[maus42]

1. Averages don't work that way. 2. Congratulations, you might not be as representative sample of the population as you thought!

This is one reason why advocate visualizing the whole distribution into a some kind of graph whenever possible.

I have institutional access, and unfortunately, the paper does not describe this distribution. Fortunately, GSS data is publicly available: http://www.gss.norc.org/

edit. Outside the year 2012, when the questionnaire logic was slightly different, the mode of the distribution of 6 available categories ranging from "not at all" to "more than 3 times a week", has been almost always been "not at all". (With second peak around "2 or 3 times a week".)