1 / 28

Measuring Religion in Britain: Richness versus Parsimony

Measuring Religion in Britain: Richness versus Parsimony. Siobhan McAndrew British Religion in Numbers University of Manchester. Need for efficient measures. Survey costs Right to privacy Typical official questions: Which, if any, is your religion?

robbin
Download Presentation

Measuring Religion in Britain: Richness versus Parsimony

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Measuring Religion in Britain: Richness versus Parsimony Siobhan McAndrew British Religion in Numbers University of Manchester

  2. Need for efficient measures • Survey costs • Right to privacy • Typical official questions: • Which, if any, is your religion? • Do you currently practise this religion?

  3. Survey data issues • Survey data are of the following types: • Continuous • Count • Ordinal • Nominal (e.g. red, blue, yellow; yes, no). • Rich array of measures in 2008 ISSP Religion III – mostly ordinal or nominal But: • Is religiosity a continuous trait (we are more or less religious)? • If yes, are there different dimensions of religiosity? • Or are there qualitatively different religious classes?

  4. Measuring religiosity • Typology approach: categorise respondents as • Religious: believe, affiliate, and attend.................28% • Fuzzily-faithful: believe, affiliate, or attend...........39% • Unreligious: do not believe, affiliate or attend.....33% • Religiosity scale based on 14 items

  5. Example of mixture distribution

  6. Example of mixture distribution ‘Unreligious’ ‘Religious’ ‘Fuzzy’ = mixture of two distinct types?? “We observe these with error”

  7. Latent classes: finding ‘ideal types’ • Social class • Cultural classes: omnivores and univores • Consumer types Religion: • Latent classes of Muslim women: religious, traditional, secular, fundamentalist (Blaydes & Linzer 2008) • Orthodox, non-orthodox, non-religious - Hagenaars & Halman 1989 • 11 belief systems (e.g. Strong believers, Secular, God + Sin, Strong but no hell/devil, etc.) - Owen & Videras 2006

  8. Or... • Is religiosity not a class but a continuous trait: • we are more or less religious • may be different types of religiosity (doctrine, practice, belief) • may be skewed, twin-peaked, fat-tailed... • If there are a small number of religious classes, continuous measures imply we are observing them with error (hence twin-peaked graph) • Or, if religiosity is a continuous trait, we may be observing that religiosity is distributed in interesting ways • Conceptually different.

  9. Structure of religiosity may relate to theory • Secularisation theory: moving from one class to another • can label classes as more or less religious • Postsecularist or rational choice approach: religiosity is rich and complex phenomenon. Distribution of religiosity may be changing in line with the particular dynamics of religious change • different dimensions of religiosity • factor scores are standardised

  10. Mean ‘religiosity’

  11. Mean ‘religiosity’

  12. Choice of ‘manifest variables’ • Belief in Nirvana • Belief in miracles • Belief in reincarnation • Religion important in upbringing • Confidence in churches • Inner peace • Sciencevs religion • Truths in many religions • R is adherent • Church attendance • Church activities • Prayer • R is spiritual • Belief in God • Belief in heaven • Belief in hell • Belief in afterlife

  13. Why? • To cover variety of potential religiosities: • Belonging/background • Activity • Vicarious religion • Orthodox/heterodox beliefs • Broadly applicable to most religious groups • Too much on belief? • Too Christian?

  14. Exploratory cluster analysis • Two-step cluster procedure in SPSS • Determines number of clusters from the data Secular: 26% Fuzzy: 44% Religious: 30% Secular: heaven, afterlife, miracles, hell Fuzzy: heaven, afterlife, hell, miracles Religious: heaven, spiritual, prayer, hell

  15. However... • SPSS throws away information • Missings (1683 of 2250!!) • treats ordered responses as not ordered • Doesn’t incorporate predictors such as age, education, sex • This makes it more likely that the clusters found aren’t ‘real’ - e.g. grouping together the definitelys, probablys and definitely nots.

  16. Exploratory factor analysis • CFA assumes both the trait and the responses are continuous & normally distributed (hmm....) • 2 main factors, 49% of variance: • Traditional religiosity (attends church, prayer, church activities, belief in God) – 28% of variance • Afterlife/supernatural (afterlife, reincarnation, Nirvana, afterlife, heaven, hell) – 22% of variance

  17. Factor 1: “Traditional religiosity”: negative skew

  18. Factor 2: Afterlife/supernatural (Weakly but significantly correlated with factor 1.)

  19. Predictors of religiosity • Traditional religiosity: being older, being female were significantly associated with this factor. Marital status, education, socio-economic status and ethnicity were not. • Afterlife/supernatural dimension: being younger, being female was significantly associated with this factor. Being non-white (p = 0.06) and not holding a professional post (p = 0.09) on borderline. Other controls not significant.

  20. Latent class analysis • Incorporates information from co-variates (e.g. sex, education, socio-economic status) to uncover latent class structure • poLCA also uses information even where the respondent did not answer all the questions – fewer cases missing. • Four latent classes found

  21. Belief/practice patterns Class I: belief in God, adherence, religion good for inner peace, heaven........................................................religious moderates(31%) Class II: no or never to belief in God, hell, heaven, Nirvana, reincarnation, church activity, prayer, spirituality, belief in the afterlife, church attendance, truth in any religion.............................................................strong secularity(24%) Class III: belief in God, truth in one religion only, confidence in churches, too much science, religion good for inner peace, belief in the afterlife, church attendance.....................strongly religious (17%) Class IV: no or never to Nirvana, miracles, church attendance, prayer, church activity, heaven..........................................secular/fuzzy (28%)

  22. Predictors of class membership Compared with Class I (religious moderates): The strongly secular: significantly more likely to be male, younger, non-ethnic background. Education, income NS. The strongly religious: significantly more likely to have an ethnic background. Not having A-levels/some college on border of significance. Age, gender, income NS. The moderately secular: significantly more likely to be younger and not earn a top-quartile income. Male, non-ethnic background on border of significance. Education NS.

  23. Uncovering religiosity? • Need similar exercise to examine religiosity as a trait (latent variable analysis) • Intuition there is also a small number of classes which are qualitatively different, which structure religiosity: but religiosity is a continuous trait • Testing requires combined latent class and latent trait model • Danger that I am searching for method that gets the results I want.

  24. Parsimony vs richness? • KISS. Latent variable analysis still tricksy • Problems with cluster analysis: FA sensitive to choice of variables, treatment of missing data • Three-fold typology approach is cheap and simple to understand • May miss complexity of the broad middle or different dimensions of religiosity • Missing responses to religious questions a chronic problem • Surveys are expensive.

More Related