Should I stay or should I go? The moderating effect of self-image congruity and trust on social networking continued use

Abstract

Recent reports show that although users of large online social networks (such as Facebook) score low in terms of satisfaction, they manage to retain, or even increase, their user base. This study sheds an exploratory light on the reuse behaviour of online social network services (SNS). Specifically, we investigate the moderating effects of self-image congruity and trust on the relationship between satisfaction and continuance to use online SNS. To capture post-adoption behaviour of SNS users, we employ the expectation–confirmation model as a core structure for our theoretical model. The model is empirically tested using survey data collected from 288 Facebook users. Results reveal a possible trend suggesting that users perceiving a high match of the SNS with their self-image will be more loyal even if they experience low satisfaction levels. The same trend has been extracted for trust and its moderating effect on the relationship between satisfaction and usage continuance intention. This paper concludes with a discussion on the theoretical and practical implications of our findings.

Keywords:

• social network services
• self-image congruity
• trust

Notes

1. http://www.theacsi.org/index.php?option=com_contentandview=articleandid=147andcatid=14andItemid=212andi=Internet+Social+Media (accessed January 15, 2014).

2. http://personalweb.about.com/od/Socialnetworkingresearch/a/Time-Spent-Social-Networking.htm

3. http://www.guardian.co.uk/technology/2012/feb/16/twitter-tweaks-privacy-guidelines

4. http://www.eltrun.gr/survey-2013/ (in Greek)

5. http://www.socialbakers.com/facebook-statistics/greece

6. http://pewinternet.org/Reports/2013/Social-media-users/The-State-of-Social-Media-Users/Overview.aspx

7. The formula for estimating the f² statistic is ( of interaction model – of main effects model) of main effects model). Values close to and above 0.02 indicate small moderating effects. Values close to and above 0.15 indicate medium moderating effects and values close to and above 0.35 indicate large moderating effects. The pseudo-F statistic is calculated using the formula where n is the sample size and k reflects the number of variables in the model.