Discussion
The status of all five hypotheses is summarised in Table 5. Discounting social time, H1 (on predicting total network size) was supported for all predictor variables. The association between social media contacts and the total offline network reflects the well-established finding that social media networks help to maintain real world social networks (Ellison et al., 2007; boyd & Ellison, 2007), while the link between social media time agrees with previous studies demonstrating links between frequency of Internet and social media use and the number of friends in social networks (Wellman et al., 2001; Kraut et al., 2002). However, no association was found with social satisfaction, possibly indicating that the positive relationship between social support and intensity or frequency of social media use (Ellison et al., 2007, Carr, Wohn & Hayes, 2016) may not translate into satisfaction. Regarding H2 (on predicting sympathy group size), partial support was found for social activity and social media contacts, possibly reflecting the volume of social media friends who may be influenced by offline friends at this more intimate level. Again no association was found with satisfaction; however, comparing the sympathy group with evidence for strong ties providing social support (Wright & Miller, 2010; Rains & Keating, 2011; Blight et al., 2015) is difficult since measures of strong ties in previous studies may overlap both sympathy and support groups in our study.