Most of the preceding discussion focused on differences in religious knowledge among members of different religious groups. But what other religious and demographic factors are linked with differences in religious knowledge? The survey shows that religious knowledge is most closely linked with years of schooling. In addition, a variety of religious traits – including overall levels of religious commitment and frequency of reading religious materials – help to explain differences in religious knowledge. Among demographic groups, men perform better than women, whites do better than blacks and Hispanics, and people who live in the East, Midwest and West score better than people in the South.
This final section of the report begins by examining differences in knowledge about religion first by education, then by religious traits and finally by other demographic and political traits besides education. It concludes with an analysis of the relative strength of the impact of each of these factors on religious knowledge while holding the other factors constant.
Education Drives Knowledge
The survey results are clear: People with higher levels of education tend to be more knowledgeable about religion. College graduates get an average of 20.6 out of 32 religious knowledge questions right. Within this group, people who have a post-graduate degree (such as a master’s degree, doctorate, medical degree or law degree) do even better, averaging 22.2 out of 32 questions correct, with 30% falling in the top 10% of all respondents (answering 26 or more questions correctly). Those with bachelor’s degrees get roughly 20 questions right on average.
Scores are significantly lower among respondents with less education. Among people who have some college experience (but no degree), the average number of correct answers is 17.5. Those whose formal education ended with a high school diploma get 13.7 questions right on average. At the low end of the education spectrum, people who did not complete high school give an average of 10.7 correct answers on religious knowledge questions, roughly half as many as those with post-graduate degrees.
In addition to asking about overall levels of education, the survey included several questions about religious education specifically. One such question asked whether people have ever taken a college-level religion course. Those who say they have taken a religion course get 22.1 questions right on average, compared with 17.9 for those who have at least some college education but who have not taken a religion course at that level, and 12.8 for those who have not been to college.
The survey also asked respondents whether they attended public, private religious or private non-religious schools in their childhood. Those who attended private school score better than those who attended public school by more than two questions
on average. However, there is no significant difference between those who attended private religious schools (17.8 questions right on average) and those who attended nonreligious private schools (18.5 questions right on average). Those who went to public schools get an average of 15.5 questions right.
Those who say they regularly attended (at least once a week) religious education classes or participated in religious youth groups as children get an average of 16.6 questions right. Those who participated in religious education and youth groups more sporadically (monthly or yearly) get fewer questions right (15.9 on average), while those who say they seldom or never participated in these activities answer the lowest number of questions correctly (14.3).
Religious Traits and Religious Knowledge
In this study, a person’s degree of religious commitment is measured by combining two questions, one asking about the importance of religion in the person’s life and the other asking how often the person attends worship services. Religious commitment has a complicated relationship with religious knowledge. People with a high level of religious commitment – those who describe religion as “very important” in their lives and who say they attend worship services at least once a week – answer 17 of the 32 religious knowledge questions correctly on average. People with a low level of religious commit-ment perform at the national norm, answering an average of 16 questions correctly. But, interestingly, people with a medium level of religious commitment get fewer questions right (14.8 on average) than people with either high or low commitment.
Moreover, statistical analyses that look simultaneously at different factors in religious knowledge reveal that the most pronounced differences are between people with the highest levels of religious commitment and everyone else. In other words, once education, religious affil-iation and other factors are controlled for, differences in religious knowledge be-tween those with low and medium levels of religious commitment mostly go away. But people with a high level of religious commitment continue to display higher religious knowledge, even when other factors are held equal.
The same analyses also show that the strength of the link between religious commitment and religious knowledge depends on the particular domain of knowledge. On questions about the Bible, for instance, there is a simple, straightforward relationship – people with high levels of religious commitment get more questions right, while those with lower levels of commitment get fewer questions right. But on questions related to world religions, there are no significant differences associated with religious commitment.
The survey shows that reading and talking about religion are related to higher levels of religious knowledge. People who say they read Scripture at least once a week, for instance, get significantly more questions right on average than those who read Scripture less often. The same pattern is seen in frequency of reading books (besides Scripture) about one’s own faith. People who are affiliated with a religion and read books about their faith at least once a week get an average of 18 (out of 32) questions right, compared with 14.5 for those who say they seldom or never read books (other than Scripture) about their faith. Similarly, reading books at least occasionally about religions other than one’s own is linked with higher levels of religious knowledge compared with those who seldom or never read about other faiths. And people who say they frequently discuss religion with their family and friends get an average of 16.9 questions correct, which is significantly higher than the number answered correctly by people who discuss religion less often.
Beliefs about God and the Bible are also associated with religious knowledge. People who say they do not believe in God or a universal spirit perform relatively well on the survey, getting an average of 18.7 questions right. By comparison, those who believe in God get significantly fewer questions right (those who say they believe in God with absolute certainty get an average of 16 questions right, while those who say they believe in God but are not absolutely certain about God’s existence get 15.3 right). This is consistent with the survey’s finding that people who describe themselves (in response to a separate question) as atheists and agnostics score better overall than people affiliated with many religious faiths.
In general, people who say they do not believe that the Bible is the word of God score higher on the survey than do those who do view the Bible as the word of God. Respondents who say the Bible was written by man and is not the word of God get 18 questions right, on average. Those who say the Bible is the word of God but should not be taken literally get an average of 16.3 questions right. And those who say the Bible is the word of God and should be taken literally, word for word, get an average of 14.5 questions right.
People who have switched religions since childhood7 score better on the survey (16.9 questions right on average) than people who hold the same faith now as they did when they were growing up (15.7 questions right on average). People who are married to someone from a different faith get 17 questions right on average, while those who are married to someone who shares their faith get 16.5 questions right on average.
Demographics, Politics and Religious Knowledge
A number of demographic characteristics are closely correlated with levels of religious knowledge. Whites, for instance, score higher than both blacks and Hispanics. Whites get an average of 17 questions right, compared with 13.8 for blacks and 13.3 for Hispanics. Consistent with this, those who were born in the United States generally score higher (16.3 questions right) than those who were born abroad (13.8).
Men get more questions right on average than women (16.7 vs. 15.3). People living in regions other than the South get more questions right on average than people living in the South (16.8 right for people living in the West, 16.3 in the East, 16.2 in the Midwest and 15.4 in the South). And those in the middle age groups (ages 30-64) get more questions right than the oldest adults (those 65 and older). Those age 18-29 get fewer questions right than those age 30-49, but are not significantly different compared with those age 50 and older.
People who have not had a child score significantly higher on the survey than parents do – an average of 17.2 questions right for those who have not had children vs. 15.5 for those who have had children. (A similar pattern is seen on the general knowledge questions.)
People who are married or living with a partner tend to score higher than those who are single (16.6 questions right vs. 15.2). However, this apparent difference is actually a function of differing levels of education and other religious and demographic characteristics; once these other factors are controlled for, there is no significant difference in the religious knowledge levels of married people and those who are single.
Politically, Republicans (17.5 questions right, on average) score significantly higher than Democrats (15.9) and independents (15.7). But here again, these apparent differences actually have a lot to do with other traits; once education, religious affiliation and other factors are taken into account, Democrats actually perform somewhat better than Republicans. Similarly, people who describe themselves as liberals answer about the same number of questions correctly as self-described conservatives (16.7 on average for liberals, 16.6 for conservatives). However, as will be described in more detail in the next section, once other factors are taken into account, liberals show slightly higher religious knowledge than conservatives do.
Assessing the Relative Impact of Religious and Demographic Factors on Religious Knowledge
The results of the survey suggest that there are many traits that seem to be linked with levels of religious knowledge – including demographic factors, such as age and race, as well as religious factors, such as religious affiliation and reading about religion. But which of these traits are strongly related to greater knowledge, and which are related only tangentially, if at all? The remainder of the report attempts to answer this question, based on a technique known as multiple regression analysis.
The multiple regression analysis begins with a statistical model that includes seven main variables: religious affiliation, religious commitment, education level, race, gender, age and geographic region. It considers the impact of each of those variables, one at a time, holding the other six constant. Then the analysis holds all seven of the main variables constant and adds into the model, one by one, a number of other possible factors in religious knowledge, such as frequency of Scripture reading, interfaith marriage and belief in biblical literalism. This produces a picture of how much each factor contributes to religious knowledge independent of the other variables. Put somewhat differently, this method of analysis imagines a survey respondent who is completely typical in all ways but one and calculates the impact of that one factor on the respondent’s level of religious knowledge.
These analyses confirm that educational attainment is far and away the single leading predictor of higher religious knowledge. Having a college degree is associated with an additional 4.3 correct answers (above the national average) even after religious affiliation and other demographic characteristics are taken into account. By comparison, having a high school education or less is linked with 2.9 fewer correct answers. Thus, with all else held equal, having a college degree makes a difference of 7.2 additional correct answers (out of 32 possible on this survey) compared with someone with a high school education or less.
Religious affiliation also remains a good predictor of religious knowledge in these models. Even after education, race and other factors are taken into account, atheists/agnostics, Jews and Mormons perform better than other religious groups on this survey. Atheists and agnostics get an additional 2.9 questions correct compared with the national average, Jews do 2.3 questions better than average and Mormons get 1.9 more questions right over the national average. Atheists/agnostics, Jews and Mormons are followed by evangelicals, who do 0.9 questions better compared with the national average. By contrast, mainline Protestants, Catholics and those whose religion is nothing in particular trail other groups even after demographic factors are taken into account.
Holding other factors constant, people with high religious commitment – those who attend worship services frequently and consider religion to be very important in their lives – get one additional question correct (compared with the national average). By comparison, those with medium and low levels of religious commitment each get 0.5 fewer correct answers.
People who took a religion course in college answer about three more questions correctly (out of 32) than those who have not taken a college-level religion course, even after controlling for overall levels of educational attainment. People who attended religious education classes or participated in religious youth groups at least once a week growing up get nearly two additional questions right, compared with those who did so seldom or never. Those who attended private religious schools as a child get an extra 1.7 questions right, compared with those who attended public school. However, there is no statistically significant difference in the scores of those who attended private religious schools and those who attended private nonreligious schools.
Reading and Talking About Religion
Holding other factors constant, people who read Scripture at least once a week get roughly two additional questions right on average compared with those who seldom or never read Scripture. Frequent Scripture reading has a stronger impact on knowledge of the Bible than on knowledge of world religions.
Among people who are affiliated with a religion, those who regularly read other books (besides Scripture) about their own faith get an additional 2.3 questions right compared with those who seldom or never read other books about their own faith. By a similar margin, people who say they regularly read books or visit websites about religions other than their own also do better than those who do not read books about other faiths. People who talk frequently about religion with family and friends get roughly two additional questions right compared with those who rarely or never discuss religion.
Scriptural Literalism and Belief in God
Holding other factors constant, people who say Scripture was written by men answer nearly three additional questions correctly, compared with those who say Scripture is the word of God and is to be taken literally, word for word. People who do not believe in God score slightly better (by 0.7 questions, out of 32) than people who believe in God but express doubts about this belief.
Interfaith Marriage and Religious Switching
After controlling for other factors, being married to someone from a different religion adds roughly one correct answer, on average, to a person’s score on the survey compared with people who are married to someone who shares their religion. Religious knowledge also tends to be higher (by one question) among people who have switched faiths than among those who belong to the religion in which they were raised.
Holding other factors constant, blacks and Hispanics score significantly lower on the religious knowledge survey than do whites (by 2.5 questions for blacks and by 2.1 for Hispanics, out of 32 questions in total). Immigrants give 1.3 fewer correct answers than the U.S.-born do, all else held equal. Men score better than women by 1.4 questions. Those who live in the East, West or Midwest all score better (by about one correct answer) than those who live in the South. The oldest group in the population scores significantly lower than other age groups, by about one question. People who have never had children get 1.3 extra questions right compared with those who have had children.
Partisanship and Ideology
Democrats score better than Republicans on the survey by a slim but statistically significant margin of half a question, all else held equal. Self-described liberals score better than self-described conservatives by 0.8 questions.
7 Switching was determined by asking respondents their present religion and the religion in which they were raised as a child, then comparing the two. (return to text)
Photo credit: Eric Swanson/Corbis