Restrictions

Our functions possess a few constraints. Basic, because our very own dataset relies heavily toward Twitter since program populating an individual reputation, it is vital to keep in mind that existing functions looking at the alignment anywhere between good user’s online and genuine profile commonly primary. While it is impractical that any particular one could well be totally other on their on the web character (as they are apt to be titled out-by their friends for like difference) studies show that pages carry out often exaggerate certain attributes of on their own on the social visualize. The research try, ergo, limited by the variance between the genuine user profile plus the depicted one. These variations are likely to be for example obvious about framework off extroversion (Amichai Burger ainsi que al., 2002; Mitchell et al., 2011).

Second, some properties within dataset has somewhat less products than others and therefore shall be translated accordingly. But not, apart from NCAA and you can faith (within this which simply Muslims got reduced count) all our metrics incorporated at the very least 100,000 possible fits and you may typically included more step 1,000,000 examples if you don’t that order regarding magnitude much more.