Not a lonely crowd? Social connectedness, religious service attendance, and the spiritual but not religious

• Compare religious service attenders, religious non-attenders (RNA), spiritual but not religious (SBNR), and neither spiritual nor religious (NSNR).
• Three social connectedness outcomes from the 2006–2014 General Social Survey and 2006–2012 Portraits of American Life Study.
• Attenders are more connected than the RNA and NSNR, but there are few differences between Attenders and the SBNR.
• Prior levels of social connectedness explain some switches, but there is little association between switches and changes in social connectedness.
• Results have implications for what it means to be spiritual, measuring religion and spirituality, and the role of formal organizations in social life.

Using the 2006–2014 General Social Survey and 2006–2012 Portraits of American Life Study, I find that on three dimensions of social connectedness: social interaction frequency, core discussion network size, and number of close ties, that religious service attenders are more connected than religious non-attenders and then either spiritual nor religious, but there are few differences between attenders and the spiritual but not religious. Difference-in-differences and fixed-effects models show little evidence that switches between categories are associated with changes in connectedness, and additional models show that prior social connectedness explains only a small amount of future switches. This paper challenges assumptions that the non-religious are a homogenous group lacking the benefits provided though the social networks of religious congregations and has implications for research on what it means to be spiritual, measuring religion and spirituality, and understanding the role of formal organizations in social life.