Communicating Innovations: Convincing Computer Phobics to Adopt Innovative Technologies

Thomas Hill, University of Tulsa
Nancy D. Smith, University of Tulsa
Millard F. Mann, University of Kansas
ABSTRACT - In a recent paper, Lepper (1985) has pointed out that the rapid development of microcomputer technology has provided a unique opportunity to investigate, in vito, the process of technology adoption. This paper summarizes the results of two experiments that were concerned with the responses of "high-tech phobics" to innovative technologies. Previous research has demonstrated that personal efficacy with respect to computers is a strong predictor of subsequent adoption of computer technology. The results of the present research show (a) that the more technologically advanced a product, the more important a factor is personal efficacy in the decision to adopt the technology, and (b) that people low in personal efficacy (as compared to people high in personal efficacy) with regard to computers tend to be more easily persuaded by expert communicators to try an advanced software product.
[ to cite ]:
Thomas Hill, Nancy D. Smith, and Millard F. Mann (1986) ,"Communicating Innovations: Convincing Computer Phobics to Adopt Innovative Technologies", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 419-422.

Advances in Consumer Research Volume 13, 1986      Pages 419-422

COMMUNICATING INNOVATIONS: CONVINCING COMPUTER PHOBICS TO ADOPT INNOVATIVE TECHNOLOGIES

Thomas Hill, University of Tulsa

Nancy D. Smith, University of Tulsa

Millard F. Mann, University of Kansas

[Address all correspondence to Thomas Hill, Department of Psychology, University of Tulsa, 600 South College Avenue, Tulsa, OK 74104; (918) 592-6000, X2248.]

ABSTRACT -

In a recent paper, Lepper (1985) has pointed out that the rapid development of microcomputer technology has provided a unique opportunity to investigate, in vito, the process of technology adoption. This paper summarizes the results of two experiments that were concerned with the responses of "high-tech phobics" to innovative technologies. Previous research has demonstrated that personal efficacy with respect to computers is a strong predictor of subsequent adoption of computer technology. The results of the present research show (a) that the more technologically advanced a product, the more important a factor is personal efficacy in the decision to adopt the technology, and (b) that people low in personal efficacy (as compared to people high in personal efficacy) with regard to computers tend to be more easily persuaded by expert communicators to try an advanced software product.

INTRODUCTION

A new business was recently described in a local newspaper of a midsized midwestern town (The Tulsa World, February 8, 1985) . For a reasonable fee, a technician will come to your home, connect your YCR (video cassette recorder), stereo, or other "sophisticated" appliance, set all switches and options, and show you how to operate your electronic gadget. As the manager of the business put it, "technology has just gotten away from people." Business is booming.

"Computer phobic" has become the common term for describing a person who avoids computers, regardless of how useful they might be. Considering the success of the business offering the service described above, it appears that many people are afraid not only of computers but of "high-tech" electronic consumer goods in general. "Hightech" phobia is probably a better, more general term for describing the anxiety and uncertainty that many of us experience when dealing with a technologically advanced product for the first time.

Most consumer research concerned with the adoption of new technology has focused almost exclusively on the identification of the life style and personality of early adopters (Danko and MacLachlan 1983, Dickerson and Gentry 1983) . The results of these studies have shown that, as one would expect, the early adopter of, for example, computers is not at all afraid of new technology; on the contrary, the early adopter perceives these innovations as a challenge (Danko and MacLachlan 1983).

Other recent research (Hill, Smith, Mann, and Roberson in press; Hill and Smith 1985; Smith and Hill 1985) has applied Bandura and his associates' theoretical framework (Bandura 1977, Bandura and Schunk 1981, Bandura, Adams and Beyer 1977) . According to this view, phobias can be explained in terms of lack of personal efficacy, i.e., the belief that a particular behavior cannot successfully be performed. If an individual thinks that he or she is unable ever to successfully interact with a computer, then it is likely that he or she will avoid such interactions, regardless of how useful it might be to learn about or use computers.

In their research, Hill and his associates (Hill et al. in press, Hill and Smith 1985) used pathanalytic methods to show that, in three independent samples of college students, computer efficacy (a) uniquely contributes to the prediction of behavioral intentions to purchase a microcomputer, (b) is independent of (not correlated with) people's beliefs about the instrumental value of learning about computers, and (c) is positively related to people's previous experience with computers. Further, Hill et al. (in press) showed that personal efficacy in general, but not interpersonal efficacy (i.e., efficacy with respect to interpersonal relationships), as measured with Paulhus' sphere specific measures of control (Paulhus 1985), is related to the use of various technologically advanced products (e.g., programmable pocket calculators, automated bank teller machines, automatic garage door openers).

Phenomenologically, personal efficacy with respect to technological innovations can be construed both as a characteristic of the person perceiving an innovation (e.g., "I can make this machine work") and as a characteristic of the innovation ("I can make this machine work"). In the research described above, personal efficacy with respect to computers was treated as a characteristic of people, i.e., as a personality variable: Some people report less efficacy with respect to computers and they are less likely to learn about or use them; others report more efficacy and are more likely to learn about or use them. However, the person who lacks efficacy with respect to an innovation will not perceive him or herself as "low in efficacy," but will perceive the innovation as complicated or complex. The more complex the innovation actually is the more likely it is that people will lack efficacy with respect to the innovation. As Rogers and Shoemaker (1971) have pointed out, complexity of an innovation - a characteristic of the innovation itself - is an important factor determining the rate of its adoption.

STUDY 1

Study 1 was designed to explore the importance of efficacy in the decision to adopt new products of varying complexity. It was expected (H1) that the more complex a product is, the more likely it is that people will feel little personal efficacy with respect to the product, and, hence, the less likely it is that they will purchase it, and (H2) that the more complex a product is, the more important a role will personal efficacy play in the decision to purchase it.

Method

Eighty-three male and female college students participated for extra credit in their courses. Data were collected in ten experimental sessions with between sis and ten participants each. The study was introduced as an advertising evaluation study. Subjects were asked to evaluate three new products based on "rough drafts" of advertisements. These products were an electric typewriter (presumably low in complexity), an electronic typewriter (presumably of moderate complexity), and a personal computer with wordprocessing capabilities (presumably of high complexity). Subjects were asked to complete a set of questionnaires containing a scale designed to measure efficacy with respect to each product, a scale designed to measure perceived instrumentality of each product, and a scale designed to measure liking of each product. All scales pertaining to a particular product were preceded by the description of the product; the order in which the products were presented was carefully counterbalanced.

Product descriptions. All product descriptions began with the following sentence: "Consider the following advertising copy describing the features of a new [personal computer with wordprocessing/electronic typewriter/electric typewriter] about to be put on the market. This product is designed to assist in the preparation of professional-looking manuscripts and term papers." This introduction was followed by a list of six features appropriate for the respective product (e.g., personal computer: 64k of memory, screen editing, etc.; electronic typewriter: one-step error correction, format memory, etc.; electric typewriter: adjustable tabs, power carrier return, etc.).

Questionnaire. The questionnaire contained seven items designed to measure personal efficacy with respect to each product. These items were worded similarly to those of the computer efficacy scale developed by Hill et al. (in press) (e.g., I will never understand how to use this product [reversely coded], I could reliably control the functioning of this product, This product would be easy to use). Eight items were designed to measure perceived instrumentality of the respective product (e.g., I could do a better job with this product, I could accomplish more by using this product, Time savings from using this product would be minimal [reversed]). Each item was accompanied by a 6-point Likert type scale, labeled "strongly disagree" or "strongly agree" at its extremes.

The items for both scales were chosen from a larger pool of items based on the criterion that the items from each subscale correlate fairly highly among themselves for each of the three products. It should be stressed that the scales were constructed before the analyses to test the hypotheses of this research were performed. This procedure resulted in reliability coefficients (Cronbach's a) for each scale within each product in excess of a = .88. Although this procedure ensured the reliability of scales, it, of course, does not ensure validity. The items -comprising the efficacy scale, however, are, as mentioned above, very similar to those comprising the computer efficacy scale used in a previous study by Hill et al. (in press). In that study scores on the computer efficacy scale were shown to validly predict students' subsequent enrollment in university computing courses, a finding which supports the validity of the efficacy scale in the current context. Nevertheless, it should be pointed out that the primary reliance on the face validity of items in the present study poses limitations to the confidence with which conclusion may be drawn.

The questionnaire also contained four items designed to measure liking of the product. The items were formulated so that they asked for the likelihood that the respondent would behave in a way indicative of product adoption (Would you recommend this product to 8 friend? If money were not a consideration, would you purchase this product? Would you like to learn more about this product? Would you like to own this product?). The reliability of this scale for each product was always greater than a = .90.

Results

In order to test whether the three product descriptions elicited differential perceptions of efficacy, a repeated measures analysis of variance was performed on the personal efficacy scores across the three products. As expected, subjects' ratings of personal efficacy with respect to the three products were significantly different, F(2,164) = 42.24, p < .0001. Subjects reported greater efficacy with respect to the electric typewriter than with respect to the electronic typewriter (means were M - 31.02 and M = 53.29, respectively, t[82] - 8.46, p < .001), and greater efficacy for the electronic typewriter than for the personal computer with wordprocessing (M = 36.34, t[82] = 5.99, p < .001). It can be concluded from this analysis that the different products, designed to vary in complexity, led to differences in subjects' ratings of personal efficacy.

In order to test whether differential personal efficacy with respect to the three products is related to differential liking of these products, a repeated measures analysis of variance was performed, treating expressed liking of the three products as three levels of the dependent variable, and treating personal efficacy and instrumentality with respect to the three products as two covariates that were repeated for each level of the repeated measure (this analysis was performed with the MANOVA routine in SPSSX). This analysis revealed that, overall, only perceived instrumentality was related to the average liking of all three products, F(1, 80) = 51.16, p < .001. This effect probably reflects the extent to which students perceive any machine that facilitates the writing of manuscripts as instrumental for facilitating their everyday life. Overall, personal efficacy with regard to the three products was not related to average liking of the products, F(1, 80) = 1.46, p > .20. However, differential personal efficacy (i.e., the difference in personal efficacy across the three products) was related to differential liking of the products, F(2, 154) = 6.02, p < .01.

Finally, in order to assess the importance of efficacy in the prediction of liking, separate zero order correlation coefficients between scales within each product were calculated. As expected, the zero order correlation between efficacy and liking increased as the product became more complex. The correlations were r (81) = .05, n.s., for the electric typewriter; r (81) = .43, p < .001, for the electronic typewriter; and r (81) = .64, for the personal computer with wordprocessing. It should be noted that the standard deviations for each of the scales for each product were virtually identical (efficacy: 5.3, 5.6, and 6.2, for the electric typewriter, the electronic typewriter, and the personal computer with wordprocessing, respectively; liking: 5.3, 4.3, 3.8, respectively), and that all correlations between instrumentality beliefs and liking were in the range from r (81) = .67 to r (81) = .72. Because the reliabilities for the efficacy scale across products were virtually identical as well, it appears unlikely that the change in correlations for efficacy with liking across products was an artifact due to differences in variances or reliability of measurement.

The results of Study 1 support the hypothesis that greater complexity of a product implies less efficacy with respect to the product, and, further, that differential efficacy with respect to different products is a predictor of differential liking (or, as measured in this study, of expressed likelihood to perform behaviors indicative of product adoption). Finally, it appears that the more complex a product, the more important a factor is efficacy with respect to product adoption.

STUDY 2

Study 1 provided further support for the notion that personal efficacy plays an important role in the adaption of innovative technology, particularly when it is complex. The question arises as to how to change personal efficacy with respect to a specific technological innovation. For example, how can one convince a "computer phobic" that he or she is capable of learning how to successfully interact with a computer? One straightforward approach to solving this issue would be simply to tell the person that it is indeed quite easy to make computers work. Of course, the trick is to get the person to believe the communicator.

The classic Yale communication studies on persuasion (Hovland, Janis and Kelley 1953) have pointed to the importance of source credibility in determining the effectiveness of a persuasive communication. With respect to innovative technologies in general, and computers in particular, the question is whether computer phobics are more likely to trust the opinion of an expert or the opinion of a relative novice, who perhaps used to be a phobic him or herself.

Actually, both predictions could be derived from different theoretical positions: On the one hand, the findings of the Yale communication group suggest that the opinion of a computer expert should carry more weight than that of a computer novice. Furthermore, previous research on locus of control and persuadability has demonstrated that people with an external locus of control (i.e., people who believe that they are being controlled rather than in control themselves) are more easily persuaded by high prestige, expert sources (Ritchie and Phares 1969, Ryckman, Rodda and Sherman 1972). Following this logic it may be argued that people who think that they are incapable of successfully interacting with computers are more likely persuaded to purchase a computer (or computer related product) by "experts who know what they are doing" than are people who feel that they are in control and quite capable of learning to use computers.

On the other hand, social comparison theory (Festinger 1954) predicts that computer phobics are more likely to trust and be influenced by the judgement of people who are similar to themselves, while people who feel confident about interacting with computers are less likely to trust the judgment of such individuals, because they are different from themselves. In practice, the issue for advertising is whether to present a computer novice attesting to the ease of use of a new computer or new computer software, or whether to introduce an expert who assures potential customers of the "userfriendliness" of the new system or software. Study 2 addresses this issue.

Method

A questionnaire designed to measure computer-specific efficacy (the same scale used by Hill et al. in press) was administered to a sample of 87 female college students. Subjects for the present study were recruited from among those who had completed this questionnaire. Because the questionnaire was administered along with various other questionnaires (in a mass testing session), subjects were not aware that the current research was concerned with computer related efficacy.

A total of 54 subjects volunteered to participate. However, of these subjects, ten (five in each experimental condition) had actually not completed the computer efficacy scale during mass testing. The data for these subjects were excluded from all analyses involving this measure.

Procedure. The entire study was conducted in 12 sessions with four to six participants per session. The experiment was conducted in the preview room of the media services department in the library of a midsized private midwestern university.

Upon arrival, the experimenter explained that the study was concerned with the evaluation of an innovative software package specifically designed for college students, called the "College Advisor." All subjects were then handed the same brief description of this program. To foster the credibility of the cover story, the description was printed on the stationary of a midwestern research firm.

The program was described as an integrated wordprocessing and database management program with special subprograms for college students to assist in the planning of classes, writing of papers, and monitoring of grade-point averages. After all subjects had read the description of the program, the experimenter explained that the sponsoring research firm was interested in college students' evaluation of the program. The experimenter then started a slide show which was presented as a "rough draft" of a potential advertisement for the program.

Manipulation of source expertness. There were two different versions of the slide show. The actual slides, which depicted four different females working at a personal computer, were the same in each condition. The slide show was synchronized with one of two different tape recording (recorded by the same individuals) of testimonials of the four women. The four women described themselves as either a computer science major, an B A student specializing in Management Information Systems, a mathematics major, and an engineering major (expert sources), or as a fine arts major, a language major, a music major, and an English literature major (non-expert sources). In both conditions, the individuals presented in the slide show commented on the ease with which the program could be learned. Subjects (groups of subjects) were randomly assigned to one of the two conditions (slide shows).

Dependent measure. At the conclusion of the slide show, subjects were asked to complete a set of questionnaires evaluating the computer expertise of the individuals presented in the slide show (manipulation check). Also, subjects were given the opportunity to sign up (and put down their address) for a six-month trial period of the College Advisor. It was explained that the program would be provided with a special microcomputer, the rental of which would cost approximately $10.00 per month. This amount was determined in pretesting to be sufficient to prevent approximately 50% of all college students from signing up for the trial period. This measure constituted the dependent measure of adoption of this new product.

In order to reduce potential experimental demand (i.e., the experimenter's being perceived as interested in subjects' signing up), subjects were handed, along with the questionnaire, an envelope addressed directly to the sponsoring research organization. Subjects were told that they should put the completed questionnaire in the envelope, seal it, and drop it in a nail box when leaving.

At the end of the experimental session, subjects were asked to complete the computer efficacy questionnaire which most had previously completed during the mass-testing session. They were then thoroughly debriefed and thanked for their participation.

Results

Overall, the manipulation of source expertness was successful. One item on the advertising evaluation questionnaire asked for subjects' rating of the computer expertise of the individuals depicted in the slide show (1 = none at all, 9 = very much). Subjects who had watched the slide show depicting the expert sources rated the expertise of these individuals significantly higher than subjects who were exposed to the non-expert sources, t (51) = 7.59, p < .0001. The means in the expert and non-expert condition were 2.70 and 6.08, respectively. However, subjects' ratings of the individuals presented in the slide show did not differ with regard to overall appeal (all F's < 2.00).

The major dependent variable of interest was whether or not subjects signed up for the trial period of the College Advisor. Overall, of the 27 subjects who watched the non-expert sources, 19 signed up for the trial period; of the 27 subjects who watched the expert sources, only 11 signed up. This difference is significant, X2 (1) = 4-90, p < .05. In order to assess the potentially differential effect of the slide shows on subjects high or low on computer efficacy, the biserial correlation coefficient between computer efficacy (both the initial measure and the measure that subjects completed after the experiment) and subjects' choices regarding the trial period was calculated within each experimental condition. In the non-expert condition, computer efficacy was not related to subjects' choices, r (25) = .05 for the premeasure of computer efficacy, and r (25) = -.03 for the postmeasure. However, in the expert source condition, both measures appeared to be related to subjects' choices, r (22) - -.38, p < .10, and r (25) = -.52, p < .01, respectively. Performing a median split on the computer efficacy scale shows that subjects who were low on computer efficacy were more likely persuaded by the expert sources than subjects high in computer efficacy; 67% of the subjects in the former group signed up for the trial period with the College Advisor as compared to 20% of the subjects who scored high in computer efficacy. The non-expert sources did not differentially influence low and high computer efficacy subjects (67% and 75%, respectively, signed up for the trial period).

The results of Study 2 support the prediction derived from the research on locus of control and persuasion. Although, given the particular scripts used in this study, the no^-expert sources were more successful in persuading subjects to sign up for the trial period, it appears that the less computer efficacy subjects reported, the more likely were they to trust an expert's opinion.

CONCLUSION

The results of the current research support the notion that personal efficacy, i.e., the extent to which a potential consumer thinks that he or she will be able to successfully "master" a "high-tech" product, is an important variable contributing to or inhibiting technology adoption. Personal efficacy with respect to a particular product is a joint function of the "objective" complexity of the product and of the psychological characteristics of the potential consumer. Study 1 was concerned with products of varying complexity. The results suggest that the more complex a product is in actuality, the more important personal efficacy is in the decision whether or sot to adopt the product. Study 2 was concerned with the effectiveness of expert and nonexpert sources in influencing individuals who possessed varying degrees of personal efficacy with respect to a particular product. The results of Study 2 suggest that experts attesting to the "userfriendliness" of a new product are more likely to convince individuals low in efficacy with respect to the product than individuals high in efficacy with respect to the product.

REFERENCES

Bandura, Albert (1977), "Self-efficacy: Toward a Unifying Theory of Behavioral Change," Journal of Personality and Social Psychology, 84 (March), 191-215.

Bandura, Albert, Nancy E. Adams, and Janice Beyer (1977), "Cognitive Processes Mediating Behavioral Change," Journal of Personality and Social Psychology, 35 (March). 125-139.

Bandura, Albert and Dale H. Schunk (1981), "Cultivating Competence, Self-Efficacy, and Intrinsic Interest through Proximal Self-Motivation," Journal of Personality and Social Psychology, 41 (September), 586-598.

Danko, William D. and James M. MacLachlan (1983), "Research to Accelerate the Diffusion of a New Innovation: The Case of Personal Computers, in Journal of Advertising Research, 23 (June/July), 39-43.

Dickerson, Mary D. and James W. Gentry (1983), "Characteristics of Adopters and Non-Adopters of Home Computers," Journal of Consumer Research, 10 (September), 225-235.

Festinger, Leon (1954), "A Theory of Social Comparison Processes," Human Relations, 7, 117-140.

Hill, Thomas, Nancy D. Smith, Millard F. Mann and Bruce F. Roberson (forthcoming), "Efficacy Expectations and Technology Adoption: The Case of Computers, n in Proceedings of the Division of Consumer Psychology, 1985 Annual APA Convention.

Hill, Thomas and Nancy D. Smith (1985), The Role of Locus of Control in the Adoption of Computer Technology, Paper presented at the Thirty-First Annual Convention of the Southwestern Psychological Association, Austin, TX.

Carl I. Hovland, Irving Janis, and Harold H. Kelley (1953), Communication and Persuasion, New Haven, Conn.: Yale University Press.

Lepper, Mark R. (1985), "Microcomputers in Education: Motivational and Social Issues," American Psychologist, 40 (January), 1-18.

Paulhus, Delroy (1983), "Sphere-Specific Measures of Perceived Control," Journal of Personality and Social Psychology, 44 (June), 1253-1265.

Ritchie, Elaine and E. Jerry Phares (1969), "Attitudes Change as a Function of Internal-External Control and Communicator Status", Journal of Personality, 37, 429-443.

Ryckman, Richard M., William C. Rodda, and Martin F. Sherman (1972), "Locus of Control and Expertise Relevance as Determinants of Changes in Opinion about Student Activism," Journal of Social Psychology, 88 (October), 107-114.

Rogers, Everett M and F. Floyd Shoemaker (1971), Communication of Innovations: A Cross-Cultural Approach, New York, New York: The Free Press.

Smith, Nancy D. and Thomas Hill (1985), A Scale for Predicting Purchase Intentions for Technologically Advanced Products, Paper presented at the Thirty-First Annual Convention of the Southwestern Psychological Association, Austin, TX.

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