The Cognitive Representation of Services Varying in Concreteness and Specificity

Laurette Dube-Rioux, University of Montreal
Dennis T. Regan, Cornell University
Bernd H. Schmitt, Cornell University
[ to cite ]:
Laurette Dube-Rioux, Dennis T. Regan, and Bernd H. Schmitt (1990) ,"The Cognitive Representation of Services Varying in Concreteness and Specificity", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 861-865.

Advances in Consumer Research Volume 17, 1990      Pages 861-865

THE COGNITIVE REPRESENTATION OF SERVICES VARYING IN CONCRETENESS AND SPECIFICITY

Laurette Dube-Rioux, University of Montreal

Dennis T. Regan, Cornell University

Bernd H. Schmitt, Cornell University

Researchers and practitioners have suggested that the inherent intangibility of services should be taken into account when designing marketing communications for services, but there is little agreement as to how this should be done. A reconciliation of the issue requires more detailed knowledge about how consumers cognitively represent services. We conceptually and empirically divide the concept of "intangibility" into the separate, but related, dimensions of concreteness and specificity. We then demonstrate that services differing on the former (but not the latter) dimension reliably call to mind salient attributes differing on both dimensions of concreteness and specificity. The results are a first step in understanding how different levels of concreteness and specificity should be used in marketing communications for services.

The elements or attributes composing services are typically more abstract and intangible than those of consumer goods (Fuchs, 1985). In fact, services have been described as "doubly intangible": not only do most services lack material substrates, but they are also difficult to grasp mentally, because consumers may be unsure what exactly they are purchasing when they receive a service (Bateson, 1979). Consequently, researchers and practitioners have suggested that marketing communications for services should take into account the inherent intangibility of services (George and Berry, 1981; Shostack, 1977).

However, researchers do not agree about how this should be done. In fact one can find two radically different views about communication strategies for services. Some researchers have argued that because services are intangible, they may be represented in the consumer's mind relatively abstractly and therefore should be advertised relatively abstractly (Rothschild, 1987). Others have argued that service advertisements should be made concrete and visual in order to overcome the inherent disadvantage of services' intangibility (George and Berry, 1981; Shostack, 1977). A reconciliation of the issue may require more detailed knowledge about how consumers cognitively represent services. Moreover, in generating the above propositions, researchers may have focused too much on contrasting consumer goods with services and may have overlooked the enormous variability that exists among services. Although services are generally more intangible than consumer goods, services differ widely among themselves on the dimension of intangibility. Some services, such as life insurance or college, are relatively "abstract" and typically associated with intangible attributes (e.g., security or knowledge). Other services, such as health services and auto repair, are more "concrete" and invoke sensory experiences. Therefore, rather than proposing one communication approach for all services, different communication strategies may be needed for services that vary in degree of intangibility.

Further complicating matters, the concept of , intangibility itself is not clearly defined when it [ comes to services. For both services and their attributes, the concept of intangibility has been frequently related to the dimension of abstractness/concreteness. But the concept of abstractness/concreteness itself is unclear, and researchers have investigated at least two different notions of it (for a review see Paivio, 1965). One definition -- which from now on we will call concreteness-- relates to accessibility to the senses (Paivio, Yulle, and Madigan, 1968). The other view -- which we will call specificity-- refers to subordination, the specificity of a word being inversely related to the number of subordinate words it embraces (Spreen and Schulz, 1966). Because the two notions of intangibility are perhaps independent (Paivio, 1965; Kammann and Streeter, 1971), an intangible attribute such as a piece of legal advice, may be at the same time both specific and abstract. i Research in marketing, in focusing on cognitive representations of durable and non-durable goods at the level of the specific brand and of the generic product category (e.g., Johnson, 1984; Johnson and Fornell, 1987), has thus far exclusively addressed the specificity view of the abstractness/concreteness concept. Yet the question of a possible effect of different levels of the concreteness of representations defined in terms of accessibility to '; the senses should also be addressed as it may be f equally relevant for both goods and services.

In this research, we study how both dimensions (concreteness and specificity) affect the cognitive representation of services. Johnson and Fornell (1987) have already shown that the specificity of a product representation can affect the level of specificity of attributes that consumers have represented in conjunction with the product. For three superordinate category types (home entertainment, domestic appliance and transportation), they investigated the level of specificity of attributes generated under three levels of product abstraction (superordinate category, product category, and brand). In a free elicitation task, they found a significant difference between the level of specificity of attributes generated at the superordinate level compared to the ones generated at the product type and brand levels, which did not differ significantly. Given Johnson and Fornell's findings that specific product attributes are linked to brands while generic attributes are linked to product categories, two research objectives arise: (i) to investigate whether cognitive representations of services have similar structures, and (ii) to extend Johnson and Fornell's research on the dimension of specificity to include the dimension of concreteness.

METHOD

Pretests

The stimuli for the experiment were selected on the basis of two extensive pretests. The purpose of pretest 1 was to generate a list of attributes for a number of services provided by the experimenter. In a free elicitation task, 42 graduate students in marketing were asked to write the first five attributes that come to mind when they think about the services. Attributes were aggregated across subjects, and the 18 most frequently mentioned for each service were retained for further study. In pretest 2, 18 psychology students rated the concreteness (1=very abstract, 7=very concrete) and specificity (1=very generic, 7=very specific) of 12 services and, in addition, the concreteness (i.e., the degree to which they can be experienced with the senses) and the specificity (i.e., to how many instances they refer) of the list of attributes that subjects had generated for these services in pretest 1. Subjects read the following explanations for the two dimensions:

Concreteness ratings: The degree to which we can experience services with our senses varies. A very concrete service is one whose features or outcomes can be seen, felt, smelled or tasted. A very abstract service is one that cannot be experienced by the senses as such. For example, "cleaning services" are very concrete; "airlines" deliver slightly less concrete services, and the services provided by "Place, Life Insurance Co." are the most abstract, i.e., they cannot be seen, felt, smelled or tasted. Specificity ratings: One may refer to services in a highly specific or in a less specific way. The more specific a service is, the more words used to describe it are restricted to only one or two actual instances of the given type of services. For example, a highly specific word refers to a very particular instance, i.e., it can be applied usually only to one specific member of a class. The less specific a word, the more it becomes applicable to a larger number of services in that class, i.e., it is more generic. You should note that "cleaning services" is presented in a very generic form since there may be all kinds of cleaning services; that airlines is slightly more specific, and "Place, Life Insurance Co." is most specific, referring to a firm selling a particular type of insurance.

The following explanations of the concepts of concreteness and specificity for attributes were provided:

Concreteness ratings: Attributes are very concrete if they represent features or outcomes of a service that can be seen, felt, smelled or tasted. Attributes are very abstract if they refer to features or outcomes not experienced by the senses as such. city ratings: Attributes are very specific if they refer precisely to identifiable features or outcomes of a service. Attributes are very generic if they refer to more general aspects of features or outcomes of a service, or to general aspects of the service itself.

Based on the results of the pretests, eight services were selected as stimuli for the experiment: two abstract/generic services (college education and legal services), two abstract/specific services (Morton College of Arts and Sciences and Thompson & Associates Divorce Specialist), two concrete/generic services (healthcare and auto repair), two concrete/specific services (Max Krauss, Doctor of Dentistry and - fix-it, Muffler / Brake Shop). Thus services were selected according to a 2 (concreteness) X 2 (specificity) factorial design with one replication.

Procedure

Thirty-four graduate students in marketing (14 females and 20 males) participated in the experiment. The average age of the participants was 28.

To confirm that we had selected through the pretests services that indeed varied along the dimensions of concreteness and specificity, subjects who actually participated in the experiment were asked to rate again all eight services in terms of concreteness and specificity. In addition, for four services, subjects were asked to check (from a list of 18 attributes that had been generated in the pretest) those five attributes which first came to mind when they thought about the particular service. The four services were selected on the basis of a replication of a Latin square design, and four different sets of four services were constructed in order to eliminate the possible confounding effects of certain stimulus combinations. Finally, subjects were asked to rate on seven-point scales the complete list of attributes for each of the four services in terms of concreteness and specificity.

RESULTS

Preliminary analysis

Were the concepts of concreteness and specificity of services successfully operationalized as two independent dimensions? First, two 2 (concreteness of the service) x 2 (specificity of the service) ANOVAs conducted on the concreteness and specificity ratings of services, revealed a strong main effect of manipulated concreteness on the concreteness ratings and a strong main effect of manipulated specificity on the specificity ratings of services (both p = < .0001). In addition, there was a significant, but weaker (p < .05) main effect of specificity on the concreteness ratings. More specific services tended to be rated as more concrete. Thus, while concreteness ratings were strongly affected by the concreteness manipulation, and to a lesser extent by the specificity manipulation, specificity ratings were uniquely and strongly affected by the specificity manipulation. Second, the average correlation between the two scales was assessed by back-transformed average Z score; it was modest (r = .26) and marginally significant (.05 < p < .10). Thus, although the two dimensions of concreteness and specificity of services were intercorrelated to some extent, the intercorrelation seemed to be low enough to justify the investigation of both variables separately. Finally, the interrelation between the two dimensions was assessed for the service attributes. The average correlation between the two scales of concreteness and specificity for attributes was .41 (p < .01), a magnitude consistent with previous findings (Spreen and Schulz, 1966; Kammann and Streeter, 1971).

TABLE 1

MEAN RATINGS OF CONCRETENESS AND SPECIFICITY OF MOST SALIENT ATTRIBUTES AS A FUNCTOIN OF SERVICES' CONCRETENESS AND SPECIFICITY

Consistency of cognitive representations

To test the hypothesis that there is consistency between the representation of services and their attributes on both the dimensions of concreteness and specificity, a multivariate analysis of variance (MANOVA) was conducted on the concreteness and specificity ratings of each subject's five salient attributes. The results revealed a main effect of manipulated service concreteness on ratings of both attribute concreteness and specificity (F[2, 70] = 14.50, p < .001). The main effect of specificity of services and the interaction were not significant (means are presented in table 1). Thus when subjects think about abstract services such as education and legal services, the attributes that first come to their mind are both more abstract and generic than when the same subjects think about concrete services, such as health care and auto repair. On the other hand, when they think about generic services, such as healthcare services, the salient attributes are neither more generic nor more abstract than when they think of a specific instance of the service, such as Dr. Max Krauss, Doctor of Dentistry.

Because the list of attributes differed between abstract and concrete services, but not between generic and specific levels of a service, this result could simply reflect the fact that the pool of attributes from which salient attributes were extracted was, as a whole, more abstract and generic for abstract services, and more concrete and specific for concrete ones. In this case, attributes randomly selected from this pool will show the same bias. To rule out this rather plausible alternative explanation, we conducted an idiographic analysis comparing, for each service, the concreteness and specificity ratings that a subject assigned to salient and non-salient attributes. An individual difference score was computed by subtracting, for each subject, the average rating of the five salient attributes from the average rating of the thirteen non-salient attributes. These difference scores indicated whether the two clusters of attributes were rated differently, e.g., as more or less concrete or specific. Nine subjects had provided concreteness and specificity ratings only for the five salient attributes. Their data were excluded from this idiographic analysis.

Average difference scores across all subjects are shown in Table 2 along with significance levels for relevant t tests. The results show that the effect of service concreteness on the ratings of salient attributes is not artifactual and may indeed indicate how subjects cognitively represent services. The idiographic analysis showed that, compared with concrete services, the salient attributes of abstract services were both less concrete and less specific (both p<.05). The idiographic analysis also confirmed the null effect of the specificity manipulation: no significant differences were found between salient and nonsalient attributes of specific and generic services. A closer analysis of this idiographic score shows that in absolute terms, at least for the list of services we presented to our subjects, abstract services call to mind salient attributes that are reliably more abstract than nonsalient ones, whereas concrete services call to mind salient attributes that are reliably more specific than nonsalient ones. Whether these results mean that in general one should highlight abstract attributes when communicating about abstract services, and specific attributes when trying to sell a concrete service, is a question we are currently studying.

TABLE 2

MEAN INDIVIDUAL DIFFERENCES BETWEEN THE MEAN RATINGS OF CONCRETENESS AND SPECIFICITY OF SALIENT AND NON SALIENT ATTRIBUTES AS A FUNCTION OF SERVICES' CONCRETENESS AND SPECIFICITY

DISCUSSION

The results of the experiment indicate that attributes that spontaneously come to consumers' minds when they think about services differ, depending on certain properties of the services. Abstract (as opposed to concrete) services, whether they are generic or specific, bring to mind both more abstract and more generic attributes. When individuals think about college education in general, or about a specific college, knowledge is more likely to come to mind than the comfort provided by a classroom or the location of the campus. The challenge faced by the advertiser of such a service is to preserve the "abstract" essence of this service-the reason why people use it--and then to communicate effectively its characteristics and benefits to the consumer.

Contrary to what we predicted on the basis of Johnson and Fornell's (1987) findings, the level of specificity of services did not affect either the level of specificity or the level of concreteness of salient attributes. Can methodological artifacts explain the failure of the specificity manipulation to affect attribute salience? Perhaps subjects could not possibly be familiar with the artificially constructed names of the service providers. Therefore, the name may not have provided any diagnostic information so that subjects may have relied solely on the category attributes. A more psychologically interesting cause of the null effect of the specificity manipulation may be intrinsically linked to the "doubly intangible" nature of services. As we said previously, most services -- and the most abstract ones in particular -- lack material substrates and are difficult to grasp mentally. As a result, consumers may be unable to perceive and to describe services in highly specific terms, even when they consider individual brands or service providers. The idiographic comparison of the concreteness and the specificity ratings of salient and nonsalient attributes seems to converge in that direction: for abstract services, salient attributes are significantly more abstract -- and not more generic-- than nonsalient attributes whereas for concrete services, salient attributes are significantly more specific but not more concrete than non salient ones. Thus, consumers may cognitively represent concrete services in a way very similar to physical goods. What seems to matter for concrete services and physical goods is the specificity of the attributes rather than their degree of concreteness. On the contrary, more abstract services are cognitively represented by abstract, but not necessarily less specific attributes.

The ultimate goal of our research is to understand how different levels of concreteness and specificity should be used in marketing communications for services, and we believe that an important first step is clarifying how services and their attributes are cognitively represented. The present study has concentrated on mental representations of services and has shown that abstract and concrete services are cognitively represented differently. One possible strategy may be to strive for a match between the specificity and concreteness of the mental representations of services and the content of marketing communications. A test of this proposition requires comparison of the effect of a match and of a mismatch between specificity and concreteness of services and attributes or benefits featured in their advertisement (or other communications), using measures of persuasiveness, recall, and recognition as dependent variables. A study on marketing communications for services that vary along the abstractness/concreteness and genericity/specificity continuum is currently in progress, using a variety of judgement and memory tasks.

REFERENCES

Bateson, John E. G. (1979), "Why we Need Services Marketing?," in Conceptual and Theoretical Development in Marketing, O. C. Ferrel, Stephen W. Brown, and Charles W. Lamb Jr (eds), Proceeding Series, Chicago: American Marketing Association, 131 - 146.

Fuchs, Victor R. (1985), "An Agenda for Research on the Service Sector", in Managing the Service Economy Prospects and Problems, Robert P. Inman (ed.), Cambridge: Cambridge University Press, 319-326.

George, William R., and Leonard L. Berry (1981), "Guidelines for the Advertising of Services", Business Horizons, July/August, 52-56.

Johnson, Michael D. (1984), "Consumer Choice Strategies for Comparing Noncomparable Alternatives", Journal of Consumer Research, 11 (December), 741-753.

Johnson, Michael D., and Claes Fornell (1987), 'The Nature and Methodological Implications of the Cognitive Representation of Products", Journal of Consumer Research, 14 (September), 214-228.

Kammann, Richard, and Lynn Streeter (1971), 'Two meanings of word abstractness", Journal of Verbal Learning and Verbal Behavior, 10, 303-306.

Paivio, Allan (1965), "Abstractness, Imagery, and Meaningfullness in Paired-Associate Learning", Journal of Verbal Learning and Verbal Behavior, 4, 32-38.

Paivio, Allan, John C. Yuille, and Stephen A. Madigan (1968), "Concreteness, Imagery, and Meaningfulness Values for 925 nouns", Journal of Experimental Psychology: Monograph Supplement, 76:1 Part 2.

Rothschild, Michael L. (1987), "Communications for Services Marketing", in Marketing Communications: From fundamentals to Strategies, Lexington: D. C. Heath, 699-726.

Shostack, G. Lynn (1977), "Breaking Free from Product Marketing", Journal of Marketing, 41 (April), 73-80.

Spreen, Otfried, and Rudolph W. Schulz (1966), "Parameters of Abstraction, Meaningfulness, and Pronunciability for 329 nouns", Journal of Verbal Learning and Verbal Behavior, 5, 459478.

Yuille, John C., and Allan Paivio (1968), "Imagery and Verbal Mediation Instructions in Paired Associate Learning", Journal of Experimental Psychology, 78:3, 436-441.

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