I am fairly new to social media, and Twitter in particular. I’m a scientist by training and am interested in differences between believers and non-believers outside of the obvious difference in whether or not they genuflect before an imaginary being. Social media provide a great avenue for doing some naturalistic research. So over the next little while I am going to try to explore the differences between believers and non-believers using the words that they use to describe themselves on Twitter. Any thoughts on what differences I will find?
For my first stab at this, I looked at the first 200 folks that showed up on Twellow (the ‘yellow pages’ for Twitter) with the #Atheist and compared them to the first 200 folks that showed up on Twellow with the #Christian (I discounted those with the actual name “Christian” if they did not also report being of the Christian faith). While reading about my findings, please keep in mind that this is not necessarily a random sample and that I did not do any statistical analyses.
After identifying my sample, I next took all of the words and placed them in an Excel file, and then removed all conjunctions, symbols, web addresses and numbers. I kept all of the descriptive words and used those for the analyses. So — drum roll, please!!! — are there any differences between how atheists and Christians describe themselves on Twitter?
Here are the top 9 words used by atheists and Christians in describing themselves:
The first thing to note is that there was some overlap between atheists and Christians, with the most frequently used words relating to to being a writer or blogger. This is not surprising as I took my sample from a social media site. There was, however, another similarity between the two groups — the most frequent word for both groups related to politics. In general, Christians tended to report themselves as being conservative (23.5 % of the sample), while atheists identified themselves as being liberal (13% of the sample). The difference between these two percentages is interesting. The Christian group was heavily centered around conservatism w(see Figure 1 below), and was the only group to specifically mention a politician – can you guess who? If god was actually a politician, it turns out that his name would be Ronald Reagan! For the atheist group there was a higher number of politically-related words, with a greater spread in political ideology. Atheists self-identified as anarchists, libertarians, conservatives, liberals and socialists – they covered the full spectrum (except for communism). However, the majority of those atheists identifying their political ideology identified as being liberal or left-of-center.
A striking difference in the most frequent words used by the two groups concerns words related to family (see Table 1 and Figure 2). Christians frequently use family-related words to describe themselves (wife, husband, father, and mother); in contrast, in the atheist group, none of these words showed up with any particular frequency (I used a cut-off of 5 out of 200 or 2.5% of the sample).
One interpretation is that family in more important to Christians than it is to atheists; I am sure this conclusion would be pleasing to the Christian camp. However, there is a hint of another interpretation. Two of the top nine frequently used words by the atheist group are “student” and “geek.” The use of these words suggests that the atheist group may be younger than the Christian group and not yet married, and this may account for the differences in the use of family-related words. Unfortunately, I did not have the ability to look at the ages of individuals in my sample, as such information was not included in most of the Twitter descriptions. It would certainly be interesting to obtain more data to examine this issue more carefully.
Another thing that I noticed when analyzing the data is that the atheist group appears to be more heterogeneous than the Christian group. Evidence for this was shown in the use of politically-related words (see Figure 1). Also, while the two groups used more or less the same number of unique descriptors (that had a frequency of at least 5 in 200), the rate of usage per word was higher in the Christian group (see Figure 3). This suggests to me a greater uniformity in the Christian group (a higher percentage is using the same word) as compared to the atheist group, but more data are needed to make any firm conclusions.
That’s all for now! I will continue to mine social media to explore other issues related to atheism and religion. Let me know what you think of these data, whether you have interpretations that differ from mine, and what other questions would be interesting to investigate using social media data.