Catholic Poll Methodology

Methodology

The CCP is a bilingual (English and Spanish) omnibus poll using academic-quality telephone polling methods, including multiple call-back attempts and statistical weighting of results using the U.S. Census Bureau’s Current Population Surveys (CPS). The CCP is designed as the preeminent Catholic poll for the United States. Polling the Catholic population is challenging as it is a very specialized sub-sample of the overall population. Even at 23 percent of the adult population, there are typically not enough Catholics interviewed in standard polls of the overall U.S. population to provide for reliable results. Also, some 35 percent of the Catholic population self-identifies as Latino or Hispanic and about 15 percent of interviews with Catholics are typically with a respondent who prefers to speak Spanish. Below, are some of the methodological features used to ensure the CCP is as representative as possible:

The national random sample includes adult, self-identified Catholics in the United States and is typically conducted in January or February each year.

The number of respondents varies from year to year but always includes at least 800 Catholics, for a margin of sampling error of no more than ±3.5 percentage points. National surveys are typically of sample sizes of at least 800 to 1,000 where the overall margin of sampling error is less than ±4 percentage points and allows for sub-group analyses where the margins of sampling error are less than ±10 percentage points. For example, with a total sample of 1,000, a sub-group that is half of the interviews, or 500 persons, the resulting margin of sampling error would be ±4.4 percentage points. The minimum number of respondents within a sub-group that can be analyzed is about 100, resulting in a margin of sampling error of ±9.8 percentage points.

Interviews are conducted in both English and Spanish to provide representation of Hispanic/Latino Catholics. It is not possible to survey the Catholic population without using Spanish language interviewing. For a more complete discussion of how critical and challenging it is to reach a representative sample of the Hispanic/Latino population see the recently published scholarly article by CARA researchers: "How Many Hispanics are Catholic? Reviewing the Evidence." by Paul Perl, Jennifer Z. Greely, and Mark M. Gray in the Journal for the Scientific Study of Religion, vol. 45, issue 3: pgs. 419-436.

The polling period typically runs for approximately two weeks. The CCP is a random digit dial (RDD) poll conducted with Computer Assisted Telephone Interviewing (CATI) and predictive dialing to residential telephone exchanges.

The questionnaire typically takes an average of 12 to 15 minutes of interviewing time. When the phone is answered, an adult member of the household (a "targeted respondent") is randomly selected by asking for the adult residing in the household who most recently celebrated his or her birthday. After a non-threatening "ice-breaker" question, a screening question asks the targeted respondent’s religious preference. If the respondent identifies as Catholic, the interview proceeds. Otherwise, the call is ended. If nobody answers the phone on the first call or if the targeted respondent can not speak at the time of the call, up to seven call-back attempts are made. If a Spanish speaker answers the phone, arrangements are made for a call-back by an interviewer fluent in Spanish.

Results of the CCP are presented with statistical weighting. The weights modify selected demographic characteristics of poll respondents to match CARA’s best estimates of the actual characteristics of U.S. adult Catholics. For example, if the poll obtains fewer males than are represented among the national Catholic population, the weights adjust for this by placing more emphasis on responses of men than women. The following characteristics are used for CCP weighting: gender, generation, race and ethnicity (Anglo, foreign-born Hispanic, native-born Hispanic, African American, and other), and geographic region. CARA’s weights are derived from previous polls that include both Catholics and non-Catholics and data from the Current Population Survey (CPS) of the U.S. Census Bureau.

CARA regularly polls at other times of the year but not using the more affordable omnibus format of the CCP. CARA also polls Catholics nationally using other methods, primarily Knowledge Networks online self-administered surveys.

Results of client commissioned questions are confidential and owned by the CARA client. CARA will not release these results unless requested by the client. However, the core omnibus questions on the demographics of the Catholic population and questions about Catholic background and commitment are publicly released. These are frequently cited in research and by the media. For example see: "Study Sees Church Rebounding From Scandal" in The New York Times, May 18, 2006.