1. Childlessness trends, religious trends: are they associated? Marion Burkimsher Affiliated to the University of Lausanne

4. Puzzles… The average number of children of women born in 1960 recorded in the Swiss census of 2000 was 1.73. But the mean TFR for the period 1980-1999 was 1.53! Why the difference? In Georgia young people are significantly more religious than older people: how unusual is this pattern? In Switzerland around 20% of women aged 40 are childless. But of those who have ‘no religion’, the proportion is 38%, whereas for the Orthodox/evangelicals the proportion is only 12% In Georgia, the association is the opposite: religious women are more likely to be childless (and more highly educated)

5. Structure of presentation Childlessness trends: Data sources Summary of trends Religious trends Data sources Summary of trends Association of childlessness and religiosity Across Europe In Switzerland and Georgia Further work

6. Published paper on religious trends: Burkimsher, Marion. 2014. Is Religious Attendance Bottoming Out? An Examination of Current Trends Across Europe. Journal for the Scientific Study of Religion 53.2 (2014): 432-445. Paper at the “revise and resubmit” stage: Burkimsher, Marion. 2014? Europe-wide fertility trends since the 1990s: turning the corner from declining first birth rates. Demographic Research.

7. Structure of presentation Childlessness trends: Data sources Summary of trends Religious trends Data sources Summary of trends Association of childlessness and religiosity Across Europe In Switzerland and Georgia Further work

8. The challenge of assessing childlessness We cannot know the proportion of women who will end up childless until they are approaching the end of their reproductive life; aged 45-50. For men it is even harder… Period measures of fertility (eg. the Total Fertility Rate, TFR) are deflated because of ongoing postponement in age at first birth. We need an assessment of the cohort fertility rate Childlessness is the complement of the first birth fertility rate (=1-TFR1). But for that we need births by biological birth order, which is not recorded (until recently) in many countries Migration (often affecting young adults in their prime reproductive ages) means that tracking vital statistics (births) within the country may not produce an accurate assessment of childlessness

9. Data sources Vital statistics: as provided by the national statistical offices and complied into the Human Fertility Database (http://www.humanfertility.org); includes fertility rates by age and cohort by each year. Data up to 2010, 15 countries > Projection of cohort TFR1 calculated by the method proposed by Myrskylä, Goldstein and Cheng (2013). The 5 year trend for each age-specific rate is extrapolated for a further 5 years, then the rate is frozen (up to cohort born 1980) Swiss census of 2000: question on number of own children European Social Survey: Wave 3 in 2006 asked question on number of own children

10. Projected rate of childlessness: declining Data source: HFD

11. Projected rate of childlessness: stability Data source: HFD

12. Projected rate of childlessness: increase Data source: HFD

13. Structure of presentation Childlessness trends: Data sources Summary of trends Religious trends Data sources Summary of trends Association of childlessness and religiosity Across Europe In Switzerland and Georgia Further work

14. The challenge of assessing religiosity Religiosity can be assessed in various ways: affiliation, attendance (prayer, belief, importance/relevance to life)… Any dimension of religiosity can change for an individual over time (whereas childlessness can only change once!) Religiosity is self-reported, so open to social norms in reporting. Often someone other than the individual completes the questionnaire / census form Religiosity can be associated with different behaviours and social- demographic characteristics in different ways, in different places and at different times

15. Data on religiosity European Social Survey and World/European Values Survey: attendance and affiliation Swiss census of 2000: affiliation Swiss Household Panel: affiliation, attendance (prayer) Generations and Gender Survey (Georgia): attendance (and affiliation)

16. 1. Cohort differentials in religious attendance as in 2012 (mostly) …but are the differences caused by age effects (people becoming more religious as they get older…)?

17. Inter-cohort comparison of religious attendance: decline Data source: ESS 2012, Austria ESS 2010

18. Inter-cohort comparison of religious attendance: stability Data source: ESS 2012

19. Inter-cohort comparison of religious attendance: growth Data source: GGS 2006

20. 2. Religious attendance of a cohort over time > Trends in religious attendance of the Baby Boom cohort (born 1950-1981) from 1990-2012 …but are the changes caused by period effects (secularisation / revival) or age effects?

21. Trends in attendance rates of 1950-1981 cohort: decline Data sources: WVS, EVS, ESS

22. Trends in attendance rates of 1950-1981 cohort: stability Data sources: WVS, EVS, ESS

23. Trends in attendance rates of 1950-1981 cohort: growth Data sources: WVS, EVS, ESS, GGS

24. 3. Individual flux in religiosity

25. 1 Never 2 Only family ceremonies 3 Only religious celebrations 4 Religious celebrations & family events 5 Few times/year 6 About once/month 7 Every 2 weeks 8 Once a week 9 Several times a week 1 Never 2 Occasional 3 Regular Recoding religious attendance (…any religion)

26. 1 Protestant / Reformed 2 Roman Catholic 3 Christian Catholic 4 Other Christian 5 Jewish 6 Muslim 7 Other 8 No religion or denomination 1 Protestant 2 Catholic 3 Other 4 None Recoding religious affiliation

27. Year-on-year change in religious attendance Data source: SHP 1999-2009

28. Individual flux in religious attendance 1999-2009 Respondents with data in >1 wave (N=9363) Data source: SHP 1999-2009

29. Comments The proportion of respondents who have stayed at the same level of religious attendance across all waves is 39% (for all respondents who have valid responses for >1 year) The proportion of respondents who have changed level of attendance at some stage is 61% The proportion of respondents who have at some time been regular attenders is 36% - cf. to annual average attendance rate of ~24% (includes respondents with only 1 wave of data) The proportion of respondents who at some stage have been ‘never’ attenders is 62%

30. Change in individual affiliation 2004 to 2009 Respondents with valid data in 2004 AND 2009 92% kept same affiliation (N=4690) Data source: SHP 1999-2009

31. Comments Religious affiliation is considerably more stable than level of attendance (or frequency of prayer) The main trend has been towards non-affiliation, slightly more marked amongst Catholics

32. Structure of presentation Childlessness trends: Data sources Summary of trends Religious trends Data sources Summary of trends Association of childlessness and religiosity Across Europe In Switzerland and Georgia Further work

33. Data source: ESS 2006 Childless proportion by religiosity: M&W aged 25-59 ordered by largest differentials Countries without birth order-specific data in HFD in ( )

34. Data source: SHP biographical data Primary determinants of childlessness in Switzerland

35. Data source: SHP biographical data Primary determinants of singleness in Switzerland

36. Women who were childless, by religion: Switzerland Data source: Swiss census 2000 38%

37. Women who were childless, by religious attendance: Georgia Data source: GGS 2006

38. Structure of presentation Childlessness trends: Data sources Summary of trends Religious trends Data sources Summary of trends Association of childlessness and religiosity Across Europe In Switzerland and Georgia Further work

39. Burning questions 1: What are the characteristics of people who change their level of religious involvement (to become more or less religiously involved)? Compare Switzerland and Georgia…. Many factors to consider… education, politics, traditionalists/modern, rural/urban, happiness/satisfaction As religiosity can change over time, which is the more usual pattern (in Switzerland): 1. becoming more religiously active > (marriage) > child OR 2. having a child > becoming more religiously active (need a long-running panel survey, like SHP)

40. Burning questions 2: Changing ‘meaning’ of religious attendance? Ritual (duty) >community Change in ‘meaning’ of childlessness? Alternative calling / health issues > for everyone > constraints / choice (possible with contraceptives) Trends are going up/down > change in behaviour Who is changing their behaviour / embracing change?

41. Burning questions 3: Are both fertility choices and religious behaviour determined by personality? (Can check for Switzerland using SHP data) Do those who take on ‘new’ behaviour norms have specific personality attributes? (But does personality vary over time / space / for the individual?) Inter-cohort differentials reflect change in opinions between generations > influence of (potential) grandparents… What avenues to explore to discover general ‘rules’ that work across time / countries / different trends?

42. Childlessness trends, religious trends: are they associated? In a static way childlessness and religious affiliation and practice are strongly associated (though not always in the same direction) But we cannot say - yet - whether the trends in religiosity and fertility are necessarily linked And if the trends are linked, then what is the underlying mechanism?

43. Thank you! For further questions and comments please contact me: drmarionb@gmail.com

45. Data source: HFD Association of change in childlessness and change in total fertility: trends from 1970-1970 cohorts

46. Proportion of women who were childless in Switzerland, 2000 Derived from vital stats From census 16.5% 20.7% Data source: vital statistics & census 2000

49. Religious attendance of young people (18-29): decline Data sources: WVS, EVS, ESS

50. Religious attendance of young people (18-29): stable/growth Data sources: WVS, EVS, ESS

52. 1Never 2 Few times a year 3Once a month 4>once a week 5 Daily/almost daily 1 Never 2 Occasional 3 Frequent Recoding prayer frequency (prayer outside religious services)

53. Individual flux in frequency of prayer 1999-2009 Respondents with data in >1 wave (N=9320)

54. Comments  The proportion of respondents who have stayed at the same level of prayer frequency across all waves is 50% ->Level of prayer is more stable than level of religious attendance  The proportion of respondents who have changed frequency of prayer at some stage is 50%  The proportion of respondents who have at some time have prayed frequently is 56% (cf. to annual average proportion of respondents who pray frequently ~45%)  The proportion of respondents who at some stage have ‘never’ prayed is 47%  The proportion who never attend (17%) is the same as the proportion who never pray (also 17%). However, of the respondents who regularly attend, 5% never pray! And of those who frequently pray, 22% never attend religious services!

55. Definition of variables Age Group – four decade-long cohort groupings based on year of birth: 1940s, 1950s, 1960s, 1970s (see discussion elsewhere on why older and younger cohorts were not included) Region – seven regions of Switzerland: Lemanic, Mitteland, NorthWest, Zich, East, Central and Ticino. It is the current place of residence of the respondent. Rural-Urban – three groupings defined by type of current commune of residence of respondent, formed from the 9 groups coded in the survey. Urban includes centres and industrial/tertiary communes. Suburban includes communes defined as suburban, rich and peri-urban. Rural includes those defined as rural and commuter communes, agricultural, agricultural/peripheral and tourist areas. Foreign nationalities – 6 different foreign nationalities were studied, those with most representation in Switzerland: German, ‘ Anglo ’ (British and American), Italian, French, Spanish and Turkish. A score of zero indicated that neither the respondent or their parents had that nationality as their first, second or third nationality. A score of 1 meant that one of those three individuals (ego, mother or father) included it as one of their nationalities; a score of 2 meant that 2 of the people did; a score of 3 indicated that all three had that nationality. Swiss nationality ( ‘ Pure Swiss ’ ) – a two-way dichotomy. If the respondent and both parents had Swiss nationality and no other nationality, then the person was considered ‘ pure Swiss ’ ; if any of the three people had any other nationality, then they were considered not ‘ pure Swiss ’. Educational level – The original survey classified respondents by levels 0 to 10, plus several for special education situations. For this analysis, the level ‘ Basic ’ was considered to be levels 0 to 4 of the original groupings, plus codes -6 and -4 for special education. ‘ Intermediate ’ covered groups coded 5, 6 and 7, and ‘ Advanced ’ as groups 8 to 10. Catholic/Protestant influence. To be considered ‘ Catholic ’ the respondent had to classify themselves as either ‘ Roman Catholic ’ or ‘ Christian Catholic ’ ; if they attend religious services a few times a year or just for religious festivals, they were considered as ‘ occasional ’ (level 1); if at least once a month, then as ‘ practising ’ (level 2). There was a similar grading for ‘ Protestants ’ ; they had to define themselves as ‘ Protestant or Reformed ’. If Religious – a two-way dichotomy. For a person to be ‘ not religious ’ they had to either say they had no confession or religion or never attend any religious services or only for family ceremonies. For a person to be considered ‘ religious ’ they had to both classify themselves as having some confession and attend religious services at least for some religious festivals. If Sibling – a two-way dichotomy. Considered positive only if the respondent has lived with a sibling, otherwise negative. Trauma – this quantified the number of traumatic events a person had experienced: separation/divorce of parents; death of mother or father before the respondent was 18. They were summed, so a maximum score of 3 was possible (though in reality a maximum of 2 was observed in the sample). Background – scores were awarded for mother ’ s and father ’ s educational level plus the status of the father ’ s job (of the original 13 levels of parent ’ s education, the bottom 2 levels were grouped into ‘ low educational level ’ and the top 4 into ‘ high educational level ’ ). A score of 0 (for low level of parent ’ s education and low status of father ’ s job) to 6 (highly educated parents and high status of father ’ s job) was possible. Sociability – 4 variables were combined and the score from each was summed. From the frequency of meeting with friends, then a score of 2 was given if this was at least once a week or every day; a score of 1 was given if the frequency was at least once per month. A score of 1 was added if the respondent was an active member of a sport/leisure club, a cultural association and/or an ‘ association caritative ’. Although other types of associations were listed in the questionnaire, they had fewer participants, so were not considered. Politics - the original coding for responses gave scores of 0 to 10. A score of 0 and 1 were recoded as ‘ left wing ’ ; 2 and 3 ‘ socialist; 4, 5 and -4 ( ‘ no strong political leanings ’ ) were coded as ‘ centre ’ ; 6 and 7 as ‘ conservative; 8, 9 and 10 as ‘ right wing ’ ; and other (negative) codes as ‘ unknown/undeclared ’. If Lived Abroad – a 2-way dichotomy for any period lived abroad.

56. Lerch,Mathias, Michel Oris, Philippe Wanner, Yannic Forney (2010). “Religious Affiliation and Mortality in Switzerland, 1991- 2004”, Population-E, 65; 02, June 2010, pp 217-250 Roman Catholic ProtestantOther Christian church Other religion No religion Life expectancy at birth (males) 76.575.676.874.375.7 Life expectancy at birth (females) 82.581.783.677.479.4