Development of a Scale For Innovativeness

Clark Leavitt, The Ohio State University
John Walton, The Ohio State University
[ to cite ]:
Clark Leavitt and John Walton (1975) ,"Development of a Scale For Innovativeness", in NA - Advances in Consumer Research Volume 02, eds. Mary Jane Schlinger, Ann Abor, MI : Association for Consumer Research, Pages: 545-554.

Advances in Consumer Research Volume 2, 1975      Pages 545-554

DEVELOPMENT OF A SCALE FOR INNOVATIVENESS

Clark Leavitt, The Ohio State University

John Walton, The Ohio State University

[This research was supported by the Leo Burnett Company, Chicago.]

[Clark Leavitt is Professor of Marketing, The Ohio State University.]

[John Walton is a Ph.D. student at The Ohio State University.]

This paper describes the construction of a scale to measure innovativeness--a psychological--trait underlying adoption of new ideas, services and products. The study utilizes as a model a psychometric methodology that has demonstrable power in personality scale construction in order to demonstrate its potential for use in marketing research.Four steps were gone through in applying the model to innovativeness: (1) defining the trait, (2) creating a large pool of items based on that definition, (3) utilizing techniques for suppression of response bias to select the best items in terms of a validity/reliability criterion, (4) evaluation in terms of convergent and discriminant validity. A fifth criterion, predictive validity, awaits future developments.

This is a report on the first of a series of studies designed to understand the individual psychology underlying the adoption of new ideas and products. One of the determinants of time-of-adoption may be the individual's standing on some broad personality trait. We are concerned here with the construction of an instrument to measure such a trait.

The use of personality variables in marketing has not had a happy history. From the time of Evans' attempts to discriminate Ford and Chevy owners to the present, the record has been marked by failure, or, at best, limited success. Recently +here has been renewed reason for encouragement as a result of attempts to relate patterns of personality traits with patterns of product choice (see Worthing et al., 1973, for reference). The interaction between traits of the person and other variables such as situational measures has also shown promise (Endler and Hunt, 1968).

The basically poor record should come as no surprise. As has been frequently pointed out (Mischel, 1968) by personality psychologists, global traits have very little predictive power in most areas. Beyond this, the choice of measures in marketing seems unnecessarily ad hoc, often involving instruments with poor or unknown technical specifications and with little thought about the nature of the product choice to be predicted.

Yet it seems premature to dismiss the possibility that a single global trait can be worth the candle in marketing, particularly in predicting time-of-adoption behavior, dependent as it is on information utilization of the kind often influenced by personality traits. Therefore, this study was done to ~ind or create an appropriate trait measure.

THE ADOPTION PROCESS

Considerable research during the past decade has begun to highlight and quantify major elements of the diffusion process of new product and other innovations. Everett Rogers (1962), in his now classic work, postulated a diffusion paradigm that was capable of generalization to diverse behavioral disciplines. Robertson (1971), using the Rogers paradigm as a point of departure, established the practical and theoretical efficacy of marketing innovation research.

The diffusion of various types of marketing innovations has been studied with the new line of ladies hats (King, 1964)j instant coffee (Frank, Massy, Morrison, 1964); auto diagnostic centers (Engel, Kegerreis, Blackwell, 1970): a detergent (Pessemier, Burger, Tigert, 1967); and touchtone telephones (Robertson, 1967), among others.

A degree of commonality can be found in the research methodologies of most of these studies: (1) the dependent variable, adoption, is operationalized by the researcher; (2) an expedient assortment of independent variables is specified; (3) correlation or regression analysis is used to measure the degree or strength of the relationship between a variety of independent variables and the dependent variable.

The results from-this paradigm have not been encouraging. Indeed, the typical finding has been that only a small amount of variance in the dependent variable is explained by a variety of demographic and personality variables used as predictors. Two alternative explanations for these data are possible. First, repeated negative findings may indicate that there is no relationship between adoption and independent consumer variables, especially at the micro or individual level. If this is the case, then research resources should be put to more fruitful uses. The second explanation is that the specification of the research variables is inadequate. Since the operationalizing of most demographic variables is fairly standard this criticism mainly concerns the personality variable.

It is difficult to accept the first explanation. Behavior adoption of innovations has been successfully studied in a variety of other disciplines (Rogers and Shoemaker, 1971). The application of this concept to marketing and consumer behavior has too much face validity to dismiss it easily.

The personality variables are the soft underbelly of the problem. It is generally recognized (see KassarJian, 1971, or Worthing, et al., 1973) that selection of personality variables as such has not been carried out according to a systematic rationale for the appropriateness of the specific measure. In particular, the authors know of no studies that have devised specific measures for the behavior under consideration with anything approaching the care that could be used in arriving at paper-and-pencil measures of personality traits.

To test the effectiveness of the strategy of customized construction of a personality scale designed to fit a specific problem, the present project has attempted to measure innovativeness--a presumed trait underlying adoption of innovations.

This study was undertaken with the thought in mind that the explication of methodology of psychometric test construction is an area of opportunity for marketing. However, it is probably worth the considerable cost and effort only where the psychological trait has a potential significance extending beyond one product area. This was one reason for using adoption behavior. Along with the specificity of each of the content areas where adoption takes place, it is possible that the newness of the innovation as such will have a general effect over all categories. Also. instruments constructed by the technique used here have been among the few to show some success in marketing (Worthing et al., 1973).

PSYCHOMETRIC METHODOLOGY

There are two important strategy considerations in scale construction. First, the manner in which the items are generated: here the choice is between a rational approach versus an empirical approach. In the former the items represent the investigator's interpretation of an idea derived from theory or clinical intuition of a psychological trait that is real or important in a conceptual sense. The empirical strategy calls for the unsystematic selection of a large number of items taken from various sources which are then sorted on the basis of their ability to predict a criterion without regard to whether or not the relationship makes any sense beyond simple expediency.

Given a pool of items derived in either way, the second strategy concerns the manner in which the final items are selected. This may involve the issue Of response bias as well as other means for maximizing reliability and validity.

A valid test is one that measures what we want to measure, all of what we want to measure, and nothing but what we want to measure (Thorndike and Hagen, 1961). The three principal types of validity include content validity, predictive validity, and construct validity.

Content validity refers to the "adequacy with which a specified domain of interest is sampled" (Nunnally, 1967). The establishment of content validity is intimately related to the choice of strategies mentioned above... a universe of content is delimited, and then items are systematically drawn from that universe. Frequently, judges or panels of experts are required to assure the adequacy of content validation procedures, especially with the rational approach.

While content validity is dependent to a large degree upon subjective considerations, predictive validity can be more objectively established. Predictive validity refers to the magnitude of the association between the scale score and the behavior or criterion that the scale is attempting to identify. To the extent that a high degree of association exists between the scale score and the criterion, then the scale is said to have high predictive validity.

Construct validity refers to the a priori meaningfulness of the variable or construct under investigation. More specifically, construct validation provides at least a partial answer to the question, Does this scale tell us something about the subjects under investigation that is broadly meaningful and significant? (Thorndike and Hagen, 1961).

When groups known to differ on the particular variable are easily identifiable, then assessment of construct validity is trivial. [See Crano and Brewer, 1973, p. 253, for a complete explication.] However, when abstract constructs such as innovativeness are studied, then construct validation is more difficult. Campbell and Fiske (1959), as part of a more generic multi-trait, multi-method technique, suggest that convergent and discriminant validity may be one way to approach the assessment of construct validity Using this approach, other measures of the same trait and sometimes other traits are hypothesized to be related positively (convergent) or be independent to an extent beyond that caused by common method variance. If these hypotheses are upheld empirically, then construct validity is affirmed.

The second important issue in scale construction is reliability. In this context, reliability has two-distinct meanings. First, reliability measures the degree to which the items that make up the test are interrelated. This represents an internal consistency approach to reliability. Reliability has also been defined as the degree to which scores from a particular administration of a given test are related to scores on a subsequent administration of that same test to those same subjects This represents a temporal stability approach. Both indices of reliability are important and should be demonstrable.

The final important issue in scale construction is the suppression of response biases. Response biases include any systematic variance in the scale score that is not attributable to the trait being measured. Two principal sources of this variance are the traits of social desirability and acquiescence. Social desirability measures the need for social approval reflected in the tendency of a subject to respond to items in a way consistent with approved social norms rather than their true feelings. Acquiescence is the tendency of subjects to agree with statements. Both of these sources of variance must be controlled.

A RATIONAL APPROACH TO SCALE CONSTRUCTION

Jackson (1971) suggests the application of four principles of scale construction that consider each of these issues. These principles are enumerated below.

1. An explicit, theoretically-based definition of a particular trait is essential prior to attempts at measurement;

2. Careful empirical selection of items for homogeneity contributes substantially to refined measurement:

3. Suppression of response biases such as social desirability and acquiescence is best undertaken at the level of item selection and scale development; and

4. Both convergent and discriminant components of validity must be considered at every stage of scale development if the final scales are to Possess these proPerties (Jackson, 1967; from PRF manual, p. 15).

First, the diffusion adoption literature was evaluated as background data to assist in the development of the range of variables that were necessary for subsequent evaluation of results.

The literature review showed that open-mindedness (Jacoby, 1971), leadership (Ostlund, 1972), self-confidence (Ostlund, 1972), venturesomeness (Robertson and Kennedy, 1968), empathy (Rogers and Stanfield, 1968), gregariousness (King and Sproles, 1973), and student activism (King and Sproles, 1973) were sometimes related to adoption.

Second, a long tradition of psychological research relating suggestibility and persuasibility to lowered self-esteem was considered. In a contrasting vein, a series of studies begun by Asch emphasized the independence and lack of conformity of subjects. These two research trends seemed to constitute two cells of a matrix that might be labeled "good independence" and "bad conformity." The other two cells seemed relatively empty: "bad independence" being related, perhaps, to negativism or--more recently--to reactance. Only the very general area of problem solving seems related to "good conformity." This would represent the ability of the individual to pick up cues or to change set in order to use the resources of the situation creatively to solve a problem.

It seemed possible that a trait existed that underlay the intelligent, creative, selective use of communication for solving problems. In other words, being persuasible where it was adaptive for the recipient of the persuasive message. Surely we sometimes learn something of value from mass media and from opinion leaders. This led to defining a trait of "innovativeness" as follows:

A person high on the trait of innovativeness is open to new experiences and often goes out of his way to experience different and novel stimuli particularly of a meaningful sort (not just thrill-seeking). Most important, he tends to make constructive use of information received whether sought or accidentally encountered. He has a low threshold for recognizing the potential application of ideas he gets from others but does not apply suggestions mechanically. Rather, he has the ability to transform information for his own use. His involvement in his own enterprises and acts is such that he looks for ways to change and improve them. Above all, he is responsive to communication in a selective, constructive way when the message has a valid relevance to his activities. He is objective in his evaluation although occasionally naive.

Psychologists have attempted to measure rigidity, dogmatism, closed mindedness, submissiveness, and suggestibility, but there have been few, if any, attempts to scale the open-minded, constructive trait that we have in mind.

A group of three experts was assembled and after discussing the definition and type of person who was high on the trait, during several sessions over a period of two weeks produced 144 prospective positive innovativeness items. In addition, a twenty-item social desirability scale was included to measure desirability response bias (Crown and Marlowe, 1964). This total set of items was then administered to 300 female respondents.

The Jackson (1967) method of item analysis was used next to maximize reliable content variability in relation to social desirability response variance. For each item, the correlation between the scale item and the total desirability score was subtracted from the correlation between that scale item and the total scale score. The square root of the remainder is the proportion of true content (innovativeness) variance for any given item. Jackson calls this value the Differential Reliability Index (DRI). Obviously, the higher the DRI, the more the item is reflective of the substantive content (innovativeness) of the item. From this administration, items with DRI's of .39 or greater were retained. This amounted to 29 positive items from the original pool of 144.

While this procedure eliminated desirability response bias, acquiescence as still a problem. Therefore, a second questionnaire was constructed containing 196 items to deal with additional problems of scale development. These items included the 29 positive innovativeness items, 33 negative innovativeness items, 14 psychological subscales (94 items), and a 40-item new-product checklist (to be discussed in another report).

The 29 positive innovativeness items were retained from the previous analysis. The negative innovativeness items were included to alleviate acquiescence response bias. A large number of negative items were constructed by reversing the positive items in a sensible way. Some of these items were literal reversals of the positive items, while the remainder were more broadly a negation of innovativeness.

Fourteen psychological scales were included for construct validation purposes. If innovativeness is an independent trait it should show only moderate relationships with these other traits when compared to the correlation between the two innovativeness forms. Descriptions of these scales and hypothesized direction of relation with innovativeness are shown in Table 1. These hypothesized relationships were drawn from the sources previously cited. The questionnaire was administered to 299 women from a variety of demographic categories.

TABLE 1

DESCRIPTION OF PSYCHOLOGICAL SCALES AND HYPOTHESIZED RELATIONSHIPS WITH INNOVATIVENESS

RESULTS

Item Selection

Items selected for the final scales were given consideration under various criteria, primarily the whole-part correlation of each innovativeness positive item with positive total and each negative item with the negative item total. To qualify for inclusion in the scale, positive innovativeness items had to correlate at least moderately high with the positive total. In a similar manner negative items needed to be correlated at least moderately with the negative item total. The 20 positive and 20 negative items which most closely met these criteria were selected for the innovativeness scale. A listing of these items is available on request. [A copy of the final forms can be obtained by writing Prof. Clark Leavitt, The Ohio State University, 1775 College Road, Columbus, Ohio 43210.]

Only scores of females are included in these data. Therefore, it is possible that these items are relevant only for women. A preliminary analysis has been undertaken for male subjects and the similarity of the male data suggests that the items will work for both males and females.

Reliability

Two internal consistency measures of reliability were calculated. These measures included the Spearman-Brown split-halves method and Kuder-Richardson formula 20. The value of these reliability coefficients for the 40-item test and alternate forms A and B are shown in Table 2.

TABLE 2

RELIABILITY DATA

The reliability coefficients for the 40-item test are quite high. Corresponding reliability values for both alternate forms are, as expected, somewhat lower. These values represent, however, acceptable levels of reliability and justify the decision to develop two forms.

Divergent validity was assessed through the use of the fourteen psychological subscales. As shown in Table 1, it was hypothesized that innovativeness should show a moderate positive relation to political control, internal/external control, confidence in others, leadership, culture vulture, competence, energetic, communication, community-minded, counterculture, well-being, and good spirits.... a negative relationship was hypothesized with helplessness and inflated self-esteem. Actual data on these relationships are presented in Table 3.

The data in Table 3 indicate that these hypotheses were upheld with a few exceptions: internal/external control and communication were slightly negatively related to innovativeness, and inflated self-esteem was positively related to innovativeness. None of the correlations are high enough to suggest identity. These results indicate that at least a moderate degree of independence is evident for Innovativeness.

CONCLUSION

The objective of this paper was to construct a valid and reliable scale to measure innovativeness--a trait postulated to underlie adoption behavior. Data presented indicate that either of two forms meet this objective. The success of these scales and, by inference, the developmental methodology outlined in this paper depend on the utility of these scales and this methodology for marketing researchers and practitioners in many different environments. The authors eagerly await these results. Meanwhile, further studies are underway to assess the effectiveness of Innovativeness in predicting adoption behavior of various kinds.

TABLE 3

DIVERGENT VALIDITY CORRELATION OF PSYCHOLOGICAL SCALE TOTAL SCORES WITH INNOVATIVENESS SCALES

REFERENCES

Asch, S. E. Social psychology. New York: Prentice-Hall, 1952.

Bentler, P. M., Jackson, D. N. & Messick, S. Identification of content and style: A two-dimensional interpretation of acquiescence. Psychological Bulletin, 1971, 76, 186-204.

Campbell, D. T. & Fiske, D. W. Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 1959, 56, 81-105.

Crano, W. D. & Brewer, M. B. Principles of research in social psychology. New York: McGraw-Hill, 1973.

Crowne, Douglas P. & Marlowe, David. The approval motive: studies in evaluative dependence. New York: John Wiley and Sons, Inc., 1964.

Endler, N. & Hunt, G. M. B. S-R inventories of hostility and comparisons of the proportion of variance from persons, responses, and situations for hostility and anxiousness. Journal of Personality and Social Psychology, 1968, 9, 309-315.

Engel, J. F., Kegerreis, R. J. & Blackwell, R. D. Word of mouth communication by the innovator. Journal of Marketing, 1969, 33, 15-19.

Frank, R. E., Massy, W. F. & Morrison, D. G. The determinants of innovative behavior with respect to a branded, frequently purchased food product. In G. Smith (Ed.), Reflections on progress in marketing. Chicago: American Marketing Association, 1964.

Jackson, D. N. The dynamics of structured personality tests. Psychological Review, 1971, 78, 229-248.

Jackson, D. N. Manual for the personality research form. Research bulletin 43. London, Ontario: University of Western Ontario, 1967.

Jacoby, J. Personality and innovation proneness. Journal of Marketing Research, 1971, 8, 244-7.

Kassarjian, H. H. Personality and consumer behavior: A review. Journal of Marketing Research 1971, 8, 409-418.

King, C. W. The innovator in the fashion adoption process. In G. Smith (Ed.), Reflections on progress in marketing. Chicago: American Marketing Association, 1964.

King, C. W. & Sproles, G. B. The explanatory efficacy of selected types of consumer profile variables in fashion change agent identification. Unpublished working paper, Krannert Graduate School of Industrial Administration, Purdue University, 1973.

Mischel, W. Personality and assessment. New York: John Wiley and Sons, Inc., 1968.

Nunnally, J. C. Psychometric theory. New York: McGraw-Hill, Inc., 1967.

Ostlund, L. C. Identifying early buyers. Journal of Advertising Research, 1972, 12, 25-30.

Pessemier, E. A., Burger, P. C. & Tigert, D. J. Can new product buyers be identified? Journal of Marketing Research, 1967, 4, 349-55.

Robertson, T. S. Determinants of diffusion behavior. American Marketing Association, Proceedings, 1967.

Robertson, T. S. Innovative behavior and communication. New York: Holt, Rinehart and Winston, Inc., 1971.

Robertson, T. S. The process of innovation and the diffusion of innovation. Journal of Marketing, 1967, 31, 14-19.

Robertson, T. S., & Kennedy, J. N. Prediction of consumer innovators: Application of multiple discriminant analysis. Journal of Marketing Research, 1968, 5, 64-9.

Rogers, E. M. Diffusion of innovations. New York: The Free Press, 1962.

Rogers, E. M. & Shoemaker, F. F. Communication of innovations. New York: The Free Press, 1971.

Rogers, E. M. & Stanfield, J. D. Adoption and diffusion of new products: emerging generalizations and hypotheses. In F. M. Bass, E. A. Pessemier, and C. W. King (Eds.), Applications of the sciences in marketing management. New York: John Wiley and Sons, Inc., 1968.

Thorndike, R. L. & Hagen, E. Measurement and evaluation in psychology and education. New York: John Wiley and Sons, Inc., 1961.

Worthing, P. M., Venkatesan, M. & Smith, S. Personality and product use revisited: An exploration with the personality research form. Journal of Applied Psychology, 1973, 57, 179-183.

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