Below in the comments David Heddle says:
Of course there is no way, that I can see, of estimating how many of those leaving the church were self-identified Christians but who were actually in-the-closet unbelievers. Perhaps (who knows?) this is a sizable group, one that is beginning to come out of the closet as the stigma of being a non-believer wanes. Those cases, however many there may be, are a win-win: better for the church they have left, better for society that they feel comfortable enough to stop the masquerade.
In the GSS there is a variable GOD which asks how confident people are their belief in God. Below are the year-by-year changes for those who are Protestant & Catholic, those with “No Religion” & the whole population. 95% confidence intervals in parentheses.
Because of the sample sizes I am provisional about my conclusions, but, it seems that:
1) The religious categories stayed relatively stable
2) A slight increase in skepticism toward belief in God was evident in the total sample
3) Relatively stability in the fraction who “Know God Exists”
#2 is explicable in light of #1 if there was simply a change in the religious categories, something which we can be much more confident about (the “No religion” segment doubled over this span). #3 makes sense if the defection toward greater skepticism was generally from lukewarm or weaker believers.