Chapter 5. Discussion
5.2 Limitations and Future Suggestions
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
5.2 Limitations and Future Suggestions
Though this study provided empirical findings, some limitations need to be taken into consideration. First of all, this study found the association between heterogeneous information exposure and attitude certainty towards controversial social issues. The finding echoes the viewpoint from the past studies that being exposed to diverse information will facilitate attitude formation and make individuals more clear about their attitudes (Lee et al., 2014; Lee & Myer, 2016). However, based on the literature review, it was equally likely that attitude certainty would lead to individuals’ selective exposure (Knobloch-Westerwick & Meng, 2009). Resonated with the confirmation bias, selective exposure was evidenced to occur for stronger attitudes than weaker attitudes.
People with certain attitudes prefer to receive information that supports their viewpoints and are resistant to changes (Brannon, Tagler & Eagly, 2007). On the other hand, empirical findings infer that constantly avoiding dissimilar viewpoints is rare (Garrett, 2006). Some individuals actively seek out heterogeneous information when they are ambiguous about their attitude (Knobloch-Westerwick & Meng, 2009). In an online behavioral investigation, even if participants hold certain attitudes on scientific issues, they still check out attitude inconsistent information (Jang, 2014). Thereby, the causal relationship between heterogeneous information exposure and attitude certainty could not be guaranteed.
The result illustrated that Taiwanese citizens who have been confronted with attitude inconsistent information could make their attitude towards controversial social issues clearer. However, the current study lacked an understanding of participants’ pre-existing attitudes, it is still unclear whether individuals’ attitude certainty was strengthened or changed. Since individual can be ambivalent about an attitude object,
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
yet hold this attitude with high certainty simultaneously (Ryffel et al., 2014), it is not adequate to interpret the exact variation of attitude certainty. This study could not elaborate on whether attitude certainty increased by strengthening the pre-existing attitudes, altering different attitudes, or making the ambiguous attitudes more certain.
Researchers should pay more attention to the potential impact of individuals’ pre-established attitudes in future studies. Additionally, researchers should consider measuring heterogeneous information exposure of a topic with not only incongruent contents but relatively neutral messages. The exposure time to issues-relevant information should be considered as well. In this way the researchers can explore whether information receivers still prefer attitude-consistent information or they will be more intrigued by relatively neutral viewpoints (Garrett, 2006).
Second, attitude formation is a complex process containing different influencers, and attitude certainty is a secondary cognition attributed to multiple antecedents. For instance, attitude certainty is influenced by individuals’ direct and indirect experiences with an attitude object (Fazio & Zanna, 1978; Wu & Shaffer, 1987), as well as the ability to resist persuasion (Petrocelli, Tormala & Rucker, 2007). In this study, attitude certainty was conceptualized as whether individuals were able to develop certain attitudes toward the selected issues. Unlike measuring to what extent people are attached to their primary attitudes (Rucker & Petty, 2004; Tormala & Petty, 2002), different operationalization might lead to distinct consequences. For example, people who hold a certain attitude on one issue might not have the same certainty on the other topic. In this way, it might influence the overall attitude certainty. Despite the concerns on the correlation between different topics and questions in combing the four items, this study considered them as index to examine how participants are able to develop certain attitudes on controversial social issues.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
In the current study, selective avoidance was measured by unfriending and hiding unwanted dissonant information and perspective on Facebook. However, scholars argued that, hiding disagreements was less emotionally involved than unfriending, and it achieved selective avoidance without terminating the social ties (Zhu, Skoric & Shen, 2017). Although hiding content will facilitate the social media algorithm to reduce unwanted information, it does not completely eliminate future exposure to dissimilar views. Given the increase of interpersonal connection on SNSs, future studies are encouraged to explore more societal and psychological factors. Moreover, survey combing with computationally tracking digital traces of the behaviors over time is also suggested.
Furthermore, this research took Facebook as the analytical context, one of the most used SNS in Taiwan. SNSs such as Facebook (98.9%), Instagram (38.8%) and Twitter (5.6%) featuring various functions are the three most commonly used social media in Taiwan (TWNIC, 2019) with different target audiences (Gil de Zúñiga et al., 2012). Facebook features not only opinion expression, but information seeking, discussions and entertainments; while Instagram is more image-based. By joining Facebook groups, users have higher opportunities to encounter individuals who are not in their contacts and have dissimilar viewpoints. On the contrary, people are prone to see the information they subjectively follow on Instagram and interact with others sharing mostly similarities. On the other hand, it has been perceived that information receiving from SNSs were distinct from other informational channels. Because SNS users are exposed to information conveyed by whom they relate to, the information tends to be more effective and consistent (Gil de Zúñiga et al., 2012). Nevertheless, to avoid context collapse on Facebook, an increasing number of users have been shifted their information seeking, opinion expression, and discussion to Instagram (Parmelee
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
& Roman, 2020). Since heterogeneous information exposure under different social media platforms varies from target audiences, gender distribution, and perspectives, it might affect the outcome. Therefore, the findings from this study should not be generalized to other social media platforms. Whereas future studies are suggested to investigate whether other SNSs contribute to increasing attitude certainty on social issues.
Finally, in terms of the analysis measurements, the data were derived from a survey conducted by the TIGCR. Due to the restriction on the overall length of the survey, items could not comprise the measuring concepts comprehensively. Hence, the use of single-item measurements could be viewed as a shortcoming. Variables measured by a single item included heterogeneous information exposure and heterogeneous discussion. Although several past studies had adopted the single-item measures, researchers have suggested that the variables should be examined more in-depth (Knobloch-Westerwick & Meng, 2009; Lee et al., 2014). Furthermore, this study relied on self-reported data reflecting respondents’ subjective perspectives might lead to a gap between individual perceptions and societal conditions (Scheufele et al., 2016).