A measure is a variable in a dataset that validly represents a theoretical concept. In practice, concepts tend not to be wholly captured by single measures; consequently, measures are usually limited and partial indicators of concepts.
In practice, researchers utilize a theory to logically derive specific hypotheses. A hypothesis should be phrased in terms of concepts that refer to observable phenomenon. A measure provides one observable indication of a concept in the form of a variable from a survey data set. Of course, measures can take many forms, but we restrict ourselves to quantified measures because the ARDA is solely a survey data resource.
If a theory properly distills a clear hypothesis, which in turn is properly stated in terms of observable concepts, which in turn are accurately indicated by available measures, a researcher can analyze the relationship of measures to ultimately provide empirical support or a challenge to common theories.
List of Measures
Several files in theARDA's Data Archive have examples of Measures.
These are measures assessing the degree to which people approve or disapprove of public displays of religion. Typical topics include prayer in schools and the display of religious symbols in public. See Froese and Mencken (2009) for details on the construction a scale incorporating different topics.
A battery of questions was developed to assess traits posited as aspects of an “authoritarian personality.” The battery examines a range of dimensions, but in general it attempts to determine levels of submission to authority and adherence to conventional or traditional values, accompanied by the belief that such values should be enforced on others. A separate personality dimension called “social dominance” has also been proposed in order to explain the sources of prejudice.
These are measures that assess what types of activities occur within the context of worship services in a given religious group. Such practices are often used to delineate between liturgical and ritualistic practices, as opposed to emotive and extemporaneous practices.
This refers to measures that assess the makeup of religious groups with reference to the race or ethnicity of members. See Dougherty and Houser (2008) for detail on the construction of an “entropy index” of racial composition.
This measure assesses the makeup of a religious group with reference to its social class, usually in the form of income or education levels. Given that the measure is aimed at understanding group-level attributes, it is found in data sets at the congregational level or in data sets that ask individuals to estimate certain things about their congregation.
Hood (1975) developed a series of questions based on Stace (1960) that were designed to tap into experiences perceived as transcending cultural and temporal conditions, indicating an individual’s level of mysticism. The original scale has been explored in a substantial amount of empirical research and remains in use currently.
The Herfindahl Index is most commonly used in economics to measure the market share of industries. Applied to religion, it assesses the “evenness” of religious groups pertaining to the overall market of adherents in a geo-political area. Its use to predict attendance or adherence rates has been controversial, since a “mathematical relationship between [the] variables” may cause potentially erroneous results (Voas et al. 2002: 212).
In 2000, Steensland and colleagues proposed a new method for classifying religious tradition which was based on both doctrine and historical changes in religious groups. The schematic divides religious traditions into black Protestant, Catholic, evangelical Protestant, Jewish, mainline Protestant, no religion, and “other” religion. The “other” category functions as a catch-all to reduce missing cases in multivariate analyses with listwise deletion, but should not be substantively interpreted, as it contains a mixture of Eastern religious traditions, Mormons, and everything in between. The classification scheme is created by using variables such as affiliation and denomination to classify respondents.
A question that asks respondents what religion, if any, their spouse ascribes to. This variable allows researchers to investigate how religious capital affects spouse choice, as well as how network religious preferences affect individuals. This variable also potentially allows researchers to approximate which religious traditions are more exclusive with regard to endogamy.
This asks respondents what religious tradition their parents ascribe to and allows researchers to investigate why individuals maintain or change from the religious tradition they were exposed to when younger.
A series of questions with multiple sub-questions asking for respondents’ level of interaction with New Age materials and ideas. These vary, reflecting the diffuse content of the subculture. Bainbridge (2004) suggested that New Age, “paranormal” and UFO beliefs were all empirically distinguishable but related belief systems. In addition to beliefs, there are measures of media consumption and experiences with regard to the New Age.
These variables measure a respondent’s self-assessment of life satisfaction with family life, job, education, etc. Subjective assessments contrast with more “objective” measures such as quality of life metrics, social status, or physical health.
Measurements of a respondent’s level of acceptance or condemnation of contact with those from a different religious of ideological persuasion. These can be measured by asking about different contexts or scenarios.
Images of God are typically measured using a battery of questions inquiring about the traits that respondents attribute to God. Typically these descriptions fall out along underlying dimensions, especially the following: level of engagement or distance from the world, wrath or anger, and love. These dimensions, as well specific combinations of these dimensions, have been found to have a substantial relationship with political and moral ideology.
These variables can be used to explore whether or not respondents believe in a literal creation story or if creation should be taught in schools and whether they accept versions of evolutionary theory presented by contemporary science.
This measures how frequently respondents attend places of worship. It is debatable how much measurement error is present in self-reported attendance, as people tend to over-estimate their participation (see Hadaway et al. 1993; Hout and Greeley 1998; Pressler and Stinson 1998; Smith 1998).
These questions measure how literally respondents read the Bible or other sacred scriptures. It is debatable whether this measure taps more of a belief or an identity. Although certainly some of both, the latter seems to be more important, as “literalists” from different interpretive communities may disagree on which parts of the Bible are to be taken literally.
What is clear is that this question taps a dimension of religion important for understanding other aspects of religiosity, political attitudes, and views about moral authority more generally. Accordingly this measure has become a standard control, along with service attendance and religious tradition. It is also used as an indicator of fundamentalism. Although some believers claim that belief in Biblical “inerrancy” is different, this too is typically taken to mean “literalism.”
Measures typically include three or four response options. Researchers sometimes employ these as ordinal categories, while others utilize them as nominal categories.
An approximation of how often respondents attended religious services and how religious they considered themselves at a given point in their childhood. This variable is used to approximate how respondents’ levels of religious involvement have changed over time and to estimate levels of religious socialization.