The measurement of religious involvment and some attitudinal covariates in ESS

1
The measurement of religious involvment and some
attitudinal covariates in ESS

• Jaak Billiet
• Member of CCT - ESS
• CeSO - K.U. Leuven
• Bart MeulemanNational Research Fund
• CeSO K.U. Leuven
• Prepared for the mini conference on measurement
  and composite scores held in Barcelona 14-15
  February 2008

2
Outline

• Equivalent Measurement of religious involvement
  and some other variables/values in ESS - R2 (and
  the religious map(s) of Europe)
• - MGSEM end equivalence- composite scores
• Is religious involvement related to social
  attitudes and values?
• Discussion

3

Measurement of religious involvement in ESS

• Religious involvement not the focus of ESS
• RI aspect of social identity diversity
• Mainly to be used as a predictor (explanatory
  variable) for other phenomena as
• - some value orientations - citizenship
  political orientations - attitudes towards
  immigration - social capital - social
  trust
• Aspect in testing theories about social change

4

Measurement

• Questions in core in R2
• - belonging to religion or denomination at
  present (Yes/No)- past belonging (Yes/No) -
  which religion or denomination? (general cntry
  specific)- how religious? (notvery 0-10 p
  scale)- participation in public religious
  services (every day-never 7 p scale) - how
  often pray apart from previous? (every
  day-never 7 p scale)(dropped from R1
  importance of religion 10p)

5
Measurement

• Three constructs for 26 countries, N 47,471 in
  random samples from cntry populations
• (1) Most simple Public religious practice a
  combination of question on actual belonging and
  participation in public religious services

6
Measurement of religious involvement

• (2) Typology used by sociologists of Religion
• Adapted to situation with only 3 indicators
• RLBLG RLBLGE RLGATND TYPE
• NOT YES 2d gen not belong
• NO 1st gen not belong
• daily-montly marginal member
• YES less marginal member
• dayly-holy days member
• In next table (1st2d not belong) versus member
• Colours for kind of dominant denomination KATH,
  PROT, ORTH, MIX, ISLAM

7
Religious map of Europa
marginal not in table
8
Religious map of Europa

• Previous figure diversity in religious past
  (previous dominance) and actual total not
  belongingNot belonging gt belonging in half of
  the countries
• Next figure first versus second generation of
  (total) not belonging and ratio 1st/2ndgives
  idea about recent de-christianisation in EU

9
(No Transcript)
10
Religious map recent (small) dechristiannisation
11
Measurement of religious involvement

• (3) Latent (metric) variable (scale 0 10)
• degree of religiosity 10p
• 0.83 religious
• involvement 0.78 frequency participation
  7p
• 0.89 frequency of
  praying 7p
• Is this measurement equivalent (factorial
  invariant over cntrys)?

12
Equivalent measurement of religious involvement

• Yes, largely equivalent measurement over 26
  countries
• How do we know? statistical tests whether the
  measurements are metric invariant (invariant
  identical slopes in all countries) condition
  for comparing relations btw cntrys and scalar
  invariant equal intercepts (means of
  indicators for cutting point on latent Y
  variable) condition for comparing latent
  means btw cntrys
• Exception Turkey not scalar invariant WHY?

Reason different indicators for Male and Female
required in Islam? See further
13
MGSEM test of scalar metric invariance
14
Parameter estimations for selected model
15
Turkey scalar metric equivalent reason?

• Observation correlations with background
  variables different gender is conversely
  correlates (Male more religious than female,
  unstandardised regression coeefficient 0.614 in
  TU and between -1.445 and -0.711 in all other EU
  countries)
• See also deviation of TU in explained variance in
  religious involvement by background variables
  (gender, urbanisation, age, education, job
  activity)(see next figure)

16
(No Transcript)
17
Closer view on Turkey

• Strange relation between religious and gender
  in Turkey (Islam) different according to
  indicators used
• Frequency of praying (1-7) and how religious
  (1-10)
• Regular practice Female 35.7 ---- Male 80.5
  (duty for Male)
• (definition of member NOT equivalent)

18
Conclusion about Turkey

• Latent scores and composite scores strictly not
  to compare with other EU countries
• One should use smaller set (only two indicators?)
  or find a fundtional alternative for public
  practice
• Rule for the composite scores (W. Saris)
• EU relig_w .249rlgdgr .19rlgatndb
  .421prayb
• TR relig_w .498rlgdgr .068rlgatndb
  .241prayb
• correlation weighted/unweighted 0.993
• Unw Mean/STD 1.468
• We Mean/STD 1.403 (smaller)

19
Mean latent score reference (0) AT
20
Composite weighed scores (factor regr weights)
corrected
21
Measurements of values and attitudes

• values1. Self-transcendence metric invariant

22
Test on scalar and metric invariance
metric invariance
23
selftran_w 0.139v30.137v80.138v190.142v12
0.141v18
24

• 2. Conservation metric invariant

25
(partial) metric invariant
conserva_w 0.122v90.130v200.132v70.1
46v50.126v14
26
(No Transcript)
27

• Attitudes towards immigrants
• 3. not_allow (No willingness to allow immigrants
  into cntry)To what extent do you think
  country should allow people of the same race of
  ethnic group as most county people to come
  in?(allow many to come and live here, allow
  some, allow a few, allow none)How about people
  of a different race or ethnic group from most
  country people? (allow many..allow
  none)How about people from the poorer
  countries outside Europe? (allow many..allow
  none) metric invariant over countries
• not_allow_w .08imsmetn .62imdfetn
  .13impcntr

28

• 4. Consequences (badgood consequences) Would
  you say it is generally bad or good for
  countrys economy that people come to live here
  from other countries?(Bad -- good 11-point
  scale) Would you say that countrys cultural
  life is generally undermined or enriched by
  people coming to liver here from other
  countrys? (underminded---enriched 11-p)Is
  country made a worse or a better place to live
  by people coming to live here from other
  countries (worse----better 11-p) metric
  invariant over countries
• conseque_w .20imbgeco .25imueclt
  .33imwbcnt

29
Is religious involvement at individual level
related to attitudes and values?
30
Table 4. Unstandardized simple regression
coefficients with religious involvement
independent variable and each attitude/value as
ependent variable (ESS R2)
31
Is RI related to social attitudes? main
findings

• Strongest relation between attitudes towards imm
  and religious involvement in SI, IS, GR, (BE,
  PT)(more religious is more negative twds
  immigrants) UK is opposite
• Individual religiosity seems to have most
  positive effect on social attitudes in UK (mixed
  cntry) and most negative in GR

32
Is RI related to social values? main findings

• Strong relations to values in all
  couintriesmainly on conservationreligious hold
  more conservative valuesstrongest relation in
  BE, FI, AT, IS, GR, IE
• Relation with self-transcendence is weaker but
  strongest in FI, IS, SE, IE and UK