was no difference between men and women in their attitudes towards using computerized products in general (t = -0.53, n.s.). Moreover, women had a more favorable attitude than did men (t= -2.91, p <.005) towards using new, computerized, self-service options.

DISCUSSION

Both information processing literature and attitudinal research have concentrated on product situations and largely ignored service situations. This study, drawing on the information processing paradigm, uses a service setting to investigate determinants of customer attitudes towards new options in service delivery. Thus, the model represents an extension of both streams of research to service settings.

The model is supported and validates the effect of prior behavior and category-based affect in service settings. The study also addresses issues of concern to services marketing researchers. It is found that attitudes towards new computerized self-service options are adversely affected by the need for interaction with a service employee, but that this effect is mitigated by greater familiarity with computerized products.

TABLE 4

T-TESTS FOR GENDER DIFFERENCES IN ATTITUDES

The findings that females have more favorable attitudes than do males towards new, computerized, self-service options, and that there are no gender differences in attitudes towards using computerized products in general, appear to contradict findings from the two earlier studies by Langerad et al. 1981 and Breakwell et al. 1987. One explanation for these results may be that the Langeard et al. study, which examined self-service versus full-service, used a sample from the general population, and students may not show gender differences for self-service options. The other explanation may lie in the fact that the Breakwell et al. study, which did use a student sample, examined attitudes towards the effects of new technology on society, rather than attitudes towards using products based on technology. It may be that females had a less favorable attitude in the Breakwell study because they were more sensitive to issues such as unemployment caused by technology. In this study, however, their attitudes towards using technology are measured, and it appears that women students, at least, seem to have more favorable attitudes towards using new service options based on technology, than do men.

The study provides several implications for practitioners as well. Those people (students, professionals, white collar workers), who are familiar with computerized products, are more likely to have favorable attitudes towards using new, computerized, self-service options. Therefore, it would be strategically sound to locate such service delivery options at sites frequented by these populations. In other words, familiarity with computerized products can be used as a basis for segmentation.

Firms may also wish to promote familiarity with computerized products in the general population as a long term strategy, since familiarity will positively affect attitudes towards these new service options. However, given that the need for interaction with a service employee does have a significant negative effect on attitudes, it may be prudent to provide full service options alongside these self-service options in locations frequented by the general public, so that customers can have a choice of service delivery.

Although the use of cross-sectional data implies that the temporal sequence of relationships cannot be guaranteed, cross-sectional data is commonly used in causal analysis, and appears to be quite acceptable if there is theoretical support to suggest the causal paths for the hypotheses.

A student sample was considered appropriate for theory development, due to the relative sample homogeneity, and for the setting used in this study, as students are a relevant sample for fast food restaurants. The results may be generalized to other urban student populations, and to some extent, to populations with similar levels of familiarity with computerized products.

For future research, the study could be replicated with a sample from the general population to explore age and educational differences in attitudes. Older or less educated people, for example, may have a greater need for interaction with the service employee, as well as lower familiarity with computerized products. These differences, in turn, may be reflected in their attitudes towards new, computerized self-service options. Also, using a random sample from the general population would further validate the model developed in this study. Future studies could also test the model with different samples across the country to explore regional and cultural differences.

The results of this study should interest any firm that is interested in providing a computerized self-service delivery option to customers. Examples of such options would include airline ticketing machines at airports, computer ordering in catalog stores, closed-circuit check-out in hotel rooms, and touch screen ordering for photo processing. This study could be a starting point for future research that investigates customer attitudes towards using other innovative service delivery options.

REFERENCES

Ajzen, Icek, Christine Timko and John B. White (1982), "Self-Monitoring and the Attitude-Behavior Relation," Journal of Personality and Social Psychology, 42 (3), 426-35.

Bateson, John E. G. (1985), "Self-Service Consumer: An Exploratory Study," Journal of Retailing, 61 (3), (Fall), 49-76.

Bagozzi, Richard P. and David J. Moore (1991), "The Role of Negative Feelings and Empathy as Mediators of the Effects of Fear Appeals," Working Paper.

Belk, Russell (1974), "An Exploratory Assessment of Situational Effects in Buyer Behavior," Journal of Marketing Research, 11, 156-63.

Bem, D. J. (1967), "Self Perception: An Alternative Interpretation of Cognitive Phenomena," Psychological Review, 74, 184-200.

Bettman, James R. and Mita Sujan (1987a), "Research in Consumer Information Processing," Review of Marketing, 197-235.

Bettman, James R. and Mita Sujan (1987b), "Effects of Framing on Evaluation of Comparable and Noncomparable Alternatives by Expert and Novice Consumers," Journal of Consumer Research, 14 (September).

Breakwell, G. M., C. Fife-Schaw, T. Lee, and J. Spencer (1986), "Attitudes Towards New Technology in Relation to Social Beliefs and Group Memberships: A Preliminary Investigation," Current Psychological Research and Reviews, Spring, 5 (1), 34-47.

Business Week (1990), "Service Pay Is Fueling Inflation," April 23, 26.

Calder, Bobby J, Lynn W. Phillips, and Alice M. Tybout (1981), "Designing Research for Application," Journal of Consumer Research, 8 (September), 197-207.

Davis, Fred D., Richard P. Bagozzi, and Paul R. Warshaw (1989), "User Acceptance of Computer Technology: A Comparison of Two Theoretical Models," Management Science, 35 (8), 982-1003.

Davis, Fred D., Richard P. Bagozzi, and Paul R. Warshaw (1991), " Usefulness Versus Fun as Determinants of Intentions to Use Computers in the Workplace," Working Paper.

Dickerson, Mary D. and James W. Gentry (1983), "Characteristics of Adopters and Non-Adopters of Home Computers," Journal of Consumer Research, 10 (September), 225-35.

Einhorn, H. J. and R. M. Hogarth (1981), "Behavioral Decision Theory: Processes of Judgment and Choice," Annual Review of Psychology, 32, 340-88.

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

Fiske, Susan T. (1982), "Schema-Triggered Affect," in Cognitive Social Psychology, A. H. Hastorf and A. H. Isen, eds., New York: Elsevier, 55-78.

Fiske, Susan T. and Mark A. Pavelchak (1986), "Category-Based Versus Piecemeal-Based Affective Responses: Developments in Schema-Triggered Affect," in The Handbook of Motivation and Cognition: Foundations of Social behavior, R. M. Sorention and E. T. Higgins, eds., New York: Guilford.

James, L. R., S. A. Mulaik, and J. M. Brett (1987), Causal Analysis: Assumptions, Models, and Data. Beverly Hills: SAGE Publications.

Johnson, Michael D. (1989), "The Differential Processing of Product Category and Noncomparable Choice Alternatives," Journal of Consumer Research, 16 (3) (December), 300-09.

Johnson, Michael D. and Claes Fornell (1987), "The Nature and Methodological Implications of the Cognitive Representation of Products," Journal of Consumer Research, 14 (September), 214-28.

Johnson, Michael D. and Christopher P. Puto (1987), "A Review of Consumer Judgment and Choice," Review of Marketing, 236-92.

Joreskog, Karl G. and Dag Sorbom (1989), LISREL 7: A Guide to the Program and Applications, second edition, Chicago, IL: SPSS Inc.

Lachman, R., J. Lachman, and E. C. Butterfield (1979), "Cognitive Psychology and Information Processing: An Introduction," Hillsdale, NJ: Lawrence Erlbaum Associates.

Langeard, E., J. E. G. Bateson, C. H. Lovelock, and P. Eiglier (1981), Marketing of Services: New Insights from Consumers and Managers, Cambridge, MA: Marketing Science Institute, Report No. 81-104.

Lovelock, Christopher H. and Robert F. Young (1979), "Look to Consumers to Increase Productivity," Harvard Business Review, (57), (May-June), 168-178.

Marketing News (1990), "Arby's Adds Technology to Boost Service," November 12, 1.

Nation's Restaurant News (1991), "Fast Feeders Turn on Video Menu Systems," March 11, 7.

Nunnally, J. C. (1978), Introduction to Psychological Measurement, New York: McGraw-Hill.

Oliver, Richard L. and William O. Bearden (1985), "Crossover Effects in the Theory of Reasoned Action: A Moderating Influence Attempt," Journal of Consumer Research, 12 (December), 324-40.

Park, C. Whan, Easwer S. Iyer, and Daniel C. Smith (1989), "The Effects of Situational Factors on In-Store Grocery Shopping Behavior," Journal of Consumer Research, 15 (4) (March), 422-33.

Silpakit, Patriza and Raymond P. Fisk (1985), "`Participatizing' the Service Encounter: A Theoretical Framework," in Services Marketing in a Changing Environment, T. M. Bloch, G. D. Upah, and V. A. Zeithaml, eds., Chicago: American Marketing Association, 117-121.

Sujan, Mita (1985), "Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumer Judgments," Journal of Consumer Research, 12 (June), 31-46.

Surprenant, Carol F. and Michael R. Solomon (1987), "Predictability and Personalization in the Service Encounter," Journal of Marketing, 51 (2), 86-96.

The New York Times (1989), "The Powerful Push for Self-Service," (April 9), S3-1.

The Wall Street Journal (1990), "Fast-Food Chains Invade College Eateries," November 27, B1.

Zeithaml, Valarie A. and Mary C. Gilly (1987), "Characteristics Affecting the Acceptance of Retailing Technologies: A Comparison of Elderly and Nonelderly Consumers," Journal of Retailing, 63 (1), 49-68.

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