Even though the extant literature investigates how and why academics use social media, much less is known about academics’ temporal patterns of social media use. This mixed methods study provides a first-of-its-kind investigation into temporal social media use. In particular, we study how academics’ use of Twitter varies over time and examine the reasons why academics temporarily disengage and return to the social media platform. We employ data mining methods to identify a sample of academics on Twitter (n = 3,996) and retrieve the tweets they posted (n = 9,025,127). We analyze quantitative data using descriptive and inferential statistics, and qualitative data using the constant comparative approach. Results show that Twitter use is predominantly connected to traditional work hours and is well-integrated into academics’ professional endeavors, suggesting that professional use of Twitter has become “ordinary.” Though scholars rarely announce their departure from or return to Twitter, approximately half of this study’s participants took some kind of a break from Twitter. Although users returned to Twitter for both professional and personal reasons, conferences and workshops were found to be significant events stimulating the return of academic users.Contents
Introduction
Review of relevant literature
Methods
Results
Discussion and Implications
Limitations
Conclusion
As academics are increasingly encouraged to participate online for a variety of teaching and scholarly purposes, the research community faces a critical need to understand how scholars are using social media in their actual day-to-day practices. Importantly, researchers’ current understanding of how scholars use social media is constrained by limited knowledge of how social media use varies or does not vary over time. Understanding temporal aspects of social media use is important because such knowledge will shed light on how prevalent social media are in scholars’ lives and how academics participate on social media over time. Investigating the factors that might explain temporal variations in social media use and understanding how often and why scholars temporarily disengage from and return to social media will enable us to make greater sense of whether social media are used differently by different academics and allow us to interrogate the place, and perhaps roles, of social media tools in contemporary scholarly cultures. This paper therefore contributes to the literatures on digital and networked scholarship, and addresses literature examining academics’ use of digital technologies.
In this paper, we answer the following broad research questions: Is academics’ participation on Twitter consistent over time? Do they take breaks and subsequently return to the platform? What factors might explain breaks and returns? We focus on Twitter for a variety of reasons that we explain below. We address the identified knowledge gap by using data mining, descriptive, and qualitative analysis techniques. In particular, we use the Twitter API to retrieve available tweets from 3,996 individuals who self-identify as lecturers, adjunct professors, assistant professors, and associate professors. Next, we examine participation over the duration of the day and week and use ANOVA and chi-square tests to investigate differences between groups. We also report descriptive statistics on scholars’ temporary disengagement from Twitter and present qualitative themes identifying reasons explaining scholars’ temporary disengagement and subsequent return to Twitter. We find that scholars use Twitter predominantly during traditional work hours and for professional endeavours. We also find that even though participants rarely announce their departure from or return to Twitter, approximately half of them took a break from Twitter. Finally, we found that scholars leave and return to Twitter for both professional and personal reasons. We proceed by reviewing literature relevant to the topic, describing our methodology and presenting our research results. We conclude by discussing the implications for research and practice.
Social media platforms (e.g., Twitter, Facebook, Instagram, and YouTube) are commonly used by faculty members at institutions of higher education. Surveys show that these platforms are being used by anywhere between 30 percent and 70 percent of faculty members (Bowman, 2015; Moran, et al., 2011; Moran and Tinti-Kane, 2013; Van Noorden, 2014). Scholars use these platforms for a variety of purposes, including for teaching (e.g., classroom communication), research (e.g., publication dissemination, as a conference backchannel), and service (e.g., community building), as well as for a variety of personal purposes. Even though extensive literature examines social media use in scholars’ professional and personal endeavors (Veletsianos, 2016), temporal dimensions of social media participation are relatively unexplored.
In this study we focus on Twitter for two reasons: it is widely used by scholars (Lupton, 2014) and the platform enables us to unobtrusively identify and examine scholars’ actual temporal participation via access to the platform’s Application Programming Interface (API). Due to the platform’s popularity with academics, researchers frequently use Twitter as a setting to explore scholars’ use of social media such as the behaviors scholars exhibit, the motivations that bring them to the platform, and the challenges that they face on social media (DeGroot, et al., 2015; Gettman and Cortijo, 2015; Jordan, 2017; Kimmons and Veletsianos, 2016; Lemon, et al., 2015; McHeyzer-Williams and McHeyzer-Williams, 2016; Ortega, 2016; Rainford, 2016; Veletsianos and Kimmons, 2016).
The current literature reveals that scholars use social media for a variety of purposes (Donelan, 2016) and that these purposes might bear a relationship to temporal dimensions of social media use. In a survey conducted by Moran, et al. (2011), for example, over 90 percent of faculty reported using social media in classroom instruction. This finding would seem to suggest that faculty are more likely to use social media during class time and during the academic year. Other surveys indicate that faculty are more likely to use Twitter when they are at academic conferences as a type of backchannel for conversation surrounding presentations and speakers (Donelan, 2016; Ross, et al., 2011). Therefore, one would expect that Twitter use spikes around conferences and, for some disciplines, around conference season. Kimmons and Veletsianos (2016) found that academics’ participation in event-based hashtags on Twitter (e.g., hashtags relating to a conference or an event of wide social interest) exhibited spikes and slumps, thereby suggesting that Twitter participation may be ephemeral rather than ongoing.
Current research on temporal dimensions of online participation is limited. In a recent synthesis of the literature on social media in education, Gao, et al. (2012) noted that the majority of the studies on the topic focused on a short period of time, and the authors suggested that researchers should conduct longer-term studies. For example, there is a paucity of research on the frequency (e.g., daily, bi-weekly, monthly) of scholars’ participation on social media. Early research on Twitter use shows that the average user may post updates to his or her account every few days, whereas more frequent users who might use Twitter as a microblogging platform may be updating it several times a day (Java, et al., 2007). Although some studies have examined frequency of use in classroom settings (Abe and Jordan, 2013; Carpenter and Krukta, 2014), only a handful of studies examine temporal dimensions of social media use in scholars’ lives.
The few studies that explore temporal dimensions of social media use among scholars focus on average frequency but do not explore changes in participation frequency over time. For example, Haustein, et al. (2014) discuss findings pertaining to frequency of activity on Twitter among 37 astrophysicists. They found that five scholars tweeted rarely (0.0–0.1 tweets per day), 11 tweeted occasionally (0.1–0.9 tweets per day), 11 tweeted regularly (1.2–2.9 tweets per day), and 10 tweeted frequently (3.7–58.2 tweets per day). Work, et al. (2015) reported similar results with regard to distributions of tweets per day in their study of doctoral students’ use of Twitter. They found that a small percentage (17.5 percent) were regular Twitter users while the majority used Twitter “occasionally (45 percent), rarely (25.6 percent), or not at all (11.9 percent)” [1]. While the results of the studies by Haustein, et al. (2014) and Work, et al. (2015) found some diversity in terms of frequency of overall tweeting activity among scholars; they assumed a constant level of social media activity over time and did not explore change in the frequency of social media use among individual scholars. In a qualitative study of social media use, Kieslinger (2015) explored influencing factors related to frequency of social media use. The author categorized scholars into heavy, targeted, and restricted users: heavy users were characterized by daily social media use for both professional and personal use; targeted users went on social media platforms for strategic purposes and limited time spent online; and, restricted users had a static online presence, chose not to participate in social media related activities, and thus spent the least amount of time online. Lack of time was one of the primary barriers that the restricted users in that study attributed to their limited social media presence.
Scholars express challenges related to time with regard to online activity in general and not just in terms of social media practices. Along with spending their time on activities deemed important by their institutions such as writing, research, and teaching, scholars now have to consider the significance of building and maintaining an online presence (Lowenthal, et al., 2016). In the existing body of literature, feelings of concern are reported relative to academic workload, work-life balance, and the potential for time spent on social media to reduce scholarly productivity and to affect scholars’ lives in negative ways (Ferguson, 2017; Lemon, et al., 2015; Veletsianos and Kimmons, 2013). For example, in their study of Ph.D. candidates and their appropriation of online tools, Dowling and Wilson (2017) reported a reticence to adopt online tools beyond those necessary for basic communication and research. Those authors noted a particular vehemence toward social media and its potential for distraction as Ph.D. candidates manage time pressures in their doctoral studies. In two studies where scholars have reflected on their personal use of Twitter, they noted an effort to use Twitter with deliberate time boundaries. In their reflection on their year spent experimenting with Twitter for professional purposes, McHeyzer-Williams and McHeyzer-Williams (2016) described using Twitter in a way that did not disrupt their regular academic work schedule. Budge, et al. (2016) each provided a narrative of their Twitter experiences and gave examples of deliberately restricting Twitter use with respect to time (e.g., not using Twitter before bedtime and allocating the time spent riding public transportation to Twitter use). The broader literature on typical users’ social media use provides some information that may be worthwhile to consider here. For instance, youth appear to be using social media as a pastime activity for when their time is not taken up with other responsibilities with large spikes in their participation around 12:00 PM, and then again trending heavily upward between 4 PM and 8 PM (Piwek and Joinson, 2016; Quan-Haase and Young, 2011). Studies of the general population found there to be a peak of usage in the hours of the evening and early night, coinciding closely with the results of usage among youth (De Choudhury, et al., 2013).
Scholars’ concerns with regards to time extend beyond online participation. Spurling (2015), for example, explored academia’s temporal rhythms and qualities of time, particularly how time is categorized, carved out based on its purpose, and fragmented in an effort to mitigate work-life tensions. For scholars who are expected to adapt to increasingly networked practices, time is compressed in an effort to achieve hyper-efficiency and is fragmented through multitasking, which has lead to reports of increased stress and lack of large chunks of time for deep and reflective thought (Menzies and Newson, 2007). Such stress may be exacerbated by the affordances of digital communications (Vostal, 2015). In response, and in an effort to combat the culture of overwork, some scholars have called for scholars to engage in “slow scholarship” (Berg and Seeber, 2016).
One way scholars respond to time concerns, is to temporarily disengage from social media. Though scholars use a number of terms to describe this phenomenon (e.g., digital sabbatical, digital detox), at its essence disengagement represents a withdrawal from social media for a specific period of time. Limited research exists on this topic, but reports suggest that scholars may temporarily disengage for a variety of reasons such as their need to re-evaluate use, reclaim their time, and address a variety of other personal or professional concerns (e.g., Zellner, 2012). Looking beyond the scholarly community, there is a broader cultural movement to encourage temporary disengagement from technology. Turkle (2011), for example, posits that contemporary culture has become cyborg-like with a tendency to tether to devices. She calls for a greater awareness about the personal and cultural impacts of technology use. Others are questioning whether they are overly attached to technology (and whether their technology habits are negatively affecting their health and well-being), which highlights a broader cultural movement to deliberately spend time untethered from devices as a sort of ongoing digital diet and detox (Ellis, 2017).
We were able to identify one study which examined temporary disengagement from social media: Schoenebeck (2014) studied Twitter users who expressed in a tweet that they would be taking a 40–day break from Twitter for Lent, a religious observance that encourages taking a break from certain habits. By combining interviews with the tweets of those posting that they would be seeking a break, Schoenebeck found that 64 percent of users who attempted to take a hiatus from social media were successful in doing so. Among those who were not successful, 31 percent acknowledged their failure in their tweets, while the rest simply returned to the platform. In describing the reasons for disengagement, respondents expressed anxieties regarding their behavior on social media sites and its effect upon their day-to-day lives, as well as a desire to spend less time in virtual environments. These reasons align with the sentiments of scholars interviewed by Menzies and Newson (2007) with regard to managing temporal tensions that occur with online technology use. Schoenebeck’s study is unique in its observance of both operationalizing breaks from social media platforms and its observance of actual use patterns paired with qualitative data to explain motivations behind social media use.
In summary, the research community lacks understanding of temporal dimensions of online participation among scholars, including frequency of use, patterns of use, and disengagement. The lack of research in the area is reflective of the broader educational technology literature, as research that explores time as a variable is limited (Barbera, et al., 2015). Though there are broader cultural conversations regarding time spent on and disengagement from social media, there is little research examining these issues in relation to scholars’ social media use. In this study we examine scholars’ temporal participation on Twitter, reasons for temporary disengagement, and reasons that might bring them back to address the identified knowledge gaps. Building on Schoenebeck (2014) and data mining methods we employed in prior research (e.g., Kimmons and Veletsianos, 2016; Veletsianos and Kimmons, 2016), we collect and analyze social media data so as to examine temporal participation unobtrusively in real-world settings. The methods described below allow us to observe real use as opposed to self-reported use or use as recalled by participants.
The overarching question guiding this study was: Is academics’ participation on Twitter consistent over time? Do they take breaks and subsequently return to the platform? What factors might explain breaks and returns? To operationalize and answer these questions we subdivided them into the following research questions:
RQ1: How does scholar Twitter participation vary over time?
RQ2: How common and what are the characteristics of scholars’ temporary disengagement from Twitter?
RQ3: What reasons might explain scholars’ temporary disengagement from Twitter?
RQ4: What reasons might explain scholars’ subsequent returns to Twitter?
We then used a combination of descriptive, inferential, and constant comparative analysis methods to analyze quantitative and qualitative data. We built on previous work that employed such techniques to answer questions relevant to education (Veletsianos and Kimmons, 2016). To conduct this study, we (1) queried the Twitter Application Programming Interface (API) to identify scholars and to extract relevant Twitter data; (2) programmatically identified temporal participation patterns, including participation breaks and returns from breaks; (3) conducted quantitative analyses of the whole dataset; and, (4) conducted qualitative analyses of a representative sample of break-associated tweets. We explain these steps in more detail below.
First, we identified roughly 1,000 Twitter users from four different faculty groups: lecturer, assistant professor, associate professor, and adjunct professor. In total, we retrieved 3,996 users by querying the Twitter API with keywords for each group (e.g., “associate professor”). The Twitter API search focuses on user-provided profile descriptions and limits results returned to 1,000 users. These users’ names, usernames, biographies, and geographical location were collected. Complete available tweet histories were also extracted using the Twitter API, which permitted the extraction of up to 3,500 tweets per user along with associated metadata for each tweet (e.g., timestamp, location, number of times the tweet was retweeted). This resulted in a very large tweet dataset (n = 9,025,127 tweets). Gender of each user was predicted by alignment of user screen names to a corpus of common English names for each gender (e.g., 64 users were named David, and 22 users were named Lisa). Ambiguous names, such as Alex or Chris, were ignored in gender comparisons (Kantrowitz, 1993). This allowed for gender identification of 75.6 percent of users.
Table 1: Gender comparison by faculty group. Female Male n Percentage n Percentage Lecturer 278 39.4% 427 60.6% Adjunct 350 43.3% 459 56.7% Assistant 316 42.1% 434 57.9% Associate 295 39.0% 461 61.0% Overall 1,239 41.0% 1,781 59.0%
Second, we programmatically identified temporal participation patterns, including participation breaks and returns from breaks. Using location data from each tweet and provided timestamps, we determined the local time that each tweet was posted. In total, 86.1 percent of tweets provided metadata necessary to determine local time (n = 7,769,252). We operationalized “participation breaks“ as the time lag between a user posting original content (i.e., excluding retweets). We then calculated break times between tweets that were 30 days in length or longer and organized breaks into three interval groups: 30–59 days, 60–89 days, and 90+ days. For example, if a user posted two tweets, one on 1 March and one on 15 May, we included them in the 60–89 day interval group. To answer research questions three and four, we also created lag tables of each break to include three tweets prior to the break, the comeback tweet, and two subsequent tweets (cf., Table 2). Additional breaks intentionally broke the lag identification, because we only wanted to connect lag tweets to a break if they were not part of another break.
Table 2: An example of the lag tables created. Break ID User ID Break length t-3 t-2 t-1 Break t+1 t+2 t+3 1 1234 30 tweet tweet tweet [break] tweet tweet tweet 2 2345 60 tweet tweet tweet [break] tweet tweet [break] 3 3452 90 tweet tweet tweet [break] tweet [break] —
Third, we ran a variety of descriptive and inferential tests (e.g., ANOVAs, chi-square comparisons) on a number of variables as described below to determine results and significance for research questions one and two.
Finally, we thematically coded the three tweets immediately prior to and the three tweets immediately following 1,200 random instances of identified breaks to answer research questions three and four. Two researchers independently read each tweet and generated open codes describing the tweets. If existing codes were incapable of describing the tweets, new codes were developed. Whenever a new code was developed it was compared to data that had already been coded to examine whether new codes could be applied to data already examined. This process of constantly comparing data to codes (Glaser and Strauss, 1967) led to each researcher developing a list of codes characterizing the data. Next, the two researchers discussed and examined the codes with a third researcher, and generated a shared code book. The two researchers then collaboratively applied the codes to all the data. The three researchers then discussed the codes once again, and generated themes capturing the codes. The researchers took a series of steps to reduce bias and establish credibility as suggested by Creswell and Miller (2000), including conducting independent analysis by multiple researchers, peer debriefing, and reporting findings in thick and detailed descriptions.
RQ1: How does scholar Twitter participation vary over time?
Overall, nearly half of tweets were posted by scholars during the work week and during work hours (46 percent; cf., Table 3) with relatively few tweets being posted during the weekend (22 percent). Scholars averaged tweeting 264 times per day on weekends, and this jumped to 380 times per day on weekdays (an increase of 44 percent), revealing that weekdays were when they were most active on Twitter. Similarly, scholars averaged tweeting 25 times per hour during work hours and only 10.6 times per hour at other times. If we factor in eight hours per day for sleeping, non-work tweeting only elevated to 19.5 times per hour, which means that scholar tweet frequency increased by 28 percent when they were working as compared to out-of-work waking hours.
Table 3: Likelihood of tweets occurring during work time by faculty group.
Note: * Denotes significance at the p < .05 level; ** Denotes significance at the p < .01 level; *** Denotes significance at the p < .001 level.Work days & hours Work days Work hours Descriptive results Lecturer 45% 78% 57% Adjunct 46% 77% 59% Assistant 48% 79% 60% Associate 47% 79% 59% Overall 46% 78% 58% ANOVA results df 3 3 3 Mean square 0.25 0.08 0.23 F 13.06 *** 13.8 *** 13.93 *** Bonferroni post hoc mean differences Lecturer — Adjunct -0.02 ** 0 -0.03 *** Lecturer — Assistant -0.04 *** -.02 *** -0.04 *** Lecturer — Associate -0.03 ** -.01 ** -0.03 *** Adjunct — Assistant -0.02 * -.02 *** -0.01 Adjunct — Associate -0.01 -.02 *** 0 Assistant — Associate 0.02 0 0.01
Furthermore, in-work vs. out-of-work tweeting varied slightly between faculty groups, with lecturers tweeting the least during work (45 percent) and assistant professors tweeting the most (48 percent). A one-way analysis of variance (ANOVA) revealed that differences between groups in all three metrics were significant at the p < .001 level. Bonferroni post hoc testing revealed that differences existed between most groups with assistants and associates being the most similar and lecturers and adjuncts being unique in most areas (cf., Table 3). Lecturers tweeted during work and during work hours with the least relative frequency, and assistants and associates tweeted with the greatest relative frequency during work and on work days.
Across the week, scholar tweeting is least frequent on Sunday and then steadily climbs to its peak frequency on Wednesday (cf., Figure 1). This climb then reverses and returns close to its Sunday level on Saturday. This reveals that tweet frequency increases by 54 percent from Sunday to Wednesday and that scholars are most active on Twitter during the middle of the week and are least active on weekends.
Figure 1: Tweet frequencies by day and work hours.
Similarly, a comparison of tweet frequencies by time of day (for each day) reveals low tweet frequencies in the early morning hours with a sharp increase beginning at around 6 AM (cf., Figure 2). Frequencies peak at 10 AM and then gradually decline until 6 PM, wherein there is a slight increase again until 9 PM until frequencies plummet to a minimum at 3 AM. Weekends reveal less tweeting frequency for all hours, but differences are most pronounced during daylight or work hours (8 AM — 5 PM). It is interesting to note that though frequencies go down in the evening, scholars are nonetheless tweeting deep into the night, with their 10 PM frequencies matching their 8 AM frequencies. In fact, 98.4 percent of scholars tweeted at least once between 12:00 AM and 4:59 AM, and 88.2 percent tweeted between 2:00 AM and 3:59 AM, revealing that almost all scholars used Twitter during hours that are typically reserved for sleeping.
Figure 2: Tweet frequencies by hour in the day for each day.
Over the course of the year, tweets were least frequent during summer months (July and August), picked up slightly in fall months (September — December), and then continued to rise in winter and spring months (January — May; cf., Figure 3). Scholars’ tweet frequencies peaked in May and then began a sharp decline back into the summer, with July only generating 53 percent of the tweets that were generated two months earlier in May. This overall pattern of frequency also matched the pattern during the work day.
Figure 3: Tweet frequencies by month.
Regarding tweet originality, operationalized as any tweet that was not a retweet, 66.6 percent of tweets outside of work were original, but this number increased slightly to 69.3 percent during work. This reveals that the type of tweeting that scholars did during work involved slightly more original sharing than did their tweeting outside of work. This shift is most clearly seen when tweet originality (percent of tweets that are original or not retweets) is viewed by tweet hour (cf., Figure 4). As this pattern reveals, scholar tweet originality rises gradually from early morning and peaks at 11:00 AM. It then stays fairly constant until 4:00 PM and then declines into the evening as retweeting takes up a greater proportion of activities. This suggests that lurking, consuming, and sharing (as opposed to creating) activities are most common in the late hours of the evening.
Figure 4: Tweet originality by time of day (Note: Scale start at 64 percent).
RQ2: How common and what are the characteristics of scholars’ temporary disengagement from Twitter?
Overall, roughly half of scholars took some kind of Twitter break (48 percent), with the most common break length being between 30 and 59 days in length (45 percent; cf., Table 4). Among those who took breaks, multiple breaks were regular, with the average scholar who took any break taking 7.3 breaks in their Twitter lifetime. Likelihood and frequency of breaks varied according to length, with 30–59 day breaks being the most common (45 percent taking an average of 5.0 breaks) and 60–89 day breaks being the least common (26 percent taking an average of 2.1 breaks). Differences were also visible between groups, with lecturers being the least likely to take a break (31 percent vs. 48 percent) and also taking the fewest breaks (6.0 vs. 7.3).
Table 4: Likelihood and frequency of Twitter breaks by faculty group. Likelihood of a break Frequency of breaks by those who took one 30–59 days 60–89 days 90+ days Any break 30–59 days 60–89 days 90+ days Any break Lecturer 28% 15% 16% 31% 4.2 1.9 2.5 6.0 Adjunct 54% 34% 35% 57% 5.4 2.2 2.7 8.1 Assistant 53% 31% 30% 56% 5.0 2.0 2.5 7.2 Associate 46% 26% 27% 49% 5.0 2.2 2.7 7.3 Overall 45% 26% 27% 48% 5.0 2.1 2.6 7.3
We conducted a chi-square test of independence to determine whether scholar rank group influenced the user taking a break. Results were significant in all cases, and Cramér’s V values ranged from 0.15 to 0.21, revealing weak to moderate associations between scholar rank and taking a break (cf., Table 5). Furthermore, a one-way analysis of variance (ANOVA) revealed that differences in frequency of breaks among scholars who took a break only varied significantly in the case of 30–59 day breaks, F(3,1808)=5.21, p = .001, and a Bonferroni post hoc analysis revealed that lecturers differed significantly from adjuncts, MD = -1.16, p = .001, and assistants, MD = -.81, p = .04, on this measure.
Table 5: Chi-square results for break likelihood differences between scholar groups. df Chi-square Cramér’s V 30–59 days 3 174.05 *** 0.21 60–89 days 3 114.41 *** 0.17 90+ days 3 94.76 *** 0.15 Any break 3 168.42 *** 0.21
Of those who took a break and later returned to using Twitter, 50 percent began using Twitter again during work (i.e., during work hours on a work day; cf., Table 6). This is slightly higher than the likelihood of other (non-comeback) tweets being posted during work (46 percent).
Table 6: Likelihood that a comeback tweet will occur during work. Comeback tweets Other tweets 30–59 days 60–89 days 90+ days Any break Work days & hours 51% 50% 50% 50% 46% Work days 81% 79% 80% 80% 78% Work hours 61% 61% 60% 61% 58%
When considered across the year, scholars most commonly initiated breaks at the end of school semesters, with the plurality occurring in December (10.5 percent), followed by July (9.5 percent) and June (8.7 percent; cf., Figure 5). Conversely, comebacks were most likely to occur at the beginning of school semesters, with the plurality occurring in January (12.1 percent), followed by September (9.4 percent) and August (8.8 percent). If we consider relationships between start and end months of specific breaks, we find that the largest break-comeback relationships existed between December–January (6.1 percent), July–August (4.1 percent), and March–April (3.8 percent), suggesting that breaks and comebacks aligned with semester end and start dates and with common dates for spring break.
Figure 5: Twitter break initiations and comebacks by month.
Regarding originality, 33.5 percent of all tweets were retweets. Among comeback tweets, however, only 24.5 percent were retweets, and among comeback tweets after 90+ day breaks, only 21.9 percent were retweets. This reveals that posting something original was a greater pull to bring scholars back to Twitter than was reposting content created by the community. Scholars might not be actively posting on Twitter, but they seem nonetheless to be aware of what others are tweeting about, essentially being present, but not active contributors, on Twitter.
RQ3: What reasons might explain scholars temporary disengagement from Twitter?
Using a representative random sample of pre-disengagement tweets associated with 1,200 breaks (i.e., the three tweets posted prior to a break, totalling 3,600 tweets), we identified six instances in which the user posted a tweet stating their intent to take a break. Overwhelmingly then, disengagement from the platform is rarely announced. In these six cases, we were able to identify explicit reasons for disengagement. We examined the content of the rest of the tweets surrounding breaks to identify potential reasons explaining disengagement. We identified another 29 instances of breaks alluding to reasons for disengagement. Following the data analysis process described earlier, we developed 12 codes summarizing these reasons for disengagement. These reasons were then developed into four themes describing the reasons that led scholars to temporarily disengage from Twitter. The four themes — which demonstrate that scholars temporarily disengage from Twitter for both professional and personal reasons — are: excessive workload, being new to Twitter, reflecting on social media use, and idiosyncratic reasons.
Excessive workload. Scholars alluded to excessive workload as the most frequent reason to explain disengagement from Twitter, and it was often associated with concerns pertaining to the use of time. Concerns around high workload were mentioned in 15 instances, and are well-documented in the literature as tensions faced by faculty members. Scholars noted spending time on writing projects and being away from Twitter in order to complete grading, writing, and other activities (e.g., “I am going through email at the end of day. fingers crossed for this to go faster”).
Being new to Twitter. Being new to Twitter was the second most-frequently reported reason that reflected temporary disengagement. Scholars mentioned this reason eight times and it seems to reflect users who self-identified as being new to the platform (e.g., “Struggling to figure out how to connect my account to my phone ... I’m new here ... ”). Users’ initial participation seemed to reflect a pattern of posting one or two initial tweets, followed by a temporary pause for a short period of time (30–90 days) before posting another update. The data that we collected are not amenable to explaining why we observed this behavior from scholars who are new to Twitter, but we hypothesize that it takes time for new users to understand how the technology fits into their practice, leading to temporary disengagement or lack of participation.
Reflecting on social media use. Some (six) users reported temporarily disengaging from Twitter after reflecting on the role of the technology in their personal and professional life, as well as after reflecting on the broader implications that social media play in their life and broader culture. Scholars disengaged temporarily after reading books on social media use or after making conscious decisions to change the ways that they used social media (e.g., “I’ve been away from Twitter for a while after reading @user’s [book title]”).
Idiosyncratic reasons. Finally, users noted individual circumstances that led them to temporarily disengage from Twitter. We grouped these reasons together as they were observed less frequently than the rest of the reasons. Such personal circumstances included injuries, family emergencies, travel, and focusing one’s efforts on social networks other than Twitter (e.g., “I have a family emergency, and I will not be responding to messages until further notice”).
RQ4: What reasons might explain scholars’ subsequent return to Twitter?
Using the same representative sample of users and tweets, we found that only two percent of returning users explicitly stated that they had returned. To capture the reasons that might have brought users back to Twitter, we read each user’s three tweets following their return and coded them to describe activities that users engaged in following temporary disengagement from the platform. We developed 61 codes describing these activities. We then collapsed these codes into five themes (cf., Table 7).
Table 7: Themes identifying reasons that might have brought users back to Twitter. Theme Theme proportion Conversational presence 38% Professional emergence 25% Passive re-emergence 12% Societal discourse presence 11% Other 14%
Conversational presence. The most common activity users engaged in upon returning was conversations with other users. Conversation accounted for 38 percent of the tweets that users posted upon returning to the platform after a period of temporary disengagement. We categorized conversation as either professional (21 percent) or personal (14 percent). We also categorized a small sample of tweets as conversational presence (three percent), but was unclear whether such presence was personal or professional. Conversational presence most often manifested as a direct tweet to another user or as a conversational tweet to a broader audience in the form of a question or request for advice. Professional conversations, such as discussing academic topics in scholars’ areas of studies, were the most prominent types of conversation that presented themselves upon the return of these scholars to Twitter (e.g., “@user Are you interested in being a guest lecturer?”). Personal conversation encompassed a variety of topics relevant to individual users. Examples of such conversation included, but was not limited to, request for help with recipes, offering condolence, telling stories, letting other users know that they missed them or were thinking about them, and so forth (e.g., “user: What main dishes would you recommend for this diet? Please don’t say chips and salsa,” and “@user1 user2 @user3 Unfortunately, my kid is sick today!”).
Professional emergence. This theme represents tweets that highlighted professional activities other than conversations. Engaging in professional promotion was a popular activity, representing approximately 14 percent of returning tweets. Professional promotion activities included the sharing of recent publications and awards, promoting job postings at individual’s home institutions, sharing calls for participation in research studies, and a variety of posts that we deemed shared or promoted individuals’ research, teaching, or scholarship (e.g., “just received an email from a student from many years ago telling me how impactful my class was to her!” and “There’s still time to apply for our faculty cluster hire in [topic]: [URL]”). Conferences and workshops accounted for approximately 11 percent of tweets posted following a break, highlighting another form of professional emergence. The rest of the tweets in this category (one percent) were tweets that related to the commencement of a new academic year.
Passive re-emergence. Returning tweets that reflected content syndicated from another social media site or Web site constituted 12 percent of the sample. We described such participation as passive re-emergence because these posts were automated and reflected participation on a platform other than Twitter. Even though users’ activities on other platforms were made visible on Twitter, we considered this activity to reflect passive use of Twitter rather than active engagement on the platform. Posts categorized as passive re-emergence commonly included content from photo sharing Web sites, LinkedIn, Facebook, and other platforms (e.g., “I entered the lottery to [hashtag] on @TodayTix! Enter here: [URL]”). We also identified tweets where articles were shared from a news Web site, but appeared to be syndicated from another platform because there was no additional comments made by the user on Twitter. Importantly, we often observed scholars actively participating on Twitter following the syndicated tweet. It should be emphasized therefore, that this theme reflects initial re-emergence on Twitter as opposed to passive participation in the long-term.
Societal discourse presence. Upon re-emergence on Twitter, scholars also posted commentary relating to current events and issues (11 percent). Such content included commentary on politics, current news, and current social issues (e.g., “Excited by Obama’s historic victory!” and “NATO’s default position is to deny the killing of civilians in [country]. Unsurprising reaction”). It should be noted that causation is difficult to identify here: Societal issues may have prompted users to return to Twitter to participate in the discourse, but it is equally likely that users decided to return to Twitter for reasons unrelated to current social events and decided to comment upon them upon returning. It is also likely that some users were consuming Twitter content while not being visibly active and that some social issues prompted them to participate in a way that made their presence visible to the data collection processes employed in this paper.
Other. Re-emergence and re-engagement with Twitter also aligned with a number of other topics that defied a common grouping (14 percent). Content included in this category was individual and varied from user to user. For example, while some users discussed personal interests such as literature, music, and cooking, others shared quotes, commented on being hacked, and discussed workload issues (e.g., “I wish I were at wimbledon watching the tennis games” and “Spent my sunday horse riding and picking strawberries. Back to work tomorrow”). In a number of instances (23), scholars interacted with companies, often asking them for help with a product or complaining about a company’s products and services.
Researchers face a critical need to develop an evidence-based understanding of scholars’ use of digital technologies over time and to understand why academics engage with digital technologies in the ways that they do. In this paper we used data mining techniques to identify nearly 4,000 academics, retrieve more than nine million tweets they posted, and identify participation breaks and returns to the platform. In analyzing the data we collected, we were able to examine scholars’ temporal participation patterns, the characteristics of temporary disengagement from Twitter, and the reasons that might potentially explain breaks from and returns to Twitter. In this section we discuss our findings and their implications for research and practice.
First, scholars’ Twitter use seems to heavily align with their work schedules at all levels of analysis (hours, days, months). Results show that the greatest portion of tweets are posted during traditional work hours in the work week. This finding suggests that Twitter is connected to, or at least used during, traditional work hours, and that such use has become ordinary. Even though the literature highlights that Twitter is used for both personal and professional reasons, this finding might suggest that use of the platform constitutes a greater part of academics’ work hours than non-traditional work hours and, therefore, potentially more integrated into their professional endeavors. These usage patterns appear to be unique. While the general population is seemingly using social media with peaks around noon, and larger usage surges in the evening (De Choudhury, et al., 2013; Piwek and Joinson, 2016; Quan-Haase and Young, 2011), scholars’ use of social media rises and falls with traditional work hours, showing a decline in the evening hours where others are seeing peak usage. This finding might indicate potential shifts in how academics are using their time with respect to technology over the last decade or so. If social media is now an ordinary part of some academics’ daily lives, in what ways should social media technologies be acknowledged at the institutional level? Have universities responded to shifts in how their faculty spend their time with respect to social media, and if so, how? Should institutions place higher value on social media as a result of some academics placing higher value on them over time? Or, is social media use part of the day-to-day, another uneventful activity that academics engage in as part of their work? These questions are important because their answers will help researchers develop a better picture of academics’ day-to-day realities with respect to social media use and technology more broadly, as well as of their institutions.
Second, this study adds to emerging literature that suggests that scholars’ social media use is complicated, partly because academics use social media for both professional and personal reasons in contextual ways (Veletsianos, 2016). For example, a significant amount of literature examining social media use in education focuses on teaching and learning, but in this study we show that teaching activities are not a driving reason that bring people back to Twitter following disengagement, even though Twitter use is connected to professional activity in a variety of ways (e.g., use aligns with work schedules, excessive workload is cited as a reason for disengagement, professional activities are cited as returns to the platform). Although we noted a variety of motivators to return from temporary disengagement, scholars’ participation for professional and personal topics was present across reasons for returning to Twitter. To understand how faculty and students use social media, we contend that we need to examine how social media are used in their broader lives. Focusing solely on pedagogical applications and uses of social media limits the sense that we can make of social media, and curtails our understanding of the social, political, economic, and historical aspects of social media use. Future research should continue examining academics’ use of social media use beyond teaching and learning. Such work may not only help researchers make sense of how and why scholars participate on social media but may also shed light on reasons why social media are used in the ways that they are in educational settings. For example, academics frequently ask students to create Twitter accounts to use in their classrooms, and even though data for what happens after the class ends are not available, we suspect that a great majority of those accounts go dormant. One reason for this may be the way that faculty approach their own social media participation. If faculty participate on social media in sporadic ways and return to social media for particular events — such as conferences — then it may be reasonable to hypothesize that student Twitter accounts created as part of class go dormant due to the ways that faculty model social media use.
Third, this study shows that one particular professional activity — conferences and workshops — is important in bringing scholars back to Twitter after temporary disengagement. This finding is consistent with prior research that noted that faculty are more likely to use Twitter at conferences (Donelan, 2016; Ross, et al., 2011) and adds further empirical evidence to highlight the ephemeral, rather than ongoing, nature of Twitter participation (Kimmons and Veletsianos, 2016). Conferences seem to have an effect on Twitter participation, but to further gauge this relationship it is important for future research to explore this topic in greater detail. In what ways is Twitter use tied to conferences? The current research seems to indicate a relationship between the occurrence of a conference and active Twitter participation, but as Twitter use consists of producing, consuming, and circulating content, we currently have little evidence to understand the multitude of ways that Twitter is used beyond what is made visible through active participation. For example, while we may observe an individual consistently posting to conference hashtags, with breaks in-between conferences, we do not know whether this individual is absent in-between conferences or whether they are consuming content on the platform. In other words, researchers need to understand the extent to which active participation (i.e., posting) is related to conference participation. Further questions that should be investigated include the following: What does Twitter activity look like following conference backchannel participation? Do scholars go back to using Twitter as before? Does their use taper off over time? How does use vary between different users (e.g., those who were using Twitter in an ongoing way prior to their conference participation vs. those who were using it ephemeral ways)?
Fourth, this study highlights the importance of examining temporal factors in educational technology research. The empirical evidence describing the nature of academics’ online participation over time is scant and is largely predicated on small-scale studies. The field of educational technology overall faces this problem, and prior literature has called for more long-term studies (Barbera, et al., 2015; Gao, et al., 2012). For example, this research shows that almost a third of academics have disconnected from Twitter for 90 days or more; this finding is illustrative of insights that can be generated by conducting long-term studies into social media use that might lead researchers to question assumptions made about social media or ask new research questions. For instance, this simple finding leads us to ask: What compels academics to return to Twitter after three months? How do academics feel about returning to the platform? How is use of social media before and after a given break similar or different?
Readers should note a number of limitations that face this study. First, the data collection methods employed do not enable us to observe passive participation (such a person reading but not posting tweets) or private participation (such as a person sending direct messages but not posting public tweets). As a result this research is only able to make claims about what is observable in the public sphere rather than what scholars’ lived practices around passive/private participation may be (cf., Veletsianos, et al., 2015). Second, while this study generates insights into the reasons that academics leave and return to Twitter, there may be many other reasons that explain these issues that have not been identified. Such reasons can be further delineated by employing alternative data collection methods (e.g., asking scholars directly). For this reason, this study recognizes that the list of reasons generated for disengagement and re-emergence are likely not exhaustive. Third, the users included in this study are the ones that self-identified using the keywords lecturer, assistant professor, associate professor, and adjunct professor in their biographies. This method limits the sample to those individuals that self-identify using these keywords and excludes those individuals that don’t self-identify, use a language other than English to identify, or use a different title (e.g., Prof, professor, reader, adjunct faculty, etc). This method also limits the sample to those that were then returned by the Twitter search API. Twitter does not provide information on selection criteria used to return results to API queries, and as a result it is unclear as to the degree to which these results are representative of all academics participating on Twitter. Finally, this study focuses on one particular platform, and as such readers should be cautious about extrapolating the results presented here to social media in general. Though we hypothesize that academics’ disengagement and return to other social media platforms may be explained by similar reasons, we are not able to provide evidence to support this notion. Though social media platforms are similar in some ways, they are distinct in others, and those differences may yield alternative results and explanations. Significantly, this is an area that future research should address as the community begins to gain a better picture of social media use in academics’ lives, as opposed to just the use of a single platform.
By leveraging the Twitter API we were able to identify a large sample of academics and download their historical tweets. We used this dataset to gain insights into academics’ participation patterns over time and to interrogate temporal uses of Twitter. While the vast majority of previous research in the literature examining academics’ use of digital technologies has generated snapshots of Twitter use, this study has enabled us to gain an understanding of time dimensions of Twitter use, both in terms of frequency (e.g., daily, weekly), but also longitudinally (e.g., over time). Results show that Twitter use is predominantly connected to traditional work hours and well-integrated into academics’ professional endeavors, suggesting that professional use of Twitter has become “ordinary.” Though scholars rarely announce their departure from or return to Twitter, approximately half of this study’s participants took some kind of a break from Twitter. Although users returned to Twitter for both professional and personal reasons, we find conferences and workshops to be significant events stimulating the return of academic users. This research not only demonstrated the value of temporal and longitudinal analyses of social media data in the context of education, but also highlights the potential that temporal data provide for examining how social media are used by students and faculty.
About the authors
George Veletsianos, Ph.D., is is a Canada Research Chair in Innovative Learning and Technology and professor in the School of Education and Technology at Royal Roads University in Victoria, British Columbia.
E-mail: veletsianos [at] gmail [dot] comRoyce Kimmons, Ph.D., is an Assistant Professor of Instructional Psychology and Technology at Brigham Young University.
E-mail: roycekimmons [at] byu [dot] eduOlga Belikov is a doctoral student at Brigham Young University.
E-mail: olgambelikov [at] gmail [dot] comNicole Johnson is a research assistant at Royal Roads University.
E-mail: digitalnicole78 [at] gmail [dot] comAcknowledgements
This research was supported by the Social Sciences and Humanities Research Council of Canada.
Note
1. Work, et al., 2015, p. 36.
References
P. Abe and N. A. Jordan, 2013. “Integrating social media Into the classroom curriculum,” About Campus, volume 18, number 1, pp. 16–20.
doi: https://doi.org/10.1002/abc.21107, accessed 17 October 2018.E. Barbera, B. Gros, and P. A. Kirschner, 2015. “Paradox of time in research on educational technology,” Time & Society, volume 24, number 1, pp. 96–108.
doi: https://doi.org/10.1177/0961463X14522178, accessed 17 October 2018.M. Berg and B. Seeber, 2016. Slow professor: Challenging the culture of speed in the academy. Toronto: University of Toronto Press.
T. D. Bowman, 2015. “Differences in personal and professional tweets of scholars,” Aslib Journal of Information Management, volume 67, number 3, pp. 356–371.
doi: https://doi.org/10.1108/AJIM-12-2014-0180, accessed 17 October 2018.K. Budge, N. Lemon, and M. McPherson, 2016. “Academics who tweet: ‘Messy’ identities in academia,” Journal of Applied Research in Higher Education, volume 8, number 2, pp. 210–221.
doi: https://doi.org/10.1108/JARHE-11-2014-0114, accessed 17 October 2018.J. Carpenter and D. Krutka, 2014. “How and why educators use Twitter: A survey of the field,” Journal of Research on Technology in Education, volume 46, number 4, pp. 414–434.
doi: https://doi.org/10.1080/15391523.2014.925701, accessed 19 November 2018.J. W. Creswell and D. L. Miller, 2000. “Determining validity in qualitative inquiry,” Theory into Practice, volume 39, number 3, pp. 124–130.
doi: https://doi.org/10.1207/s15430421tip3903_2, accessed 17 October 2018.M. De Choudhury, M. Gamon, S. Counts, and E. Horvitz, 2013. “Predicting depression via social media,” Proceedings of the Seventh International AAAI Conference on Weblogs and Social Media, pp. 128–137, and https://www.aaai.org/ocs/index.php/ICWSM/ICWSM13/paper/view/6124, accessed 17 October 2018.
J. M. DeGroot, V. J.Young, and S. H. VanSlette, 2015. “Twitter use and its effects on student perception of instructor credibility,” Communication Education, volume 64, number 4, pp. 419–437.
doi: https://doi.org/10.1080/03634523.2015.1014386, accessed 17 October 2018.H. Donelan, 2016. “Social media for professional development and networking opportunities in academia,” Journal of Further and Higher Education, volume 40, number 5, pp. 706–729.
doi: https://doi.org/10.1080/0309877X.2015.1014321, accessed 17 October 2018.R. Dowling and M. Wilson, 2017. “Digital doctorates? An exploratory study of PhD candidates’ use of online tools,” Innovations in Education and Teaching International, volume 54, number 1, pp. 76–86.
doi: https://doi.org/10.1080/14703297.2015.1058720, accessed 17 October 2018.M. Ellis, 2017. Digitox: How to find a healthy balance for your family’s digital diet. Oxfordshire: CultureTransform Limited.
H. Ferguson, 2017. “Building online academic community: Reputation work on Twitter,” M/C Journal, volume 20, number 2, at http://journal.media-culture.org.au/index.php/mcjournal/article/view/1196, accessed 17 October 2018.
F. Gao, T. Luo, and K. Zhang, 2012. “Tweeting for learning: A critical analysis of research on microblogging in education published in 2008–2011,” British Journal of Educational Technology, volume 43, number 5, pp. 783–801.
doi: https://doi.org/10.1111/j.1467-8535.2012.01357.x, accessed 17 October 2018.H. J. Gettman and V. Cortijo, 2015. “‘Leave me and my Facebook alone!’ Understanding college students’ relationship with Facebook and its use for academic purposes,” International Journal for the Scholarship of Teaching and Learning, volume 9, number 1, article 8.
doi: https://doi.org/10.20429/ijsotl.2015.090108, accessed 17 October 2018.B. G. Glaser and A. L. Strauss, 1967. The discovery of grounded theory: Strategies for qualitative research. Chicago: Aldine.
S. Haustein, T. D. Bowman, K. Holmberg, I. Peters, and V. Larivière, 2014. “Astrophysicists on Twitter: An in-depth analysis of tweeting and scientific publication behavior,” Aslib Journal of Information Management, volume 66, number 3, pp. 279–296.
doi: https://doi.org/10.1108/AJIM-09-2013-0081, accessed 17 October 2018.A. Java, X. Song, T. Finin, and B. Tseng, 2007. “Why we Twitter: Understanding microblogging usage and communities,” WebKDD/SNA-KDD ’07: Proceedings of the Ninth WebKDD and First SNA-KDD 2007 Workshop on Web Mining and Social Network Analysis, pp. 56–65.
doi: https://doi.org/10.1145/1348549.1348556, accessed 17 October 2018.K. Jordan, 2017. “Understanding the structure and role of academics' ego-networks on social networking sites,” doctoral thesis, Open University http://oro.open.ac.uk/48259/, accessed 17 October 2018.
M. Kantrowitz, 1993. “Name corpus: List of male, female, and pet names,” at http://www.cs.cmu.edu/afs/cs/project/ai-repository/ai/areas/nlp/corpora/names/, accessed 17 October 2018.
B. Kieslinger, 2015. “Academic peer pressure in social media: Experiences from the heavy, the targeted and the restricted user,” First Monday, volume 20, number 6, at http://firstmonday.org/article/view/5854/4580, accessed 17 October 2018.
doi: http://dx.doi.org/10.5210/fm.v20i6.5854, accessed 17 October 2018.R. Kimmons and G. Veletsianos, 2016. “Education scholars’ evolving uses of Twitter as a conference backchannel and social commentary platform,” British Journal of Educational Technology, volume 47, number 3, pp. 445–464.
doi: http://dx.doi.org/10.5210/fm.v20i6.5854, accessed 17 October 2018.N. Lemon, M. McPherson, and K. Budge, 2015. “Academics doing it differently: Wooing, hooking up and spinning stories,” Journal of Perspectives in Applied Academic Practice, volume 3, number 2, pp. 15–25.
P. R. Lowenthal, J. C., Dunlap, and P. Stitson, 2016. “Creating an intentional Web presence: Strategies for every educational technology professional,” TechTrends, volume 60, number 4, pp. 320–329.
doi: https://doi.org/10.1007/s11528-016-0056-1, accessed 17 October 2018.D. A. Lupton, 2014. “‘Feeling better connected’: Academics’ use of social media,” News & Media Research Centre (10 June), at http://www.canberra.edu.au/about-uc/faculties/arts-design/attachments2/pdf/n-and-mrc/Feeling-Better-Connected-report-final.pdf, accessed 17 October 2018.
L. J. McHeyzer-Williams and M. G. McHeyzer-Williams, 2016. “Our year on Twitter: Science in #SocialMedia,” Trends in Immunology, volume 37, number 4, pp. 260–265.
doi: https://doi.org/10.1016/j.it.2016.02.005, accessed 17 October 2018.H. Menzies and J. Newson, 2007. “No time to think: Academics’ life in the globally wired university,” Time & Society, volume 16, number 1, pp. 83–98.
doi: https://doi.org/10.1177/0961463X07074103, accessed 17 October 2018.M. Moran and H. Tinti-Kane, 2013. “Social media for teaching and learning,” Pearson Learning Solutions, at http://www.pearsonlearningsolutions.com/assets/downloads/reports/social-media-for-teaching-and-learning-2013-report.pdf, accessed 17 October 2018.
M. Moran, J. Seaman, and H. Tinti-Kane, 2011. “Teaching, learning, and sharing: How today’s higher education faculty use social media,” Babson Survey Research Group, at https://files.eric.ed.gov/fulltext/ED535130.pdf, accessed 17 October 2018.
J. L. Ortega, J2016. “To be or not to be on Twitter, and its relationship with the tweeting and citation of research papers,” Scientometrics, volume 109, number 2, pp. 1,353–1,364.
doi: https://doi.org/10.1007/s11192-016-2113-0, accessed 17 October 2018.L. Piwek and A. Joinson, 2016. “‘What do they Snachat about?’ Patterns of use in time-limited instant messaging service,” Computers in Human Behavior, volume 54, pp. 358–367.
doi: https://doi.org/10.1016/j.chb.2015.08.026, accessed 17 October 2018.A. Quan-Haase and A. L. Young, 2010. “Uses and gratifications of social media: A comparison of Facebook and instant messaging,” Bulletin of Science, Technology & Society, volume 30, number 5, pp. 350–361.
doi: https://doi.org/10.1177/0270467610380009, accessed 17 October 2018.J. Rainford, 2016. “Making internal conversations public: Reflexivity of the connected doctoral researcher and its transmission beyond the walls of the academy,” Journal of Applied Social Theory, volume 1, number 1, pp. 44–60, and at https://socialtheoryapplied.com/journal/jast/article/view/14, accessed 17 October 2018.
C. Ross, M. Terras, C. Warwick, and A. Welsh, 2011. “Enabled backchannel: Conference Twitter use by digital humanists,” Journal of Documentation, volume 67, number 2, pp. 214–237.
doi: https://doi.org/10.1108/00220411111109449, accessed 17 October 2018.S. Y. Schoenebeck, 2014. “Giving up Twitter for Lent: How and why we take breaks from social media,” CHI ’14: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp. 773–782.
doi: https://doi.org/10.1145/2556288.2556983, accessed 17 October 2018.N. Spurling, 2015. “Differential experiences of time in academic work: How qualities of time are made in practice,” Time & Society, volume 24, number 3, pp. 367–389.
doi: https://doi.org/10.1177/0961463X15575842, accessed 17 October 2018.S. Turkle, 2011. Alone together: Why we expect more from technology and less from each other. New York: Basic Books.
R. Van Noorden, 2014. “Online collaboration: Scientists and the social network,” Nature, volume 512, number 7513 (13 August), pp. 126–129.
doi: https://doi.org/10.1038/512126a, accessed 17 October 2018.G. Veletsianos, 2016. Networked scholars: Social media in academia. New York: Routledge.
G. Veletsianos and R. Kimmons, 2016. “Scholars in an increasingly open and digital world: How do education professors and students use Twitter,” Internet and Higher Education, volume 30, pp. 1–10.
doi: https://doi.org/10.1016/j.iheduc.2016.02.002a, accessed 17 October 2018.G. Veletsianos and R. Kimmons, 2013. “Scholars and faculty members’ lived experiences in online social networks,” Internet and Higher Education, volume 16, pp. 43–50.
doi: https://doi.org/10.1016/j.iheduc.2012.01.004, accessed 17 October 2018.G. Veletsianos A. Collier, and E. Schneider, 2015. “Digging deeper into learners’ experiences in MOOCs: Participation in social networks outside of MOOCs, notetaking and contexts surrounding content consumption,” British Journal of Educational Technology, volume 46, number 3, pp. 570–587.
doi: https://doi.org/10.1111/bjet.12297, accessed 17 October 2018.F. Vostal, 2015. “Speed kills, speed thrills: Constraining and enabling accelerations in academic work-life,” Globalisation, Societies and Education, volume 13, number 3, pp. 295–314.
doi: https://doi.org/10.1080/14767724.2014.959895, accessed 17 October 2018.S. Work, S. Haustein, T. D. Bowman, and V. Larivière, 2015. “Social media in scholarly communication. A review of the literature and empirical analysis of Twitter use by sshrc doctoral award recipients,” Study commissioned by the Social Sciences and Humanities Research Council, at https://crc.ebsi.umontreal.ca/en/publications/social-media-in-scholarly-communication-sshrc/, accessed 17 October 2018.
A. Zellner, 2012. “Digital detox, email vacations, and #twabbaticals,” GradHacker (12 April), at http://www.gradhacker.org/2012/04/13/digital-detox-email-vacations-and-twabbaticals/, accessed 17 October 2018.
Editorial history
Received 12 March 2018; revised 12 September 2018; 17 October 2018.
Copyright © 2018, George Veletsianos, Royce Kimmons, Olga Belikov, and Nicole Johnson. All Rights Reserved.
Scholars’ temporal participation on, temporary disengagement from, and return to Twitter
by George Veletsianos, Royce Kimmons, Olga Belikov, and Nicole Johnson.
First Monday, Volume 23, Number 11 - 5 November 2018
https://firstmonday.org/ojs/index.php/fm/article/download/8346/7606
doi: http://dx.doi.org/10.5210/fm.v23i11.8346
