Abstract
Responding to recent calls to transcend social media platforms when examining media effects, and using the social information processing model to predict and explain results, this multi-method study first uses a US national survey (N = 325) to examine perceived effectiveness of social support and relational closeness via paralinguistic digital affordances (PDAs; e.g., “Likes,” “+1s,” and “Upvotes”)—the one-click tools for phatic communication—between social media platforms. Results of the survey reveal some significant between-platform differences in perceived effectiveness of social support provided by a PDA, but no significant differences in the relational closeness of ties across platforms. These findings were used to design and conduct focus groups (N = 36) to understand why the identified differences exist. Focus groups reveal that although social support is exchanged across all platforms, different dimensions of social support are sought and received depending on the platform and the network audience that platform accesses. In addition, the focus groups revealed meaningful differences in the nature of network relationships between the platforms, if not the degree of closeness. Taking the two studies together, it seems the adoption and continued use of a platform is an idiosyncratic function of both the social and the technological. Findings underscore the importance of conducting cross-platform studies and demonstrate the value of using PDAs as a convenient cross-platform comparison tool, as they are one of the few common features across social media.
Keywords
Introduction
Research into the processes and effects of social media use has expanded greatly in the past 10 years. However, in that expansion, most studies focus on only one social media platform, leading to potential problems in generalizability (Rains & Brunner, 2015). Additionally, there are indications users engage different platforms for different purposes (Hayes, Carr, & Wohn, 2016), which calls into question the validity of using effects demonstrated on one social media as predictive of another. But what processes or effects may transcend specific media, and which may be unique to a particular platform?
Utilizing the social information processing model (SIP; Fulk, Steinfield, Schmitz, & Power, 1987) to predict and frame results, this multi-method research responds to Rains and Brunner’s (2015) call to understand the differences across and between platforms in one key area of research: social support. Social support is noted as an important reason people engage in social media use (Joinson, 2008; Lin & Lu, 2011), and provides a foundation to explore how interpersonal effects observed on a particular social media platform might, or might not, be generalizable to other platforms.
Study 1—Social Support Across Social Media
Background
Communication in Social Media
Social media are “Internet-based, disentrained, and persistent channels of masspersonal communication facilitating perceptions of interactions among users, deriving value primarily from user-generated content” (Carr & Hayes, 2015, p. 7). Approximately 70% of global Internet users—74% within the United States—actively use social media (Pew Research Internet Project, 2014), with some of the most trafficked platforms being the social network sites (SNSs) and microblogging sites like Facebook, Twitter, QZone, and Google+, and the professional networking site LinkedIn (Statista, 2014). While an increasing amount of research is examining Twitter and other platforms, Facebook receives much of the attention in scholarly research thus far, and few studies compare across social media platforms (Rains & Brunner, 2015).
Analyzing extant social media research, Rains and Brunner (2015) found that two-thirds of studies were limited to a single platform, and Facebook was that platform in more than three-quarters of the studies. Rains and Brunner argued this is problematic for multiple reasons, the most salient being conclusions reached from a body of scholarship dominated by one platform seriously weakens our capability to fully understand the uses and effects of social media across platforms. Assuming effects seen on one platform are generalizable to another is, in itself, problematic. While social media platforms share common characteristics, each platform has unique tools and affordances, and there are indications that users see differences in the uses of and gratifications gained from each (Hayes et al., 2016; Smock, Ellison, Lampe, & Wohn, 2011).
Even early scholarship into these then-emergent media suggests social platforms are not used uniformly, resulting in multiple sources of interactions and disparate effects. Several studies comparing Facebook to MySpace use indicated the total time spent on the sites and number of logins varied significantly (Hampton, Goulet, Rainie, & Purcell, 2011; Kujath, 2011), perceptions of truthfulness of disclosures differed (Dwyer, Hiltz, & Passerini, 2007), and the primary function of the sites for users varied (Jansen, Sobel, & Cook, 2011). Similarly, the generalizability of data gleaned from Facebook users to users of other platforms (e.g., Google+, Twitter, Cyworld) has been called into question (Hargittai, 2007; Rains & Brunner, 2015; Vasalou, Joinson, & Courvoisier, 2010). While there have been multiple indications that uses and effects might be platform-specific, to date, few studies have specifically compared interpersonal effects and processes across platforms. Hayes et al. (2016) found users may choose different social media platforms to achieve different gratifications, and derive different value from interaction experienced on each. Bayer, Ellison, Schoenebeck, and Falk (2015) found interactions on Snapchat were associated with lower social support than Twitter or Facebook. Despite the empirical documentation of such differences, these studies lack a theoretical explanation for why differences exist, and specifically what would account for differences in social support—or its correlate, relational closeness—across platforms. To take the first steps toward more carefully examining differences between platforms, we use the SIP model to help understand how users make sense of cues in meditated interactions, and how similar cues in various media may lead to different outcomes.
The Social Information Processing Model
The SIP model (Fulk et al., 1987) was posited to explain sources of influence in organizational adoption of communication technologies, but it has proven utile in addressing individual adoption and uses of technologies as well (e.g., Schmitz & Fulk, 1991; Treem, Dailey, Pierce, & Leonardi, 2015; Wang, Meister, & Gray, 2013). Fulk and colleagues (1987) proposed the adoption of technologies could be explained by examining the social environment of the user or organization, and use of a given technology could be explained by four factors, both subjective and objective: (a) the objective characteristics of the media, (b) past experience with and knowledge of the media, (c) individual differences and preferences, and (d) social information provided by network ties. Decisions about adoption and use of a particular media are determined by a complex combination of these elements, and we posit decisions about which social media to use in a particular circumstance (whether seeking social support or for something more objective, such as seeking professional guidance) is also determined by these factors. Thus, SIP guides us to expect differences in usage, expectations, and interpretations of social media actions (including communication) across platforms.
Social Support
Communicative in nature, social support is help, love, or care provided through interpersonal relationships and the interactions these relationships entail (Albrecht & Goldsmith, 2003; Cobb, 1976). Received support, or the “perception of how much supportive behavior has been has been received from social network members in the recent past” (MacGeorge, Feng, & Burleson, 2011, p. 321) via social media has seen increased focus in scholarship (Rains, Peterson, & Wright, 2015). Social media allow individuals to specifically seek social support and then to interact masspersonally with both the strong and weak ties within their social networks (Carr & Hayes, 2015).
Social Support via Social Media
Relational ties can provide social support via social media in various ways, including supportive behaviors (e.g., clicks and shares) and verbal cues (e.g., comments and direct messages) (Blight, Jagiello, & Ruppel, 2015; Hayes et al., 2016; Krämer, Rösner, Eimler, Winter, & Neubaum, 2014; Rozzell et al., 2014). When an individual receives these supportive behaviors and messages from someone in response to a status or post seeking social support, these messages are evaluated in respect to the situation (Blight et al., 2015). The evaluation of these messages can impact perceptions of the situation effective in reducing rates of depression and stress (Wright, 2012; Wright & Bell, 2003) and loneliness (Shaw & Gant, 2002), and can increase quality of life (Leung & Lee, 2005). While much work has been done on Facebook regarding social support, less is known about the ability of other platforms to facilitate its processes, leaving a void in the literature regarding both the generalizability of the work done on social media and social support, and the differences among the platforms in their provision. Moreover, research on social media and social support has not looked into the role specific features play in the provision, receipt, and perceived effectiveness of social support. Since social media are collections of multiple features—the combination of which varies by platform—it is important to take a more granular examination of the effects of a particular feature, such as the now-ubiquitous Likes and +1’s of social sites.
Paralinguistic digital affordances (PDAs) are “cues in social media that facilitate communication and interaction without specific language associated with their messages” (Hayes et al., 2016), and use “affordances” as features, consistent with Gaver’s (1991) technologist perspective. 1 They only require a single click and are delivered in response to social media content. PDAs reflect one of the few commonalities across social media platforms. Typified by Likes (Instagram, Facebook, Snapchat, and Twitter’s renamed Favorites), +1s (Google+), and Upvotes (Reddit and Imgur), PDAs are one of the most commonly used tools of social platforms, and are simple means of communication activated by a click, but their communicative value may be more complex than their simplicity would imply. They have become a form of valuable social currency; so much so that it is newsworthy when a minor Australian celebrity throws a tantrum over receiving too few of them when posting seeking emotional support (Gillard, 2015). Hayes and colleagues (2016) explored the meanings individuals attribute to the receipt of PDAs across several social media platforms, finding individuals often assign meaning to PDAs far beyond their original platform-intended meaning, and can interpret the sending and receiving of them as more or less personally meaningful depending on the circumstances. Particularly notable in their findings was that users can interpret the same PDA in various ways—including for social support—idiosyncratically, considering not only the site and the cue, but also the purpose of the posted message and the nature of the individual whom provided the PDA, suggesting the receiver’s relationship with the sender affects perceptions of the meaning associated of the cue (Carr, Wohn, & Hayes, 2016). While PDAs are not always used or interpreted as a signal of social support, users can perceive social support upon receiving PDAs and even find the support they receive through PDAs beneficial, especially if they are seeking social support through their posted content (Carr et al., 2016; Rozzell et al., 2014; Wohn, Carr, & Hayes, 2016).
Relational Closeness
In addition to differences in perceived effectiveness of social support facilitated by various social media, it is likely support is coming from network members of varying relational closeness, depending on the platform. Relational closeness refers to the strength of a relational tie, or an individual’s perception of the inclusion of another in her or his sense of self (see Aron, Aron, & Smollan, 1992). Often associated with notions of tie strength, interpersonal relationships are frequently distilled into bifurcated notions of strong/weak or close/distant in research, particularly in relation to social support. For example, strong or relationally close ties are often said to provide greater levels of emotional and tangible support while weak or relationally nonclose ties provide greater information support due to their ability to span diverse relational networks (Granovetter, 1973; Putnam, 1995; Rozzell et al., 2014).
Social media have been lauded for their ability to allow users to readily access diverse relational ties, enabling communication with close and nonclose ties alike and across diverse relational contexts (Wang & Wellman, 2010). Though substantive literature addresses the potential for social media (particularly Facebook) to collapse relational contexts and bring together previously disparate relational ties (e.g., Hogan, 2010; Marwick & boyd, 2011), less work has articulated and explored potential differences in relational ties between platforms.
There is strong reason to believe the diversity of one’s relational ties is not homophilous across all social media. By their nature, many social media constrain the social contexts and thus the relational diversity of their users (Donath & boyd, 2004; Litt, 2012). For example, LinkedIn serves as a professional context for job networking and connections, likely limiting relational ties and interactions to weaker or more relationally distant ties such as colleagues and former or potential employers, and without a strong expectation or need for interaction with close friends or family that is a defining characteristic of social media such as Facebook. Given various social media platforms have different user bases, norms of connectivity, and relational contexts in which individuals interact, it may be expected that relational closeness of an individual’s network likewise varies, so some social media facilitate more relationally close ties while other social media facilitate a greater proportion of relationally nonclose ties.
Research Questions
Given social media’s documented facilitation of social support from varied relational ties (Carr et al., 2016; Rozzell et al., 2014; Wohn et al., 2016), and the ubiquity and frequency of use of PDAs across most social media platforms, we propose investigating the differences in perceived effectiveness of support from varied relational ties between platforms by looking at interpretation of PDAs provided by an individual’s network in response to a social media post. Perceived effectiveness of support goes one step beyond perceived social support received, enabling us to determine if the recipient of the social support actually found it to be useful.
RQ1. Are there differences in the perceived effectiveness of the social support perceived by individuals receiving PDAs across different social media platforms?
RQ2. Are there differences in relational closeness of network ties providing PDAs across different social media platforms?
Method—Study 1
Survey and Respondents
As a part of a larger data collection on PDAs, a US national sample of social media users was surveyed to assess their perceived social support from and relational closeness with ties across platforms. The sample (Nrespondents = 325) was recruited using the Qualtrics research firm’s survey pool, and compensated according to their agreement with Qualtrics. Respondents (216 females, 66.5%) had a mean age of 45.81 (standard deviation [SD] = 14.35; range: 18–80) years. Respondents were not asked to disclose their socioeconomic status or ethnicity.
Procedure
From a list of the five most popular social media platforms in the United States at the time of collection, respondents were first asked to identify which they had most recently used. Facebook was reported as the most recently used social medium by 64.6% (n = 210) of the sample, followed by Twitter (n = 30; 9.2%), Instagram (n = 29; 8.9%), Pinterest (n = 28; 8.6%), and LinkedIn (n = 28; 8.6%). Facebook use was disproportionally overrepresented, χ2(4) = 404.37, p < .001; however, after excluding Facebook, the remaining four media were normally distributed, χ2(3) = .10, p = .99. Although Facebook was disproportionately used more often than other social media, this distribution did not affect analysis as (a) data were consistent with reported national social media usage patterns (Pew Research Internet Project, 2014) and (b) sufficient sample sizes were obtained for the four non-Facebook media to enable between-group analyses (see Cohen, 1988).
Based on the social media platform most recently used by the respondent, the survey engine adapted a standard instrument, varying wording to reflect the respondent’s identified medium and its features. Each respondent was asked to describe the last two posts they had made on the indicated platform and to report the total of PDAs that post had received, and then describe the relationship they had with last three unique individuals providing a PDA to the posts. Through this procedure, 975 dyads were identified. After removing duplicated relational ties for each respondent, each identified an average of 4.14 (mode = 6) unique network ties who had provided a PDA for their two posts. Respondents were then asked a series of questions regarding the social support perceived from receipt of the PDA.
Measures
Perceived effectiveness of social support from a PDA was an adaptation of the “perceived effectiveness of social support” scale by Rozzell et al. (2014). This scale was originally designed to measure the perceived effectiveness of social support received through online comments; for the adaptation, we changed “comment” to a reference to the specific PDA of the platform. For example, if the respondent chose Facebook, they were given directions to select the option that best represented how they felt about the use of the “Like” button in reaction to their post. Anchor points included, “Positive/Not Positive,” “Encouraging/Not Encouraging,” “Not Insulting/Insulting,” and “Not Supportive/Supportive,” “Not helpful/Helpful,” “Hurtful/Not Hurtful.” The scale was reliable (Cronbach’s α = .89), and greater means indicated respondents perceived the social support via PDAs received was more effective. Relational closeness was assessed using Aron et al.’s (1992) inclusion of others (IoS) measure. Though a single-item measure, the validity of this measure has been demonstrated in other studies (Rozzell et al., 2014; Wright & Miller, 2010) and consists of seven sets of increasingly overlapping circles representing the self and the relationship with another individual, which higher values indicating greater relational closeness.
Results—Study 1
The research questions posed in Study 1 sought to probe differences in perceived social support from and relational closeness to ties across various social media platforms. Of initial concern before assessing the research questions was the use of multiple data points from individual respondents, potentially violating statistical assumptions of independent observations. Thus, following Bickel’s (2007) recommendation, unconditional intraclass correlations (ρ) were calculated to assess within-group dependence. Ranging from zero to one, ρ-values indicate the variable’s variability is attributable to mostly within-participant within-person variance when close to zero and attributable to mostly between-person or group differences when close to one (Kreft & De Leeuw, 1998). The intraclass correlations for relational closeness (ρ = .65) and perceived effectiveness of social support (ρ = .72) suggested that the majority of variability in these variables were attributable to between-respondent variation, and analysis progressed.
Welch–Statterwaith t-tests were used to determine whether these variables of interest were evenly distributed across platforms. The Welch test is more robust to violations of standard assumptions in parametric statistics than standard analysis of variance (ANOVA) tests, such as unequal cell sizes (Hayes, 2009). As survey items addressed specific interpersonal ties and social support, analyses were conducted for the individual dyads (Ndyad = 975) represented in the data to better represent the nuance and diversity of individual relationships.
Respondents reported that perceived effectiveness of social support via PDAs across all platforms (Msupport = 5.62, SDsupport = 1.24) was significantly greater than the scale midpoint of 4, t(973) = 40.74, p < .001 (2-tailed). The Welch test revealed significant between-platform differences in perceived effectiveness of social support via PDAs, Welch’s ν = 4.60, p = .001, F(4, 969) = 5.33, p < .001 (2-tailed), η2 = .02. Though the test revealed differences in perceived effectiveness of social support based on social medium, it does not directly assess the directionality or specific media that may be driving those differences. Thus, a series of independent sample t-tests (see Table 1) were conducted to determine whether specific between-platform differences existed. Given the multiple comparisons being made, a Bonferroni correction was used, so that the familywise level of significance was set at p < .005. At this level of significance, only Twitter and Facebook were significantly different with regard to perceived effectiveness of social support via PDAs, with respondents perceiving Facebook as having higher levels of social support effectiveness.
Between-Groups Welch t-Test Comparisons of Social Support and Relational Closeness by Social Medium.
Degrees of freedom and t-values are listed for each paired t-test. Differences in degrees of freedom reflect differences in cell sizes, missing data for specific respondents, and reporting of t-test assuming unequal variance following a significant Levene test.
p < .05, 2-tailed. **p < .01, 2-tailed. ***p < .001, 2-tailed.
At the traditional p < .05 level of significance, differences also emerged between Twitter and (a) Instagram and (b) Pinterest, as well as between Facebook and LinkedIn. Thus, it seems PDAs on Twitter are considered the least effective in delivering social support compared to Instagram, Pinterest, and Facebook.
The second research question addressed difference in the relational closeness of ties across social media. Again, Welch t-tests were used due to the robustness of these nonparametric statistics against such a violation, and enabling accounting of the dependence of each relational tie on a specific respondent. Respondents generally reported perceiving relational closeness across all platforms (Mcloseness = 3.96, SDcloseness = 2.13) that was not significantly different than the scale’s midpoint of 4, t(968) = −.53, p = .60 (2-tailed). The second Welch test did not reveal significant between-platform differences in relational closeness, Welch’s ν = .60, p = .66, F(4, 968) = .55, p = .70 (2-tailed), η2 = .002. Again, a series of independent sample t-tests were conducted to determine whether specific between-platform differences existed, revealing no significant differences based on platform (see Table 1).
Finally, because of the analysis of several dyadic relationships with each individual respondent, there was a violation of the assumption of independence which may have confounded results. Therefore, a multilevel analysis was conducted to explain variation in respondents’ reported relational closeness by regressing the individual-level indicator of social media platform used, while controlling for individual-level within-participant responses across their six dyadic responses. The analysis did not reveal significant differences between respondents in relational closeness based on medium, F(4, 318) = .26, p = .90, b = .28, se = .39, p = .47. These results indicated no significant differences in relational closeness of ties among platforms, which was slightly above-average across all media, F(1, 318) = 633.41, p < .001, b = 3.51, se = .37, p < .001 in the multilevel analysis.
Study 2—SIP and Communication Patterns in Social Media
Background
The findings of Study 1 indicate differences in the perceived effectiveness of social support received across various social media; but relatively stable relational closeness of the aggregated ties across the same platforms. Yet prior work has generally supported the notion of a positive relationship between social support and relational closeness (e.g., Blight et al., 2015; Krämer et al., 2014; Rozzell et al., 2014). How or why, then, may perceived effectiveness of social support received via different media vary even as the relational closeness of ties accessed by those channels not differ? One likely explanation is individuals are using the various social media differently, influenced not only by the features of the sites themselves but also by the sites’ users and norms of use to access social support from their relational networks. Thus, we turn again to the SIP model and expand our conceptualization of social support to guide additional research to help explain the results of Study 1.
Dimensions of Social Support
Cutrona and Suhr (1992) distinguished among five dimensions of social support. Emotional support refers to the consolation or comfort provided via expressions of concern or caring. Informational support is the provision of advice, facts, or feedback to help an individual solve a problem. Appraisal or esteem support refers to the aid provided in the form of evaluative feedback to help an individual assess her or his own performances. Instrumental or tangible support is the provision of goods or services to aid another. Finally, network support addresses an individual’s perception of belongingness in a social network of those sharing similar concerns. Studies show people increasingly seek emotional and informational social support online through both message boards and social media (Craig & Johnson, 2011; Rains & Keating, 2011; Wright & Miller, 2010). However, recent research suggests all dimensions of social support are not accessed or provided uniformly via social media (Blight et al., 2015; Krämer et al., 2014). It seems likely that differences in users’ access to or perception of various social media is explained by differences in their adoption and usage of the various platforms, suggesting a model of technology adoption may help explain differences in the dimensions of social support accessed in each.
SIP (Fulk et al., 1987) expects users’ adoption and use of a communication technology (such as social media) is a function of their own subjective perceptions, but also influenced by the features and nature of the medium as well as the perceptions of others about the technology and their uses. Returning to the disparities among the findings of Study 1 and prior literature, the differences may be accounted for by considering some of the subjective and objective differences users perceive between both (a) the various social media themselves and (b) how their diverse relational networks utilize each site. In addition, reflecting Krämer et al.’s (2014) call to not collapse all forms of social support into a unidimensional construct, it may be that while all platforms facilitate some form of social support, they do not facilitate all forms equally. Guided, then, by the dimensions of social support (Cutrona & Suhr, 1992) and by SIP and its theoretical mechanism expecting that functionality and social norms guide use of a communicative technology, we ask a third and fourth research question:
RQ3. Which types of social support are provided by each social media platform?
RQ4. How are the social support provided and relational closeness among social media platforms affected by (a) the site features, (b) users’ past experiences, (c) individual preferences, and (d) others’ perceptions of the medium?
Method—Study 2
Focus Groups
To probe this open question and help understand the findings of Study 1, focus groups were used to facilitate insights about attitudes and behaviors through semi-open-ended discussion. Such discussion can generate insights not possible in a survey methodology, as participants’ responses build on each other (Krueger, 2009). This multi-method procedure gives increased validity and explanatory power to the results by helping to triangulate and further interpret findings from Study 1 (see Frey, Botan, & Kreps, 1999).
Participants
Focus group participants (N = 36, 20 female) were recruited from undergraduate classes at a mid-sized Midwestern university in exchange for extra credit. Three focus groups lasting approximately 60 min were conducted, consisting of 12 participants each. Participants ranged in age from 19 to 24 (M = 21.47, SD = .84) and represented diverse races.
Procedure
Focus groups were held in a campus setting and were audio recorded. A short questionnaire inquiring about social platforms used and preferred, as well as participant demographics, was administered before discussion began. The guiding protocol was constructed using the concepts and variables of social support, relational closeness, and the SIP model, and questions included inquiries related to the five types of social support, the composition of network connections on each social platform, and motivations for use of each platform.
Results—Study 2
RQs 3 and 4 sought explanations for the differences observed in Study 1 by examining what types of social support are provided by each social media platform, and then determining how relational closeness and social support were impacted by site features, users’ past experiences, preferences, and others’ perceptions of the medium. Across all three focus groups, participants reported using multiple social platforms on their intake surveys. Nearly all participants had and used Facebook, Snapchat, Instagram, and Twitter accounts, reflecting the four dominant platforms subsequently discussed most, but the focus groups additionally addressed Tumblr, Pinterest, and LinkedIn. The platforms discussed, particularly Snapchat, were slightly inconsistent with the platforms identified in Study 1, reflective of usage patterns of a national sample versus a college student sample.
To analyze the focus group data, participant responses were coded for themes accessing the five dimensions of social support, and also for the four factors of SIP. As discussion of both the dimensions and factors occurred at various points across the focus groups, the identified themes were then organized into narratives around that particular factor or dimension.
Social Support Between Platforms
Analysis of RQ1 in Study 1 determined differences existed in the effectiveness of social support across platforms. RQ3 in Study 2 sought to understand why those differences exist. Participants consistently agreed social support could be received in general on social media, but introduced (unprompted by the moderator) that seeking it was inappropriate on some platforms and it might depend on the situation and the type of support, emotional, informational, instrumental, or appraisal, sought.
Emotional Support
After the moderator explained emotional support, participants noted the social norms of some platforms allowed for posting in search of social support, as long as that user did not do it too much (The concept of “chronic bitchers” was brought up in all groups. Apparently if a person seeks social support too often on social media, they will experience diminishing returns). Participants discussed sometimes just needing to be heard on a problematic or emotional issue, and that the PDAs they received on posts seeking support would make them feel better. Facebook, Tumblr, and Twitter were specifically mentioned as places where seeking emotional social support was ok, but a couple of participants noted it depended on what you used Twitter for and who your network was, with many commenting they used Twitter for professional networking. Other platforms, especially Snapchat, Instagram and LinkedIn were seen as platforms you would not use for emotional support. As Lucy, 2 22, pointed out, “No one likes a sad Snap or a gloomy Instagram.” The professional nature of LinkedIn also made it inappropriate. While the discussion did include personal examples in which the participants themselves might seek social support, the groups obviously felt more comfortable talking about other social media users rather than themselves, indicating there may be some stigma associated with either seeking emotional support or sharing it on social media. Participants also noted PDAs could be interpreted as form of emotional support—they both sent them and received them in that way, depending on the context of the interaction, in terms of both content and platform, and they were frequently sent as a way to celebrate, or commiserate, with users’ successes or failures.
Informational Support
When asked specifically about what platforms they would use to seek informational support, participants overwhelmingly mentioned Facebook, as that is where their biggest social networks reside. Gillie, 22, noted, “I have friends on Facebook from all over the country, so when I travel, I can put something up about a city and get some good recommendations.” Another participant mentioned Twitter, through the use of hashtag #AskTwitter and then another hashtag related to the inquiry, can be great at finding solutions to problems, as can use of hashtags in general for starting or participating in broader conversations. Other platforms, particularly Snapchat and Instagram, were mentioned as places your friends would just make fun of you, or simply wouldn’t be particularly helpful, in response to an informational inquiry. Pete, 21, mentioned, “Trying to get help from your friends on Snapchat is just asking for fake shit.”
Instrumental Support
Participants chaffed at this line of inquiry, indicating social media would not be the place one would seek either material or non-material instrumental support, unless a private, direct message was involved—something that a public PDA would not be involved in. Loan seeking especially was rejected, and though a couple of participants brought up GoFundMe, the fundraising website frequently linked on social media, multiple participants voiced that they were “sick of those” in response. Participants seemed much more likely to reach out directly to friends if they had an instrumental support request, and mentioned one of the reasons was that it was embarrassing when a public request for help made on social media was ignored.
Appraisal Support
To seek objective feedback to aid in self-evaluation, Snapchat was the overwhelming favorite across groups. While Facebook was mentioned in the discussion of informational social support for the broad and large audience it provides, participants discussed the close and somewhat smaller network they have on Snapchat when considering appraisal support. They noted while they had fewer total connections on Snapchat, those connections were more relevant to their current lives. Phoebe, 21, mentioned, “there are no rando[m] followers on Snapchat, you really know almost everyone…so if I had to ask about something more personal that’s where I’d go.” In addition, Snapchat “Snaps” can be more directly targeted to an even narrower group of friends, allowing for targeted support seeking. Instagram was mentioned as another likely source of appraisal support, but participants mentioned the social norm of Instagram is of positivity, and lots of (sometimes automatic) “Liking,” so they didn’t trust it for objectivity, though many mentioned that the PDAs they receive in response to appraisal seeking posts (e.g., fashion questions) meant something to them.
Network Support
The concept of network support was probed in all three focus groups; but did not emerge as a concept of interest. It may be that, by the nature of being a member of various social media, individuals felt innately connected and belonging or it could be that college students, being young adults and having a unique social status that involves little diversity in networks, were not yet in need of such support. Ultimately, though, network support was not a dimension perceived as useful or explicitly sought or obtained via social media by participants.
Social Support Summary
While all the platforms seem to facilitate both sending and receiving some sort of support, different platforms provide different types of social support, both through PDAs and through other tools, answering RQ3 and helping unpack the significant findings of RQ1.
Relational Closeness
When asked about the closeness of their networks and whether weaker ties were grouped on one platform versus the others, an interesting dynamic emerged. While participants, gave a resounding “Yes!” to the question of whether there were different audiences for their content on the different platforms, and noted Facebook in particular had a lot of weak ties on it (“rando[m]s from high school” was a common theme), those who the participants were relationally close to seemed equally distributed across the platforms, reflecting the lack of significant results to RQ2. Although the relative relational tie diversity across the platforms was stable, the difference here was the actual nature of the relationship with the network member, not the strength of tie. For example, for focus group participants, Facebook is where they connect with close family members, along with a large and diverse network of both strong and weak ties, but Snapchat and Instagram are where their close friends are. As Camie, 19, pointed out, “Everyone is on Facebook, so you have everyone [in your network] there, but then you get just your real friends on Snapchat and Insta and you can be more open.” The relative strength of tie, as measured in Study 1, would be the same between the close friends and close family, but users’ interactions, disclosures, and types of support sought would be different reflecting the different nature of the relationship reflected in each platform’s network. It was consistently noted, however, that close (peer group, strong tie) friends were part of the participants’ network on each platform. The only platforms that seemed to be composed of distinctly weak ties were LinkedIn and Pinterest, but given the professional and creative natures of those platforms respectively, this was not surprising.
Reasons for Adoption and Use
Fulk et al. (1987) discussed four factors explaining adoption of a given technology, including the objective characteristics of the media, past experience with and knowledge of the media, individual differences and preferences, and social information provided by network ties. When asked how and why they adopted a particular platform, and what influenced their continued usage of that platform, a variety of reasons for adoption and continued use emerged, but they did seem to be relatively similar across the platforms.
The objective characteristics of the medium seemed to matter most on the platforms that had distinctly unique purposes—LinkedIn and Pinterest in particular. Facebook, which was ubiquitously used across the focus groups, seems to keep people coming back by facilitating new functions users find valuable, such as photos, groups and messaging, but, paradoxically, was not noted for having particularly valuable technological affordances, even though participants kept mentioning the tools that kept them on the site. Snapchat’s ability to easily target “Snaps” to a narrower group of followers was brought up in each group as a particularly valuable tool, with Fred, 24, pointing out, “you can limit the audience for stuff on Facebook, but no one does. On Snapchat it is kind of expected that you aren’t going to be sending the same Snaps to everyone.”
Past experience with the media did have some impact on participants’ current use—some were used purely out of habit or obligation, especially Facebook, Twitter, and Pinterest. Mitzi, 23, noted she had, “spent so much time building a feed and followers (on Twitter), that I have to keep it up.” Others agreed with her, especially regarding the more visual sites Instagram and Pinterest—those sites needed the most care and maintenance, as “no one will notice if you don’t update Facebook for a few days, but if you don’t upload new pictures to Insta, especially, people will think you’re dead” (Lucy, 22).
Individual preferences only seemed to be a factor for a few people across the groups. Some noted they had stopped using Twitter because it was “driving them crazy” with retweets and sponsored tweets in particular, but beyond that, most of the participants in these groups didn’t identify preferences as a reason to use one or the other, it seemed to be more a herd mentality—all their friends were using or adopting something, and they would eventually get on board, even if they thought the platform was stupid at first.
Social influence seemed to be the strongest reason for adoption for most platforms—when asked what the last new platform they adopted was and why, nearly every participant mentioned Snapchat and the fact that they perceived everyone was on it when they got to college, and they subsequently adopted Snapchat. They adopted Facebook and Instagram in high school for the same reasons, but maintained their presence on each for different reasons (habit and prior experience, respectively). Many adopted LinkedIn because it was recommended by professors or career centers, and others adopted Pinterest because they heard friends talking about finding useful content on it, or people shared “pins” with them and they wanted to see more. Thus, in response to RQ4, it seems an idiosyncratic mix of reasons users both adopt and remain active on various sites, but this process does seem most influenced by the factors of past experience and social ties. Participant Lucy, 22, summed it up well with the statement, “It kinda depends on the site—but these sites are for wasting time and being social, and only sometimes for a real purpose, so we are going to be on the sites that our friends are who don’t annoy us.”
General Discussion
This research responded to the prior call (Rains & Brunner, 2015) to overcome a limitation of extant literature, which has generally collapsed or considered various social media tools as interchangeable. Drawing from mixed-methodological results, findings indicate not all social media are the same, and may facilitate different communicative outcomes (in the present study, social support) based on the perceived utility of the communicative tool and others’ attitudes toward the medium. Findings from both studies have implications for how we should consider social media and the boundaries around studies of particular platforms, as well as perceived effectiveness of social support across social media.
The present findings applied SIP (Fulk et al., 1987) to understand and explain how individuals select among the multitude of available social media platforms to communicate and, specific to the present research, obtain social support. SIP posits four factors (medium characteristics, past experiences, individual preferences, and social influence) leading to an individual adopting or continuing to use a technological medium. In addressing RQ4, Study 2’s focus groups clearly articulated the roles all four of these factors, in varying degrees, play in governing not only how users interact online, but the nature of the networks with whom they interact. Perhaps unsurprisingly given the social nature of social media, social influence was one of the strongest factors impacting adoption and continued use. Focus group participants also indicated certain social media were better-suited for particular dimensions of social support than others (e.g., Twitter as an appropriate and effective channel for seeking information support, but not for instrumental or emotional support).
Relatedly, focus groups noted the homophily of relational networks across social media platforms (i.e., similar proportions of stronger and weaker ties across all media), helping explain the surprising lack of differences in relational closeness across platforms identified in RQ2. Although the relative diversity of relational ties across media is stable, the nature of relational ties and networks varied widely across platforms. Recalling Phoebe’s reaction to appraisal support, some networks may have close, narrower groups of peers with fewer ties representing fewer types of relationships (e.g., friends, family, coworkers, classmates, acquaintances) while other sites might reflect more diverse networks comprised of more relational contexts and types of relationships. However, it seems that overlap in strong peer ties exists across all sites (i.e., they follow their close friends across all platforms, but not other strong ties).
In this way, our findings reinforce and demonstrate prior assertions (e.g., Wellman & Wortley, 1990) that the types of relationships and relational networks a medium facilitates may be more impactful on communicative patterns and effects than the size or diversity of the network, and the manner in which relational closeness is considered is important. Certain media may be more normatively appropriate—either due to the nature and norms of the site, or the composition of relational networks on that platform—than others for achieving communicative goals or relational exchanges, such as the various dimensions of social support. Therefore, the impact of the actual audience (close friends, close family, various weak ties) on a user’s social media activity should not be underestimated. At the same time, it could be the actual audience is incongruent with individuals’ imagined audience (Litt, 2012). Though our study could only assess the latter, future research may want to investigate and differentiate these two constructs.
With respect to social support, our findings indicate social media can facilitate various dimensions of social support through PDAs and other tools, though some platforms may be better-suited than others for provision and accessing of specific types of support. While PDAs are not always meant or perceived as social support, social media users who are purposely seeking specific dimensions of social support perceive the PDAs they receive as being supportive (Carr et al., 2016). As MacGeorge, Feng, & Burleson (2011) explicated social support as the “perception of how much supportive behavior has been received from social network members in the recent past,” at least to these respondents and participants, PDAs can represent an indication of support received from their online social network. Channels where users maintain broader, more disparate networks such as Twitter and Facebook may be better for accessing informational support (consistent with Putnam, 1995), while channels with tighter more closely connected social circles such as Snapchat may be better for accessing appraisal support without the fear of negative of dissonant responses to support-seeking. Moreover, the finding of various channels lending themselves to different dimensions of social support reinforce recent calls (e.g., Krämer et al., 2014) to not collapse social support as a unidimensional construct, and rather to consider how individual media may (or may not) facilitate a specific type of social support. The focus groups echoed findings from Williams (2006), noting instrumental support may be difficult to provide online, not due to relational closeness but rather because the nature of the system itself does not lend itself to provisions of instrumental support—one cannot help you move a couch via WhatsApp.
Limitations and Future Research
This research sought to empirically and qualitatively assess the communicative role of PDAs with respect to social support and relational closeness between platforms. Social support, specifically emotional support, received through social media is a well-researched process, but future work should endeavor to include varied dimensions of social support as well as the specific feature that facilitates the support, as it seems all platforms are not created equal in the affordance of the distinct dimensions of social support. In addition, future research should consider measuring relational closeness in a way that accurately reflects the diverse nature of the relationships, familial, peer, and professional, individuals may have across platforms. Finally, scholars should consider the arguments of Rains and Brunner (2015) and work to not focus specifically on one platform. While work involving only one platform is not inherently bad, it will become increasingly important as users migrate to new platforms to consider differences between media as scholars continue to seek ways to build social media theory.
Some of our calls for future scholarship are manifest in the limitations of the present research. In Study 1, a unidimensional measure of social support was used to assess the effectiveness of social support obtained via PDAs, limiting the granularity with which social support dimensions could be assessed between platforms as well as the support afforded by other features. In our focus groups, participants thought platforms such as Instagram were not appropriate places to seek emotional support; however, lack of evidence should not be interpreted as support for the null. This does not mean that people do not seek or give emotional support on Instagram—it was just not a type of behavior we saw among our participants, or there may have been some social stigma in openly discussing it. In addition, the focus group participants were much more comfortable discussing how others acquire social support through social media, and where much more reticent when prodded to discuss how they specifically receive it. Thus, the semi-public forum that is a focus group may not be the best way to access individual social support sought or received.
It is also important to consider the qualitative work from Study 2 is not meant to be representative of the general population. Rather, it serves as a first attempt to explain the findings from Study 1. The focus group results should be interpreted within the context of its sample (i.e., young adults who are college students in the Midwestern United States). This younger sample may also be why certain applications, such as Snapchat (which did not emerge in the general adult sample as a “most recently used social media”), were commonly discussed in the focus groups.
Conclusion
We conducted two studies investigating the role of social media in affording social support, and found that people feel different types and levels of social support based on the medium. While this seems like an extension of McLuhan’s (1994) old premise that the medium is the message, we found that differences across platforms were less of an artifact of different features or technological affordances, but a more complex combination of hardware (e.g., the system, underlying algorithms) and audience (e.g., network composition, social influence, and social norms). Because individual social media are so different in each of these aspects, our study underscores the importance of conducting cross-platform studies, as specifically called for by Rains and Brunner (2015), and demonstrates the value of using PDAs as a convenient comparison tool, as they are one of the few common features transcending platforms.
Rather than attributing media effects to a particular medium, our findings suggest one should highlight not only the features being examined in the particular social media (Smock et al., 2011), but the network factors and perceived social norms of the platform as well. If social norms, and the social group creating them, are such a strong influence to people’s perceptions and behaviors, current studies are not doing an appropriate job of capturing such norms, which are not only different across platforms, but also different within each social media across time as well as across individuals, depending on their network composition. Returning to the original question posed in this work, interpersonal effects observed on a particular social media platform may not be generalizable to other platforms due to the different audiences users have on each.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.