Correlation is a mathematical term concerned with examining and expressing the observed relationship between two factors. Among young people there is an observed positive correlation between age and height. As young people grow older they tend to grow taller. There is a clear correlation, but the correlation is not perfect. Some 8-year-olds may be taller than some 12-year-olds.
There is a well-established scientific research tradition in the psychology of religion that is concerned with testing and mapping the correlates of religion during childhood and adolescence. Such a scientific approach is able to adjudicate between conflicting theories and opposing points of view. Typically, studies of this nature attempt to measure at least two variables (say, sex and frequency of church attendance or religious beliefs and attitude toward abortion) among a well-defined group of people. The correlation coefficient is a measure of the strength with which the two variables are or are not related.
Correlations vary between 0 and 1. A correlation of 0 demonstrates that there is no association between the two variables, while a correlation of 1 demonstrates a perfect relationship. In the social scientific study of religion a number of important relationships between religion and other variables are reported around the level of 0.2, expressed as follows, r = +0.2 or r = -0.2. The positive correlation shows that as one variable increases (say, frequency of church attendance), another variable increases (say, altruistic behavior). The negative correlation shows that as one variable increases (say, frequency of church attendance), another variable decreases (say, acceptance of substance use). Correlations between religious variables (say, prayer and church attendance) are often reported at around, r = +0.5. This means that frequency of prayer increases in line with frequency of church attendance but that the relationship is not perfect. Some people who pray never go to church, while some people who go to church never pray.
Correlation coefficients also need to be tested against probability levels, expressed in studies as p. The probability level checks whether the relationship could have occurred by chance or whether it is "statistically significant." The three key levels of significance are expressed as p < .05 (THE RELATIONSHIP COULD HAVE OCCURRED BY CHANCE 5 TIMES IN A 100), P < .01 (THE RELATIONSHIP COULD HAVE OCCURRED BY CHANCE JUST ONCE IN 100 TIMES), AND P < .001 (THE RELATIONSHIP COULD HAVE OCCURRED BY CHANCE JUST ONCE IN 1,000 TIMES). THE LOWER THE PROBABILITY OF THE RELATIONSHIP OCCURRING BY CHANCE, THE GREATER IS THE LEVEL OF CONFIDENCE THAT CAN BE PLACED IN THE FINDINGS.
Correlational studies cannot of themselves determine the direction of causality in a relationship. Often, however, there are good theoretical reasons for postulation of a direction. For example, if religious beliefs were found to go hand in hand with attitudes toward abortion, it is more plausible to suggest that religious beliefs influence attitudes toward abortion than to suggest that attitude toward abortion influences religious beliefs. Current research is able, on the basis of such theories, to distinguish between two kinds of correlates of religion: the factors that help to shape religiosity and the factors that are, in part, shaped by religiosity.
According to the research findings, there are three well-established predictors of the religious and spiritual development of young people in Western Christian societies (where the majority of the research has been conducted): sex, age, and parental influence. Consistently females are found to be more religious than males. Religiosity declines during the years of schooling. Parental influence is a determining factor on whether children and adolescents show an interest in religion and spirituality. Other significant predictors include personality, type of school attended, and the peer group.
According to the research findings there are also well-established consequences of young people being religious. These include a less permissive attitude toward drugs, alcohol, and tobacco, a greater sense of empathy for others and prosocial behavior, a clearer sense of purpose in life, and often a greater sense of personal happiness.
Correlational studies also have to be aware that some apparent relationship may in fact be caused by a third factor. For example, sometimes the correlation between church attendance and another variable may be a consequence of sex differences between males and females, caused by the simple fact that women are more likely than men to attend church. For that reason it is wise to take sex differences into account, either by computing correlations for males and for females separately or by using "partial correlation" procedures, which are able to "partial out the influence of sex differences before computing the correlation coefficient. In conclusion, the empirical evidence demonstrates that there are verifiable and scientifically grounded patterns of relationships that predict and can be used to promote positive religious development and positive youth and human development.