Browse 114 concepts used in the study of religion, review how survey researchers measured them in the past, and quickly compare the results of more than 7,600 survey questions.
The archive is a collection of surveys, polls, and other data submitted by the foremost scholars and research centers in the world. Review and analyze data online, or download free of charge.
Examine the religious composition, religious freedoms, demographics, constitutional clauses, survey findings and multiple social and political measures for 250 nations.
View maps of the United States and individual states for hundreds of variables, including congregational membership, census data, crime statistics and many others.
Generate congregational membership reports for any county, state and urban area in the United States using data collected by the Religious Congregations & Membership Study.
The profiles chart schisms and mergers, document membership trends, offer basic descriptions, and link to additional resources for more than 400 past and present American religious groups.
Browse dozens of topics from a major national survey of religious congregations. See how the responses vary by the size, religious family and region of the congregation.
Browse dozens of topics covered by major national surveys. See how the responses vary by demographic categories and, when available, how they change over time.
View maps of the United States and individual states for hundreds of variables, including congregational membership, census data, crime statistics and many others.
Greeley, A. (2020, December 3). Survey of Chicago Catholics, 2007.
Summary
This 2007 telephone survey examined the attitudes of 524 former and current Catholics living in the Chicago Archdiocese, which encompasses Cook and Lake counties and is home to 2.5 million Catholics. Current Catholics were asked a wide range of questions about their views of the church, as well as aspects of Catholic identity and commitment. Former Catholics were asked about reasons for leaving the church.
The study employed a probability sample from all telephone households in Cook and Lake counties. The sampling frame for the survey was constructed from three separate components. From a commercial list vendor we obtained listings (with telephone numbers) of residents of Cook and Lake counties; these were identified as "Catholics" or "not Catholics," using commercial information in some unknown manner. Though our results do not depend on the assumption that this categorization is correct, we used the categorization to determine the optimum sampling rates in different parts of the population. We deemed it likely that the "Catholic" list would in fact contain a high concentration of Catholics, and that therefore the cost of calling and screening these numbers would be relatively low per respondent. The second part of the commercial list (designated as "non-Catholic" below) consisted of all other residents of Cook and Lake counties for whom telephone numbers were available; from this list we selected cases with a lower sampling fraction, as the expected cost per respondent could be expected to be higher as we anticipated a lower proportion of Catholics in this list. Finally, as residents whose telephone numbers appeared on the lists might be expected to differ substantially from residents whose numbers did not appear on the lists, we selected a further sample of numbers by randomly generating telephone numbers from all possible telephone numbers in Cook and Lake counties (designated as "RDD" below). As resolution and screening costs were expected to be much higher for this group (far more non-productive telephone calls), we selected from this frame with a much reduced sampling fraction.
This process of sample design produces interviews considerably more efficiently than sampling with equal probabilities from all potential telephone numbers in the two counties. However, if the Catholics in each of the three sub-frames differ from each other, it is necessary to introduce weights into the analysis in order to have the results reflect the true population structure rather than the configuration of the respondents in the sample.
Principal Investigators
Andrew Greeley, Research Associate, NORC/University of Chicago
There are three weights given in the data set. At a minimum, the baseweight (BSWT) should be used; this adjusts only for the differential probabilities of selection from the three parts of the frame. The weight that should be used for most analyses is called samp_weight (SMPWT). This adjusts additionally for within-household selection; otherwise the estimates will be biased towards the views of Catholics in smaller households. Both pop_weight and samp_weight take into account the various aspects of the design and implementation; the difference is that samp_weight is normalized (standardized) to add to 524 (the achieved sample size) across the sample.
Response Rate
The overall response rate for the survey was 40 percent.