So to me polls are what they are...If two or more come to the same conclusions...they’re either right for the same reasons or wrong for the same reasons. Whatever.
Or happenstance, within an estimated error.
I am partial to the CUNY study because of statements like this:
One of the distinguishing features of this survey, as of its predecessor in 1990, is that respondents were asked to describe themselves in terms of religion with an open-ended question. Interviewers did not prompt or offer a suggested list of potential answers. Moreover, the self-description of respondents was not based on whether established religious bodies, institutions, churches, mosques or synagogues considered them to be members. Quite the contrary, the survey sought to determine whether the respondents themselves regarded themselves as adherents of a religious community. Subjective rather than objective standards of religious identification were tapped by the survey.
This absolutely puts the control completely in the hands of the respondents, and leaves any outside influence or data out of the situation at all. The open ended opening question is about as unbiased as one can get in establishing the facts of those surveyed.