The homepage is the most important page for every website because it is a company’s face to the world of the Internet. Visitors who gain a bad initial impression of a homepage may choose to go to other websites (Nielsen & Tahir, 2002). From a marketing perspective, the immediate impact of a poor homepage design is not only a lost opportunity to attract potential customers, but also a damaged company image (Singh & Dalal, 1999).

Many homepages are crowded with information and advertisements (Rau, Gao, & Liu, 2007). This crowding may be caused by designers who believe that users should see as much information as possible immediately when they visit a website. To date, however, little experimental evidence of how people evaluate a website with a high-density homepage has been collected. Therefore, in this study we empirically investigated the effects of homepage screen density on users’ attitudes and behavioral intentions towards a website.

In addition, we investigated the role of individuals’ personality types in regard to the extent to which homepage screen density influences website evaluations. In previous studies personality has been identified as a critical factor in the design of information systems (e.g., Benbasat & Taylor, 1978; Wheeler, Hunton, & Bryant, 2004), and an important factor in determining users’ behavior on the Internet (Huang & Chang 2008). Thus, we argue that individual preferences for the design of homepage may be influenced by personality. This knowledge will assist designers in creating homepages that suit the personality types of their target audiences.

Theoretical Background and Hypotheses

The Effects of Screen Density

One of the fundamental issues affecting designers’ decisions about how much information should be put on the homepage is screen density (Williams, 2000). This is defined as the percentage of the number of characters divided by the total amount of spaces available on a computer screen (Staggers, 1993). Homepage screen density is defined here as a percentage of the total available browser space that is occupied by texts, links, graphics, and animations.

Previous research on screen density has been focused on users’ visual search performance, and its effects on subjective screen satisfaction (e.g., Rau et al., 2007; Staggers, 1993). The findings appear to be contradictory. In some studies it was found that participants preferred high-density screens (Spool, Schroeder, Scanlon, & Snyder, 1998), while others preferred a lower density (Ross, Morrison, & O’Dell, 1988). In other studies the relationship was found to follow an inverted U-shaped curve, with too little or too much information on a screen depressing preference and performance (Coll & Wingertsman, 1990). Besides the inconsistent findings, most research has been focused on text-only screens, and the monitors used were monochrome or low-resolution color CRT terminals. As technology has changed dramatically since these studies were carried out, it is questionable whether past findings can be generalized to include homepages that contain visual elements other than text, or to Internet surfing activities other than searching. Consequently, in this study we investigated the influence of homepage screen density on users’ website evaluations.

The Moderating Role of Personality Types

It has long been assumed that the design of information systems (IS) should consider the psychological characteristics of users (Shneiderman, 1980). One such individual personality difference that has received considerable attention in IS research is cognitive style (e.g., Benbasat & Taylor, 1978; Wheeler et al., 2004). Cognitive style is defined as “consistent individual differences in preferred ways of organizing and processing information and experience” (Messick, 1976, p. 5). Although several classifications of cognitive style exist, the most widely adopted is based on Jung’s (1923) theory of psychological types as operationalized by the Myers-Briggs Type Indicator (MBTI). The MBTI consists of four bipolar psychological dimensions: Introversion/Extroversion (I/E), Sensing/Intuiting (S/N), Thinking/Feeling (T/F), and Judging/Perceiving (J/P) (Myers & McCaulley, 1985). Researchers can use the individual dimensions or combinations of these to investigate specific characteristics of individuals in relation to unique situations (Wheeler et al., 2004). Since there is no literature to support an association between the I/E dimension and the evaluation tasks in this study, we did not examine it. Nor did we investigate the J/P dimension, as it is correlated with the T/F dimension (Carlyn, 1977), and does not contribute independently to the understanding of personality (Abrahamian, Weinberg, Grady, & Stanton, 2003). Therefore, we focused on two dimensions, S/N and T/F, that relate directly to people’s preferences for information gathering and processing. They could, therefore, also be expected to affect a person’s perception and evaluation of the design of a homepage.

The S/N dimension relates to ways in which a person perceives information. Sensing types prefer to acquire and verify information through their physical senses. These people tend to focus on objective facts and details. Intuiting types prefer to receive information as it exists in totality, and are impatient with details. According to Jung’s type theory (Gould, 1991; Myers, 1987), sensing types are more patient with details and even seek some repetition of information. Thus, they like a high-density homepage which contains more information. However, intuiting types dislike devoting time to precision and become impatient with routine details. Therefore, they would avoid a high-density homepage. This led us to the following hypothesis:
Hypothesis 1: The homepage screen density will have a greater influence on website evaluations for Intuiting types than for sensing types.

The T/F dimension relates to ways in which a person processes information or makes decisions. Thinking types tend to be more analytical, impersonal, and objective. In contrast, feeling types tend to make decisions based on personal and group values. According to Jahng, Jain, and Ramamurthy (2002, p. 184) “Rather than seeking precision they prefer more persuasive information and are generally more empathetic”. According to Jung’s type theory (Gould, 1991; Myers, 1987; Nutt, 1993). In the context of website browsing, a moderate-density homepage is expected to be better suited to the information preferences of feeling types. However, thinking types seek more precise information and engage in analytical processing. Therefore, they would be less influenced by the differences between high- and moderate-density homepages. Hence it is proposed that:
Hypothesis 2: The homepage screen density will have a greater influence on website evaluations for feeling types than for thinking types.

The perceiving mental function (S/N), which contains sensing and intuiting functions, and the judging mental function (T/F), which contains feeling and thinking functions, can be combined into four cognitive styles: sensing-thinking (ST), sensing-feeling (SF), intuiting-thinking (NT), and intuiting-feeling (NF) (Mitroff, 1981). In some sense, ST and NF types are extremes in term of their information perceiving, processing, and decision making (Jahng et al., 2002). It has been demonstrated that a person’s cognitive style may influence his or her preferences and decision making (Haley & Stumpf, 1989). In e-commerce environments, Jahng et al. found that the effectiveness of a product’s information presentation varied according to consumers’ cognitive styles. Accordingly, an individual’s cognitive style may help explain the effects of homepage screen density on his or her website evaluations. Thus, we propose that:
Hypothesis 3: The effect of homepage screen density on website evaluations will vary according to users’ cognitive styles.

Method

Experimental Design

A controlled laboratory experiment was used to test the study hypotheses. The design of the experiment involved one within-subject variable: homepage screen density. In the past, a wide range of density levels have been used. In this study, two conditions only of homepage screen density (moderate and high) were manipulated. Using previous studies (Coll & Wingertsman, 1990; Staggers, 1993) and the design patterns of several commercial homepages, we set 75% as high-density and 50% as moderate-density. Two versions of a homepage were constructed. They were based on a professionally developed homepage in order to maintain external validity. This homepage was selected for two reasons. First, it was a typical high-density commercial homepage. Second, we expected that few users in Taiwan would have visited this website previously, as it is located in China. This could eliminate any influence of previous experience. The moderate-density homepage was manipulated by decreasing the amount of detailed information and the number of banner advertisements. The two versions were identical except for the amount of information displayed on them.

Participants and Procedures

Participants (129) were junior and senior undergraduates, aged between 18 and 24 years, enrolled at the National United University in Taiwan. The numbers of male (47%) and female (53%) participants were roughly equal and all were familiar with Internet browsing.

Participants were tested while in class, typically in groups of 30-40, during 30-minute testing sessions. All computers in the laboratory were Intel-based Pentiums with identical configurations and the same version of Microsoft IE. The tested homepages were presented on a 17 inch LCD monitor set at a resolution of 1024 × 768 pixels. Upon entering the laboratory, participants were given detailed verbal instructions and were then asked to view and evaluate the two versions of the tested homepages sequentially. In order to control for the potential order effect, the sequence of the two versions was randomized for each participant. After evaluating the website, participants were asked to fill out a questionnaire, which included the MBTI and demographic information. When they had completed the laboratory experiment, participants were thanked for taking part and debriefed.

Measures

Personality type The traditional Chinese version of the MBTI Form M was used to measure participants’ personality types. There is strong support for the reliability and validity of the MBTI (see review by Murray, 1990). In order to decrease respondent fatigue, items related to a person’s S/N and T/F scores only were included. The items were then scored to classify participants’ personality types for each of the two dimensions.

Attitude The attitude toward the website was measured with three 6-point semantic differential questions (good/bad, favorable/unfavorable, and like/ dislike) adopted from Coyle and Thorson (2001). The reliability coefficient for this measure was .84.

Behavioral intention Adopted from Coyle and Thorson (2001), this measure addressed intention to return to the website by asking participants to indicate their agreement or disagreement with the following two 6-point Likert items: (1) It is very likely that I will return to this website; and (2) I will return to this website the next time I want to buy a digital camera. The reliability for this measure was .75.

Results

One-way repeated measures analysis of variance (ANOVA) was conducted to examine the manipulation. The results revealed a significant difference in the expected direction. Participants exposed to the high-density homepages (M = 4.33) were significantly more likely to agree that the homepages were crowded with information than were those who had been exposed to the moderate-density homepages (M = 2.91, F = 132.96, p < .001).

A repeated measures multivariate analysis of variance (MANOVA) was conducted to examine the effects of homepage screen density on website evaluations. This analysis revealed a significant main effect for homepage screen density (F = 25.99, p < .001, Wilks’s λ = .71) across two dependent variables: website attitude and behavioral intention. Both the mean levels of website attitudes and behavioral intentions for moderate-density homepages (M = 4.15; M = 4.48, respectively) were significantly higher than for the high-density homepages (M = 3.25, F = 52.28, p < .001; M = 3.77, F = 37.13, p < .001, respectively). This suggests that most participants preferred websites with a moderate-density homepage.

A series of MANOVAs were then conducted to test the three hypotheses. In Hypothesis 1 we predicted that the levels of homepage screen density would have a greater influence on website evaluations for intuiting types than for sensing types. As anticipated, the 2 (density: high vs. moderate) × 2 (personality type: sensing vs. intuiting) repeated measures MANOVA revealed a significant interaction effect (F = 3.55, p < .05, Wilks’s λ = .95) across two dependent variables: website attitude (F = 6.91, p < .01, partial η² = .05), and behavioral intention (F = 5.88, p < .05, partial η² = .04). The main effect for S/N personality type was not significant (F = .45, p > .6). As shown in Figure 1, there were substantially greater differences in attitude and intention between high- and moderate-density for intuiting types than for sensing types. This suggests that levels of homepage screen density significantly affect the website evaluations of intuiting types, but it does not have the same substantial effect for sensing types. Therefore, Hypothesis 1 was supported.

To test Hypothesis 2, a 2 (density: high vs. moderate) × 2 (personality type: thinking vs. feeling) repeated measures MANOVA was conducted. The main effect for T/F personality type did not reach significance (F = 1.73, p > .1). As expected, the results indicated a significant interaction effect (F = 4.42, p < .05, Wilks’s λ = .93) for website attitude (F = 6.88, p < .01, partial η² = .05), but not for behavioral intention (F = 1.79, p > .18). Figure 2 indicates a substantially larger difference in website attitude between high- and moderate- density for feeling types than for thinking types. It also suggests that levels of homepage screen density significantly affect the website attitudes of feeling types, but it does not have the same substantial effect for thinking types. Thus, Hypothesis 2 was partially supported.

In Hypothesis 3 we predicted that the effect of homepage screen density on website evaluations could vary according to users’ cognitive styles. As expected, the 2 (density: high vs. moderate) × 4 (cognitive style: ST, SF, NT, and NF) repeated measures MANOVA revealed a significant interaction effect (F = 2.93, p < .01, Wilks’s λ = .87) across two dependent variables: website attitude (F = 5.18, p < .01, partial η² = .30), and, to a lesser degree, behavioral intention (F = 2.42, p = .07, partial η² = .23). The main effect for cognitive style was not significant (F = .97, p > .44). Table 1 displays the means and standard deviations across each treatment and dependent variable. Our findings suggest that the effect of homepage screen density on website evaluation differs according to users’ cognitive style. Thus, Hypothesis 3 was supported.

Discussion

In this study we investigated the influence of screen density of commercial homepages on users’ attitudes and behavioral intentions. In our study we supported the assertion that the density of the homepage screen does influence users’ attitudes and behavioral intentions toward a website. A controlled laboratory experiment was conducted to examine how different levels of homepage screen density influence users’ website evaluations. The findings indicate that participants prefer websites with a moderate-density homepage. This suggests that in order to achieve a more favorable website evaluation by users, designers should avoid cramming too much information onto a homepage of very limited size. Thus, website designers should carefully consider whether they need to limit the amount of detailed information or the number of advertisements displayed on the homepages.

We also examined the important role of personality in website evaluations. The findings offer additional support for the concept that Jung’s personality type theory is useful in understanding and predicting how users may evaluate websites. As posited, the difference between moderate- and high-density homepages exerted less influence on sensing types and thinking types than it did on intuiting and feeling types. In addition, the effect of homepage screen density on website evaluations varied according to participants’ cognitive styles. These findings corroborated the concept that there are individual differences in acquiring and processing information (Jung, 1923). The implication for website designers is that although a moderate-density homepage design may generally be a prerequisite for positive website evaluations by users, the influence of individual difference in personality types must also be considered. According to our findings, the sensing, feeling, and sensing-thinking (ST) types are less sensitive to the effects of homepage screen density. Thus, a high-density homepage can still be useful when targeting these types of users.

Although the findings from this study yield insights into the moderating role of Jung’s personality types on website evaluations, several limitations should be noted when interpreting these results. One limitation of the current research is that only two levels (moderate and high) of screen density were examined. Our research could be widened by investigating the investigation of the influence of different levels of homepage screen density on website evaluations. Another potential limitation of the experiment is that we examined only one type of website. However, generalizing beyond this type of website may be problematic. For example, people with high product involvement might follow the central route and process the product information on the website differently (Elliott & Speck, 2005). Therefore, we recommend that in future studies the moderating effects of product involvement be considered in order to arrive at a better understanding of the impacts of homepage screen density on website evaluation. Finally, only one design feature and Jung’s personality types were investigated. We suggest that future researchers consider the effects of other homepage design features, such as navigation structure, reading manipulation, and presentation style. In addition, other personality traits, such as tolerance of ambiguity (Frenkel-Brunswik, 1948) and need for closure (Kruglanski, 1980), that have been found to influence the manner in which people process information and form attitudes need to be explored in the future.

Abrahamian, E., Weinberg, J., Grady, M., & Stanton, C. M. (2004). The effect of personality-aware computer-human interfaces on learning. Journal of Universal Computer Science, 10(1), 27-37.

Benbasat, I., & Taylor, R. N. (1978). The impact of cognitive styles on information systems design. MIS Quarterly, 2(2), 43-54.

Carlyn, M. (1977). An assessment of the Myers-Briggs Type Indicator. Journal of Personality Assessment, 41(5), 461-473.

Coll, R., & Wingertsman, J. C. (1990). The effect of screen complexity on user preference and performance. International Journal of Human-Computer Interaction, 2(3), 255-265.

Coyle, J. R., & Thorson, E. (2001). The effects of progressive levels of interactivity and vividness in web marketing sites. Journal of Advertising, 30(3), 65-77.

Elliott, M. T., & Speck, P. S. (2005). Factors that affect attitude toward a retail web site. Journal of Marketing Theory and Practice, 13(1), 40-51.

Frenkel-Brunswik, E. (1948). Tolerance toward ambiguity as a personality variable. American Psychologist, 3, 268.

Gould, S. J. (1991). Jungian analysis and psychological types: An interpretive approach to consumer choice behavior. Advances in Consumer Research, 18, 743-748.

Haley, U. C. V., & Stumpf, S. A. (1989). Cognitive traits in strategic decision-making: Linking theories of personalities and cognitions. Journal of Management Studies, 26(5), 477-497.

Huang, J. H., & Chang, C. C. (2008). The role of personality traits in online consumer complaint behavior and service recovery expectation. Social Behavior and Personality: An international journal, 36(9), 1223-1232.

Jahng, J. J., Jain, H., & Ramamurthy, K. (2002). Personality traits and effectiveness of presentation of product information in e-business systems. European Journal of Information Systems, 11(3), 181-195.

Jung, C. G. (1923). Psychological types. New York: Pantheon Books.

Kruglanski, A. W. (1980). Lay epistemo-logic-process and contents: Another look at attribution theory. Psychological Review, 87(1), 70-87.

Messick, S. (1976). Personality consistencies in cognition and creativity. In S. Messick (Ed.), Individuality in learning (pp. 4-22). San Francisco: Jossey Bass.

Mitroff, I. I. (1981). Is a periodic table of the elements for organization behavior possible? Human Systems Management, 2(3), 168-176.

Murray, J. B. (1990). Review of research on the Myers-Briggs Type Indicator. Perceptual and Motor Skills, 70, 1187-1202.

Myers, I. B. (1987). Introduction to type: A description of the theory and applications of the Myers- Briggs Type Indicator (4th ed.). Palo Alto, CA: Consulting Psychologists Press.

Myers, I. B., & McCaulley, M. H. (1985). Manual: A guide to the development and use of the Myers-Briggs Type Indicator. Palo Alto, CA: Consulting Psychologists Press.

Nielsen, J., & Tahir, M. (2002). Homepage usability: 50 websites deconstructed. Indianapolis, IN: New Riders.

Nutt, P. C. (1993). Flexible decision styles and the choices of top executives. Journal of Management Studies, 30(5), 696-721.

Rau, P. P., Gao, Q., & Liu, J. (2007). The effect of rich web portal design and floating animations on visual search. International Journal of Human-Computer Interaction, 22(3), 195-216.

Ross, S. M., Morrison, G. R., & O’Dell, J. K. (1988). Obtaining more out of less text in CBI: Effects of varied text density levels as a function of learner characteristics and control strategy. Educational Communication and Technology: EC & TJ, 36(3), 131-142.

Shneiderman, B. (1980). Software psychology: Human factors in computer and information systems. Cambridge, MA: Winthrop.

Singh, S. N., & Dalal, N. P. (1999). Web home pages as advertisements. Communications of the ACM, 42(8), 91-99.

Spool, J. M., Schroeder, W., Scanlon, T., & Snyder, C. (1998). Web sites that work: Designing with your eyes open [Summary]. Proceedings of CHI’ 1998 ACM Conference on Human Factors in Computing Systems, 147-148.

Staggers, N. (1993). Impact of screen density on clinical nurses’ computer task performance and subjective screen satisfaction. International Journal of Man-Machine Studies, 39(5), 775-792.

Wheeler, P. R., Hunton, J. E., & Bryant, S. M. (2004). Accounting information systems research opportunities using personality type theory and the Myers-Briggs Type Indicator. Journal of Information Systems, 18(1), 1-19.

Williams, T. R. (2000). Guidelines for designing and evaluating the display of information on the Web. Technical Communication, 47(3), 383-396.

Appreciation is due to reviewers including

Ana Jiminez-Zarco

Open University of Catalonia

Barcelona

Spain

Email

[email protected]

Tzong-Ke Yang, Senior Lecturer, Department of Information Management, National United University, No. 1 Lienda Road, Miaoli City 36003, Taiwan, ROC. Phone: +886-37-381823; Fax: +886-37-330776; Email: [email protected]