Comparing College Students' Value-, Outcome-, and Impression-Relevant Involvement in Health-Related Issues

Abstract

The purpose of this study was to examine the role of receiver involvement in the context of health communication. Students (N = 277) completed Cho and Boster's (2005) measures of value-, outcome-, and impression-relevant involvement across 6 health behaviors, including cigarette smoking, organ and tissue donation, sunscreen use, alcohol use, sexually transmitted disease testing, and nutrition. Confirmatory factor analyses across all 6 health topics provided evidence of the 3-factor structure conceptualized by Johnson and Eagly (1989) and measured by Cho and Boster (2005). When health behaviors were regressed onto value-, outcome-, and impression-relevant involvement, outcome- and value-involvement, generally speaking, emerged as significant predictors. Results and implications of considering health campaign audience members' levels of involvement are discussed in the domain of preventive medicine.

Notes

¹Statistical analyses were performed to distinguish whether attendance in a particular group made a significant difference for participants when answering the survey items. A multivariate analysis of variance was run for each of the six health behaviors, with group membership as the independent variable and the three types of involvement as the dependent variables. For five of the six health behaviors, group size was not found to significantly affect the three types of involvement measures. However, for cigarette smoking, a small, but significant difference between groups was found for only the outcome-relevant involvement measure, F(26,270) = 162.55, p = .019. However, because it was found that some groups had either a majority of either smokers or nonsmokers and this group size significance occurred only for one type of involvement within one specific health behavior, the finding is considered an anomaly within the data. As such, it would appear best to view this as an unintended discovery that is doubtful to be replicated in future studies.

²Neither the order in which the six topics were presented nor the order of the questions for each topic were counterbalanced on the survey questionnaires. This opens the door to potential order effects. Participants may have been influenced by the order in which the topics were presented, and future research using these scales should balance the topics and questions to ensure that does not happen.

³An in-depth examination of the missing values did not indicate that participants experienced any fatigue. For the 119 questions asked of 277 participants, there were only a total of 56 missing values. Fourteen of the total missing values were from one participant who skipped an entire page of the questionnaire and was subsequently removed from that specific topic's analyses. For the remaining 42 missing values, 31 of those values occurred sporadically within the first half of the questions and the last 11 occurred in the final half of the questionnaire, indicating that subjects did not experience levels of tedium that affected their later responses.

⁴Impression-relevant involvement scale items: “Talking about my beliefs concerning [topic] has little effect on what others think of me”; “If I express the right kind of opinion on [topic], people will find me more attractive”; “People may judge me on the basis of the opinion that I express in public about [topic]”; “The kind of opinion that I express in public about [topic] has little effect on what others think of me”; “The impressions that others have of me are very much affected when I talk with them about my position on the topic of [topic].” Outcome-relevant involvement scale items: “The decision to [topic] has little impact on my life”; “My quality of life would not change depending on the decision to [topic]”; “My well-being has little to do with the decision to [topic]”; “It is difficult for me to think of ways [topic] impacts my life”; “Making the decision to [topic] affects my daily life”; “It is easy for me to think of ways [topic] influences my well-being”; “My life would be changed by my decision to [topic]”; “All in all, the effect of [topic] on my life is little.” Value-relevant involvement scale items: “The values that are the most important to me determine my decision to [topic]”; “My position on [topic] reflects who I am”; “My beliefs about how I should live my life determine my position on [topic]”; “My decision to [topic] is based on the core principles that guide my life”; “My decision to [topic] is based on the values with which I try to conduct my life”; “Knowing my position on [topic] is central to understanding the kind of person I am.”

⁵Reported factor loadings for items in italics resulted from the original analysis in which all scale items were included. All other loadings were computed in a subsequent confirmatory factor analysis wherein italicized items were removed.

⁶Removal of these items also improved the fit of the data to the hypothesized three-factor model. Factors were considered salient only if the corresponding eigenvalue exceeded 1.0 (Thompson, 2004). For STD testing, sunscreen use, and cigarette smoking/tobacco use, four factors initially emerged. On removal of the two negatively worded impression-relevant involvement items, three-factor solutions were confirmed. Five factors were originally identified in the scales for organ and tissue donation, alcohol use, and nutrition. After removing the two impression items, four-factor solutions emerged for all three scales. The fourth factor for both organ donation (λ = 1.04) and nutrition (λ = 1.05) just exceeded the criteria of 1.0. One item on the outcome-relevant involvement scale for alcohol use failed to load significantly on any factor: “It is easy for me to think of ways the legalization of alcohol influences my well-being.” This item was removed, resulting in a three-factor solution for alcohol use. The χ² goodness-of-fit statistic was significant for all analyses (The χ² goodness-of-fit test statistic tests the null hypothesis that any differences between the proposed model, that is, factor structure, and the data are due to sampling error.)

⁷The authors readily admit some domain ignorance in four of the six health topics studied. The six health topics were chosen due to their current status in the health communication literature and the potential utility of the knowledge gained by learning which type(s) of involvement are predictive of students' health outcomes. Related to the former point, consider that the most recent issue of Health Communication (2006, Vol. 19, No. 1) included articles on alcohol (Yanovitzky, Steward, & Lederman, 2006), smoking (Smith & Wakefield, 2006), STDs (Rouner & Lindsey, 2006), and nutrition (Williams-Piehota, Pizarro, Silvera, Mowad, & Salovey, 2006).