From Elon Journal of Undergraduate Research in Communications VOL. 5 NO. 2
Who You Are Affects What You Buy: The Influence of Consumer Identity on Brand Preferences
Participants included 65 people, 17 males and 48 females, who were recruited through online posts to social media sites and through email. They were asked to complete the short survey by clicking the link in the message. Participants consisted of students from a southern liberal arts university as well as a graduate student from another institution, and non-students familiar to the researcher. There were five freshmen (7%); eight sophomores (12.3%); nine juniors (13.8%); 34 seniors (52.3%); one graduate student (1.5%); and eight non-students (11.6%). There were 43 whites (66.2%); 11 blacks (16.9%); one Asian (1.5%), seven Hispanics (10.8%) and three who did not indicate their race (4.6%). Compensation for completion of the questionnaire was not provided.
A total of 18 items were adapted from previous studies so that they can operationalize self-image congruity, social identity and brand preference. Refer to Appendix A, especially questions 4, 5 and 6, for a complete list of the items used in the study.
The first question asked participants to choose one of the five brands offered and then use that brand to answer the following questions. The list of brands included Lululemon, Heineken, Mac, Pampers and TOMS shoes. The researcher arbitrarily chose these brands to reflect the different personalities and traits the sample may embody. Most of these brands were prevalent on the college campus as well, as most students consumed these brands or were aware of them, with the exception of Pampers. Pampers was chosen to test a brand that was typically inconsistent with participants’ social identities that influenced brand preference. Furthermore, the researchers coded each brand with five to seven adjectives, which were devised from brand positioning statements and consumer profiles, as shown in Appendix B. For example, a few characteristics that described Lululemon included “Athletic,” “Healthy,” and “Relaxing.” Participants were shown the adjectives listed in Appendix B in questions 2 and 3, to prime them into thinking about the brand’s personality traits and perceived images. These questions were also used to bring awareness to participants’ traits associated with themselves and the brand, transferring the brand’s perceived image from their subconscious to their consciousness. This allowed them to consider these adjectives when completing the rest of the survey, which focused on self-image congruity, social identity and brand preference.
Self-image congruity was assessed using items that asked about participants’ brand consumption and their self-image. A set of six items asked participants to indicate the degree to which they agreed or disagreed with statements regarding their self-image congruity with the brand. Several of these items were adapted from Hohenstein, Sirgy, Herrmann and Heitmann (2007) or developed using similar guidelines. Using a Likerttype scale, participants rated each question on the scale of strongly disagree, disagree, neither agree nor disagree, agree or strongly agree. These items can be found in question 4 under Appendix A.
Brand preference was operationalized with a set of six items also adapted from the work of Hohenstein et al. (2007). Participants also rated the degree to which they agreed/disagreed with each statement using the same rank-order assessment. The response options ranged from strongly disagree to strongly agree. These items can be found in question 5 under Appendix A.
Finally, six items were used to measure social identity, of which the same rank-order used in the previous questions was used for this question. Response options ranged from strongly disagree to strongly agree. The operationalized social identity items can be found in question 6 under Appendix A.
Question 7 asked individuals to indicate how familiar they were with the particular brand they chose. Response options were very familiar, familiar or never heard of it. This was used to gauge how well participants were aware of the brand’s images, and whether this influenced the relationship between the self-concept and brand preference.
Demographic information was gathered about their gender, grade level, and racial origin. These questions were optional; however, all respondents who completed the questionnaire offered their demographic information.
The survey questionnaire on Survey Monkey was tested on a few participants for comments and edits before final release. Once revisions were completed, the link to the survey was sent out via social media websites (Facebook and Twitter) as well as through email. Participants in a convenience sample completed the survey. The survey took no longer than five minutes to complete.
Among 69 participants who began the survey, only 65 completed it properly. Each item under Question 4 measuring self-image congruity was assigned a numerical value, for example, 1 for strongly disagree; 2 for disagree; 3 for Neither A/D; 4 for agree; and 5 for strongly agree. All item scores were added to create a total score for this variable of self-image congruity. All items under Question 5 were treated in the same way to calculate a total score of brand preference. All items under Question 6 were used to calculate a total score of social identity. The brand preference total score was used as the dependent variable in all analyses, whereas the self-image congruity score and social identity score served as the independent variables.
Descriptive statistics regarding brand name were computed to determine the percentage-breakdown for each of the five brands listed. Of the 65 participants, eight chose Lululemon (12.3%); 11 for Heineken (16.9%); 27 for Mac (41.5%); three for Pampers (4.6%); and 16 for TOMS Shoes (24.6%). The majority of the respondents were either very familiar (n = 38, 58.5%) or familiar (n = 27, 41.5%) with the brand they chose. Of the brands provided, the majority of males chose Heineken (n = 9, 52.9%) and Mac (n = 6, 35.3%), while the majority of females chose Mac (n = 21, 43.8%), TOMS Shoes (n = 15, 31.3%) and Lululemon (n = 7, 14.6%).
A simple linear regression was conducted to predict the extent to which the self-image congruity score predicted the brand preference score among all participants. Similar to previous research, results found that self-image congruity strongly predicted brand preference, F(1, 63) = 20.74, p < .05, R2 = .24, β = .50.To further examine this relationship, supplementary linear regressions were performed to examine how sex and racial group influenced the relationship between the self-image congruity and brand preference.
To examine whether race has any impact, participants were divided into two groups. Those who identified as “Black,” “Hispanic,” “Asian,” “American Indian,” or “Other” were classified as one group called “Minority.” To examine the extent to which the Minority group accounted for the self-image congruity and brand preference relationship, a linear regression was conducted. Results indicated a statistically significant relationship between the two variables among the minority group: F (1, 20) = 18.54, p < .05, R2 = .46, β = .69. A second linear regression analysis examining those who identified as white also yielded a significant relationship, suggesting that this subset also accounts for a statistically significant relationship between self-image congruity and brand preference as well, F (1, 41) = 6.04, p < .05, R2 = .13, β = .36.
Another set of regression analyses was conducted to examine differences between males and females’ influences on the self-image congruity and brand preference relationship. Results found for males: F (1, 15) = 6.16, p < .05, R2 = .24, β = .54, and females: F (1, 46) = 8.14, p < .05, R2 = .13, β = .39. This suggested a statistically significant relationship between the self-image congruity and brand preference relationship for males and females each.
To test the second hypothesis, which predicted that consumers would prefer brands consonant to their in-group associations, a linear regression analysis was conducted. Results indicated that social identity significantly predicted brand preference, F (1,63) = 13.79, p < .05, R2 = .17, β = .42, indicating that those with higher social identity scores also had higher brand preference scores. To further examine this association when accounting for racial and sex differences, supplementary simple regressions were run on subgroups, such as whites, minorities, males and females.
Scholars have mentioned the disparity between social identity influences on white and minority individuals, suggesting that black, Hispanic and Asian individuals tend to be more influenced by their in-group than white individuals and their in-group (Mihalcea & Catoiu, 2008). This current study sought to find support for this claim by conducting a second linear regression analysis by separating the Minority group (n = 22). Results indicated that this subgroup shows a statistically significant relationship between social identity and brand preference F (1, 20) = 13.34, p < .05, R2 = .37, β = .63. A linear regression analysis on whites, however, did not yield significant results, F (1,41) = 3.02, p > .05.
To examine whether sex made a difference in the relationship between the social identity and brand preference, a linear regression analysis was run on males (n = 17) and females (n = 48) separately. A statistically significant relationship was found between these two variables among males: F (1, 15) = 6.69, p < .05, R2 = .26, β = .56, but not females: F (1, 46) = 3.39, p>.05.Continued on Next Page »