Secularization in Poland and Spain after the democratic transition: A cohort analysis

Abstract

Spanish

French

The objective of this article is to determine the extent to which the evolution of religiosity in Spain and Poland in their post-democratic transition periods has been affected by the process of generational replacement. For Spain data are drawn from several surveys carried out by the Spanish Centre for Sociological Studies (CIS) between 1980 and 1996. For Poland the data come from the Polish General Social Survey and ISSP covering 1992–2008. Results show two radically different patterns of religious change. The fall in religious practice in Spain observed throughout the first 16 years after the political transition was due mostly to the inter-cohort change that affects each new generation born after 1950. In the case of Poland, post-transition change is less marked and due mainly to decline in religious practice on the individual level. The study also observes that the cohorts of Poles born during and after the fall of communism are significantly less religious than older cohorts.

Keywords

Cohorts Poland post-authoritarian societies religiosity Spain

There is a great deal of empirical evidence that demonstrates that in most European countries institutional religion has experienced a sharp decline over the past few decades (Halman and Draulans, 2006; Voas, 2009). The advance of secularization in many European societies is distinctive in light of the still vigorous religious life on other continents (Bruce, 2002; Caplow, 1985; Filali-Ansary, 1996; Sasaki and Suzuki, 1987). This has led some authors to suggest a reformulation of the secularization thesis, indicating that Europe is not so much a pioneer of a universal phenomenon, but rather an exception (Berger, 1999; Berger et al., 2008; Davie, 2007). In any case, it should also be noted that comparative studies have revealed that the progress of secularization varies greatly across European countries, so the decreased social significance of religion on this continent cannot be considered homogeneous or unilinear (Bruce, 2002; Casanova, 1994; Pickel, 2009). Over the last few decades the field of sociology of religion has put considerable effort into providing an accurate explanation for the diversity of secularization trajectories in Europe. Some researchers have suggested that political transformations determine the framework within which religious life develops on a national level (Davie, 2000; Martin, 1978). Therefore, it is particularly interesting to analyse the development of religiosity in European countries that have experienced profound changes in their political systems, as is the case in Poland and Spain.

Until now analyses of religiosity in both countries have predominantly centred on the immediate effects (period effects) of the change in political context linked to the transition to a democratic system (Mariański, 2011; Pérez-Agote, 2010; Requena, 2005; Requena and Stanek, 2012; Tomka, 2011). However, societal change can occur not only because individuals change their behaviour in a specific moment, but also because of an ongoing ‘massive process of personal replacement’ in which older cohorts are succeeded by younger ones with different patterns of behaviour or attitudes (Ryder, 1965). Distinguishing between period effect and cohort effects is fundamental to understanding the sources of social change in both countries.

The general aim of this article is to determine the extent to which religious evolution in later cohorts varies relative to the kind of relationship that existed between the authoritarian state and the institutionalized religion in two contrasting European cases: Spain, which for a large part of the 20th century saw a close alliance between the Catholic Church and the Francoist political dictatorship, and Poland, where the Catholic Church was clearly in conflict with the communist regime. To be more specific, this article examines how the replacement of older generations by newer ones that were socialized during the decline of the authoritarian regimes or in a non-authoritarian political context has affected religiosity in both countries during the first 16 years after their political transitions. Our analytical strategy is twofold: first, we measure the differences in levels of religiosity among cohorts socialized during the authoritarian regimes and during the beginning of democratic periods; second, by using decomposition techniques and logistic regression modelling we analyse how religiosity has been affected by the change in generational composition in both societies over approximately one and a half decades after the end of the Franco regime in Spain and the fall of communism in Poland.

Background

Despite increasing criticism, the secularization theory still offers a useful explicative framework of the current processes of religious change (Bruce, 2011; Norris and Inglehart, 2004). This theoretical perspective claims that the diminishing significance of religion in social and individual life is a consequence of an inherent tension between modernity and religion (Müller, 2011; Pickel, 2011). In this sense, secularization is understood to be part of a broad societal process of modernization involving a series of interrelated phenomena, such as technological change, the development of the modern state, the spread of rational-scientific thought, the process of urbanization and the disappearance of small local communities (Bruce, 2002). Given a context of modernization, the secularization thesis asserts that more recent cohorts are less religious than earlier cohorts at the same age. In other words, the advances of modernization affect successive cohorts differently, leading younger generations to have lower levels of religiosity than their predecessors (Firebaugh and Harley, 1991; Voas, 2010). Under the circumstances of an ongoing modernization, the demographic process of cohort replacement itself fosters the secularization.

The intergenerational variation in the levels of religiosity of successive cohorts can be attributed to both structural changes and contextual transformations that occur in a given society (Voas and Doebler, 2011). First of all, the differences in the degree of secularization between generations are linked to the transformations in the composition of the successive cohorts brought about by increased levels of education, larger percentages of urban population or greater participation by women in the labour market (De Graaf and Te Grotenhuis, 2008). Second, along with changes in the composition of the population, religiosity is affected by societal and contextual factors that cause religion and its institutions to lose the capacity to influence how certain institutions function (the state, public administrations, educational system, mass media, etc.), as well as to control and determine the behaviour, beliefs and attitudes of individuals (Chaves, 1994; Dobbelaere, 2002). Although the social significance of the religious sphere has decreased for the entire population to a greater or lesser degree, the change is sharpest among the youngest generations. Individual’s experiences at ages of greatest cognitive development, when they are not yet fully mature, play a decisive role in their later ideas and behaviour (Hagenaars, 1990; Mannheim, 1952), particularly in their religious beliefs and practices. From this perspective, intergenerational metabolism itself is the decisive argument to explain this social change: it is not necessary for a radical change on the individual level to occur in order for societies to change, as younger generations start to take the place of the older ones.

Many studies carried out in Western European countries, such as the United Kingdom, the Netherlands and West Germany, indicate that the advance of secularization is increasingly evident as successive cohorts are socialized in contexts in which the link between religion and other important social spheres has been weakened (De Graaf and Te Grotenhuis, 2008; Kaufmann et al., 2012; Voas, 2009, 2010; Wolf, 2008). Nevertheless, the continuity of the secularization process associated with the succession of cohorts subject to societal processes of modernization is not so clear in societies which undergo profound and sudden political and social changes (Wohlrab-Sahr, 2011). The contextual factors of a given historical moment, such as coercive secularization or democratization of social life, could accelerate or delay the process of religious change associated with cohort replacement. For example, some recent studies indicate that one of the causes of the religious revival in Romania and the former GDR is caused by generational change as new cohorts socialized within a context of religious freedom replace those educated during the period of forced secularization (Lois, 2011; Voicu and Constantin, 2012). Therefore, countries that have transitioned from authoritarianism to democracy provide particularly interesting case studies in which to observe and analyse this interaction between cohort effects and period effects. Democratic transition after a prolonged authoritarian regime represents a decisive change in the context in which successive generations are socialized and can be considered a key factor in the acceleration or delay of secularization processes. This contextual change could be even more decisive if in the past there had been a close positive link or manifest antagonism between religion and politics.

During the second half of the 20th century Spain and Poland could be considered paradigmatic cases of two models of authoritarian systems in which institutional religion played radically opposing roles. In Spain, the close alliance between the Catholic Church and the Franco regime translated into a religious hegemony imposed on many aspects of social life in which Catholicism became a fundamental symbolic instrument to legitimize the dictatorship (Box, 2010). The Catholic Church acquired extraordinary influence over large swathes of the public sphere and private life, including a monopoly on teaching religious doctrine and on moral direction. Although in the 1960s and 1970s the ideological pressure had already started to diminish, the death of Franco in 1975 and the subsequent transformation of the political system in Spain substantially changed the relationship between the Catholic Church and the state. Although it maintained certain privileges over other religions (for example, funding for educational and pastoral activities), the political and social influence of the Catholic Church substantially diminished (Casanova, 1994; Linz, 1993).

Historically, in Poland the strong association between Catholicism and national identity was consolidated after the country was partitioned by the three neighbouring empires. In the late 1940s the imposition of a totalitarian communist regime monopolized by a single party led to a forced social, economic and cultural transformation in Poland (Varga, 1995). The fact that it was a foreign power imposing the totalitarian and atheist regime converted Catholicism into an ideology of national unity once again, with the Church as its main guardian. This situation converted the Church into a space in which civil society developed, further reinforced by its open and inclusive attitude towards all currents of political opposition, regardless of their ideologies (Mach, 2000). However, the start of the political transition in 1989 and 1990 created a new situation for the Polish Catholic Church in which it gradually lost its position as the repository of national identity.

The different positions of the Catholic Church within the political regimes in Poland and Spain have had very different implications for the later development of religious behaviour in those countries. For instance, the Spanish detached themselves from Catholicism as political coercion was relaxed. Period effect analysis shows that secularization in Spain was especially intense during the political transition in the second half of 1970s (Requena, 2005; Requena and Stanek, 2012). In contrast, in post-communist Poland the memory of the Catholic Church’s opposition to the regime helped maintain high levels of religious practice in the first years after the fall of communism (Requena and Stanek, 2012). Nevertheless, as the new system stabilizes, it is to be expected that the integrating function of Catholicism and its symbolic attraction will decrease and cease to slow the religious change associated with modernization, which manifests in a gradual decrease in levels of religiosity (Bruce, 1999; Martin, 2011).

Given these antecedents, this work analyses the religious change that occurred in Spain and Poland in the years following their democratic transitions. In addition to comparing the magnitude of the change in both countries, our specific objective is to figure out how much of it can be due to individual religious change (within cohorts) and how much can be attributed to generational replacement (across cohorts). To do this we studied the change in religious practice over 16 years, starting two years after the democratic transitions – 1980 in Spain, two years after the democratic Constitution of 1978 was approved; and 1992 in Poland, two years after the first completely democratic elections were held in 1990.¹ Our expectations, based on the thesis of the relevance of the interrelation between religion and politics for religious change, are: (1) that in equivalent periods of time, there has been less religious change in Poland than in Spain; (2) that after the democratic transitions generational change has been stronger than individual change in Spain, where the Catholic Church was strongly implicated with the authoritarian political regime and younger cohorts socialized during the decline of the Francoist dictatorship have distanced themselves from institutional religion since their youth; (3) that in Poland, where Catholicism was a crucial element in the opposition and resistance to the totalitarian political regime, generational change has been lower and possibly weaker than individual change because younger generations perceived a positive link between Catholicism and political freedom, did not feel the political urge to move away from religion and, accordingly, their contribution to societal religious change should have been less intense than in Spain; and (4) that the religious behaviour of cohorts socialized after the democratic transition is significantly different than that of cohorts socialized in authoritarian contexts.

Data and methods

The empirical information that we worked with is taken from a series of cross-sectional data based on national sample surveys in both countries. The data about Spain come from four studies carried out by the Centre for Sociological Research (CIS) covering 1980, 1985, 1992 and 1996. The data available for Poland cover 1992, 1997, 2003 and 2008 and were obtained from the Polish General Social Survey (Polski Generalny Sondaż Społeczny, PGSS), with the exception of the 2003 study which was carried out within the ISSP (International Social Survey Programme).² The selection of data sources was aimed at creating a comparative framework for both countries, so for each country four points were selected separated by approximately equal intervals of time (between four and six years). Although, obviously, it would be possible to analyse more recent data for Spain, the 16 years considered in both cases allow the intensity of religious change to be gauged over equivalent periods of exactly the same duration. The data used for Spain come from CIS Data Base studies 1237 (N = 2489), 1456 (N = 2485), 2001 (N = 2486), and 2230 (N = 2485). The data for Poland have been taken from PGSS 1992 (N = 1647), PGSS 1997 (N = 2401), ISSP 2003 (N = 1277) and PGSS 2008 (N = 1203).

For theoretical as well as methodological reasons, the present analysis is based on religious practice, understood as taking part in mass. On the one hand, religious practice is one of the most conspicuous aspects of any religion, and it can be interpreted as a ‘ritual for maintaining the institutional creed’ (Pérez-Agote et al., 1993). This makes it ideal for measuring change in church-oriented religiosity. On the other hand, using religious practice is also relevant from the methodological point of view because it is the most visible aspect of religiosity and, consequently, the easiest to research through surveys, producing the most reliable kind of information for researching change over time. This study identifies as practising Catholics those people who go to mass at least once a month, which is the criterion normally used in studies on religiosity (Müller, 2009). The indicator of religious practice is available in all the surveys mentioned above and the scales used to measure attendance to mass in Spain and Poland are similar, facilitating the homogenization process.

After presenting a summary of descriptive statistics which compares the magnitude and speed of religious change in both countries, two main analysis techniques are used in our study. First, using the cumulative data set for all the years examined, we carried out a decomposition based on a simple regression model to distinguish the two proximate sources of overall or aggregate change in religious practice in both countries: net change among individuals and population replacement (Firebaugh, 1997). Due to the categorical and binary nature of our dependent variable, we used logistic regression to estimate for each country a model of the odds of religious practice between practising Catholics (PC) and the rest of population (1 – PC) according to the following equation:

P C / (1 - P C) = \exp [b_{0} + b_{1} (Y E A R) + b_{2} (C O H O R T) + e] .

This regression model permits us to estimate intra-cohort and inter-cohort change throughout the periods considered. In the formula, b₀ is the model intercept, b₁ is the estimated slope for yearly change controlled for cohort, b₂ is the estimated slope of the difference between adjacent cohorts and e is the error term. The above equation takes the exponential form and estimates the odds of religious practice. It also can be expressed in logarithmic form in order to estimate the logit (the natural logarithm of the odds) that an individual will be a practising Catholic:

logit (P C / [1 - P C]) = L n (P C / [1 - P C]) = b_{0} + b_{1} (Y E A R) + b_{2} (C O H O R T) + e .

Firebaugh (1997) showed that overall change can be decomposed through the following two expressions: b₁ (YR_T – YR₁) for the estimated contribution of change within cohorts or individual change; and b₂ (C_T – C₁) for the estimated contribution of change across cohorts, where YR_T is the year of the final survey, YR₁ is the year of the first one, C_T is the average year of birth for the sample in the last survey and C₁ is the average year of birth for the sample in the first one. These two components can be added to calculate overall change. Since we used logistic regression, we decomposed the logit (that is, the odds ratio logarithm) of religious practice between two years (the overall change) into the two aforementioned components using the following equation:

Ln [P C_{T} / (1 - P C_{T}) / P C_{1} / (1 - P C_{1})] = b_{1} (Y R_{T} - Y R_{1}) + b_{2} (C_{T} - C_{1}) .

The target of this first step is mainly exploratory: knowing with some precision how much of the change can be attributed to individual religious change and to cohort composition during the specified periods in both countries and, to the extent possible, disregarding age effects. Unfortunately, it appears that there is no fool proof statistical formula for distinguishing age, period and cohort effects due to the well-known identification problem that plagues this kind of analysis (Glenn, 2005).⁴ Therefore it was necessary to employ informal means of inspecting data and assessing those effects. For reasons explained below, we think that the data from Spain and Poland legitimately allow us to ignore age-specific effects in the explanation of aggregate change of religiosity in both countries.

Second, in order to separate the effect of intergenerational change from other structural factors, a logistic regression model for each country was estimated, which, in addition to the year of survey (YR) and cohort (C),⁵ included a series of control variables representing sex (S), education (ED), employment status (ES) and the size of the municipality (SM) in which the people surveyed resided. The model can be expressed as follows:

[P C / (1 - P C)] = \exp [b_{0} + b_{1} (C) + b_{2} (Y R) + b_{3} (S) + b_{4} (E D) + b_{5} (E S) + b_{6} (S M)] .

The secularization thesis predicts that urbanization and a general increase in the level of education are crucial factors in the diminishing social significance of religion (Wallis and Bruce, 1998). Indeed, the improvement in the general level of education and the processes of urbanization in Spain and Poland throughout the second half of the 20th century could be considered major determinants of religious change in both countries. In addition, there is extensive empirical evidence that sex continues to determine the differences in levels of religious participation in modern societies (Tilley, 2003). The aim of this second exercise is simply to analyse cohort and period effects when the structural factors that advance the process of secularization are held constant.

Results

In the 16 years after the democratic transitions in Spain (1980–1996) and Poland (1992–2008) there was a significant drop in levels of religious practice. However, the decrease was much more pronounced in Spain where the number of practising Catholics dropped 14% while in Poland it only dropped 7% – meaning that the odds ratios of being a practising Catholic in Spain were 0.56 between the end and start of the period, compared to 0.71 in Poland. In both countries the change in attending mass over the time periods analysed, that is, the difference between this indicator at the start and end of the periods, is statistically significant (Z test, p < .000 in both cases). But given that the level of religiosity at the beginning of these periods was lower in Spain than in Poland, not only was the absolute loss of religiosity greater in Spain but also the relative loss (−28.7%) compared to Poland (−9.5%). In other words, during these 16 years practising Catholics in Spain decreased at an annual rate of −2.1%, while in Poland the decrease was only −0.6%.

Table 1 shows the basic data related to change in levels of religious practice in both countries. Because the focus of this work is on cohort replacement, the data are organized to compare the total change with the change in each of the cohorts included in the Spanish and Polish samples. Simply observing these percentages suggests that religious change was not only greater in Spain than Poland, but also that the impact of generational change was stronger in the former: that is, the average change among cohorts was greater.

Table 1.

Within cohort, cross-cohort and total change in percentage of practising Catholics in Spain (1980–1996) and Poland (1992–2008).

Cohorts	Spain				Poland				Changes
	1980		1996		1992		2008		Spain	Poland
	%	N	%	N	%	N	%	N	1980–1996	1992–2008
Before 1901	45.7%	35
1901–1910	71.8%	177	33.3%	15					−38.4%
1911–1920	66.9%	299	67.2%	116	78.7%	127	100.0%	2	0.4%	21.3%
1921–1930	61.2%	405	62.0%	287	73.1%	201	76.5%	81	0.8%	3.4%
1931–1940	56.4%	408	53.1%	356	72.6%	307	80.3%	132	–3.3%	7.7%
1941–1950	40.5%	489	41.2%	337	76.0%	242	71.2%	184	0.8%	−4.8%
1951–1960	26.9%	472	24.5%	413	75.7%	313	65.6%	273	–2.5%	−10.2%
1961–1970	27.9%	129	16.2%	505	72.9%	258	69.9%	196	–11.7%	−3.0%
1971–1980			20.0%	431	74.6%	118	59.1%	193		−15.5%
1981–1990							60.4%	212
All cohorts	49.0%	2414	34.8%	2029	74.5%	1566	67.5%	1273	–14.1%***	−7.0%***
Average within cohort change (weighted by size)									–4.1%	−2.7%
Estimated change in religious practice within and across cohorts: logistic regression
Overall change: ln [(P_T/(1 – P_T)/(P₁/(1 – P₁)] = b₁ (YR_T – YR₁) + b₂ (C_T – C₁) =									−0.566	−0.362
Within-cohort: b₁ (YR_T – YR₁) =									−0.155*	−0.231*
Cross-cohort: b₂ (C_T – C₁) =									−0.411***	−0.131***

***

p < .000; ^** p < .01.

Note: Practising Catholics are those who attend mass at least once a month.

Source: CIS Data Base (studies 1237 and 2230) for Spain and Polish General Social Survey 1992 and ISSP 2008 for Poland.

Table 1 also includes the decomposition using logistic regression of the logit of religious practice into its two components: individual or intra-cohort change and intergenerational change. This decomposition reveals that the Spanish have experienced much more inter-cohort change (–0.411) than individual change (–0.155). For some reason, the religious distance among adjacent cohorts was large in Spain, while individual religious switching was not very important in this period. Interestingly, things were quite different in Poland, where change within cohorts (–0.231) contributed more to the total change than across cohort change (–0.131). In Poland, the level of church attendance of adjacent generations was not as different as in Spain, and there was much more individual religious change than generational renewal.

It is worth mentioning that in both countries the slopes for cross-cohort and individual components of change are both negative (Spain: b₁ = −0.010 and b₂ = −0.033; Poland, b₁ = −0.014 and b₂ = −0.009). In other words, during the post-transitional periods in Spain and Poland population replacement reinforced individual change, and vice versa. On the one hand, according to the same-sign slope rule (Firebaugh, 1997), it means that overall change outpaced individual change, something which usually produces a pervasive sensation of very fast religious change. On the other hand, the fact that both coefficients are negative practically allows significant age effects in the data for these periods to be discarded or at least to conclude that age effects are smaller than cohort effects. Under the same-signs rule, when both slopes are negative age effects should be smaller than cohort effects whenever aggregate change outpaces individual change.

Moreover, the fact that the aggregate level of religious practice has significantly dwindled in both countries, despite the prima facie positive relationship between age and religiosity, strongly suggests powerful cohort effects, at least in Spain. Careful intra-cohort data inspection (Figure 1) will be helpful here. In Spain, the lack of change in religious practice within the several cohorts points to the fact that the fall in religious practice has resulted mostly from cohort succession. The pattern is much more complex in the Polish case where we already know that intra-cohort change outpaces inter-cohort change. In any case, it should be noted that only Polish cohorts born in 1921–1930 and in 1931–1940 have experienced increased church attendance during this 16-year period, whereas none of the remaining cohorts analysed in this study (that is, Poles born between 1941 and 1980) has increased religious practice. This pattern can be indicative of period effects in the Polish case.

Figure 1.

Percentage of practising Catholics in Spain (1980–1996) and Poland (1992–2008) by birth cohort.

Table 2 shows two logistic regression models which attempt to expand previous analyses of the intra-cohort and inter-cohort change in Spain and Poland, adjusting for factors which may have affected religious behaviour such as sex, educational level, size of municipality and employment situation.⁶ In other words, we estimate how church attendance probabilities vary in relation to consecutive cohorts and periods while keeping other variables related to structural and individual characteristics constant. As expected, log-odd coefficients estimated for Spain show a steady trend of decline in religiosity across cohorts. Each Spanish generation born between 1921 and 1960 is considerably less religious than the previous one. This observation confirms the results of the bivariate analysis already presented in Figure 1 and of the decomposition analysis. On the other hand, our model shows no significant differences between the two youngest Spanish cohorts, socialized during the political transition and immediately afterwards (1961–1970 and 1971–1980), even though both are still considerably less religious than those born in the 1950s and earlier. The coefficients estimated for survey years show a relatively small additive period effect once the rest of factors are controlled. Specifically, between 1980 and 1992 the overall level of religious practice remained steady and differences between each survey year are not statistically significant. In terms of period effects, the probability of church attendance only decreased, but not very much, in the final years (1992–1996).

Table 2.

Log-odds from logistic regression on religious practice in Spain and Poland (dependent variable: attending mass at least once a month).

	Spain		Poland
Cohort (ref. Spain: 1911–1920; ref. Poland: 1921–1930)
1921–1930	−0.33***	(0.101)
1931–1940	−0.48***	(0.101)	0.14	(0.129)
1941–1950	−1.14***	(0.104)	−0.06	(0.127)
1951–1960	−1.83***	(0.109)	−0.14	(0.127)
1961–1970	−2.15***	(0.117)	−0.11	(0.133)
1971–1980	−2.07***	(0.144)	−0.25*	(0.136)
1981–1990			−0.56***	(0.161)
Survey year Spain (ref. 1980)
1985	0.12	(0.068)
1992	−0.11	(0.071)
1996	−0.22**	(0.071)
Survey year Poland (ref. 1992)
1997			−0.07	(0.081)
2003			−0.13	(0.096)
2008			−0.24**	(0.093)
Women	1.05***	(0.055)	0.71***	(0.060)
Educational level (ref. no education)
Primary/first grade of secondary	−0.17	(0.095)	0.09	(0.204)
Second grade of secondary	−0.27***	(0.085)	0.05	(0.105)
Tertiary	−0.54***	(0.098)	−0.03	(0.087)
Employment situation (ref. non-active)
Employed	−0.30***	(0.059)	−0.13	(0.075)
Unemployed	−0.57***	(0.094)	−0.57***	(0.112)
Habitat (ref. under 10,000)
10,001–50,000	−0.50***	(0.07)	−0.62***	(0.103)
50,001–100,000	−0.74***	(0.102)	−0.85***	(0.074)
over 100,000	−0.79***	(0.06)	−1.02***	(0.158)
N		9,005		6,022
Intercept	1.27***	(0.133)	1.36***	(0.198)
Log likelihood	–5150.555		–31378.036
Pseudo R²	0.159		0.064

Standard errors in parentheses.

***

p < .000; **p < .01; *p < .05.

Source: CIS Data Base for Spain and Polish General Social Survey and ISSP 2003 for Poland (see text for specific studies).

Consistent with the previous exercise’s results, the logistic regression model for Poland displays a much more irregular pattern of religious practice when compared with Spain (cf. with Figure 1). First, there are no considerable differences in religious attendance among cohorts born before 1971. Second, differences between coefficients indicate that Poles born after 1970 are much less likely than older cohorts to attend mass on a regular basis. Apart from this generational gap, our results also show that Poles born in the 1980s are relatively less religious in comparison to the generation born in the previous decade. This may suggest that among Poles who have grown up in democracy each new generation tends to be less involved in religion than the previous one. This trend can be important to the extent that it may represent a potential secularizing force which could cause a general religious decline in Polish society over the coming years. Finally, our analysis also confirms that much of the change in religious practice is due to period effects (intra-cohort change). The evolution of the log-odds of church attendance over time reflects a successively decreasing trend. In addition, estimates show that period-related decline was especially intense between 2003 and 2008. To sum up, in the new political context of post-transitional Poland two overlapping processes are taking place: a negative period effect on religious participation, combined with a negative cohort effect, which exclusively affects the religious involvement of post-communist cohorts.

Discussion

The objective of this article is to analyse the dynamics of religious change after the recent democratic transitions in Spain and Poland under the following hypotheses: (1) that the intense interaction between religion and politics has dominated religious change in both countries and (2) that the democratization of their political systems has had different influences on the changing levels of religiosity in both societies. The comparison between Spain and Poland is particularly interesting due to the fact that both countries have a long and deeply rooted Catholic tradition and both have had authoritarian systems imposed upon them during long periods of the second half of the 20th century (Casanova, 1994). However, the institutionalized religion has played very different roles in the recent histories of both countries. It is to be expected that these differing roles would have produced equally differing processes of religious change that are driven by very different generational dynamics in each case.

In order to empirically support our interpretation of religious change in these countries, we have analysed cohorts and attempted to specifically determine how much of the religious transformation can be explained by change that simultaneously affects the entire population (period effect) and how much depends on the succession of cohorts socialized in authoritarian and post-authoritarian contexts (cohort effect). If the interpretation of religious change that we present in this work is correct, in the years following the transition to democracy religious change should be more intense in Spain than in Poland; the generational effects should have been greater than the period effects in Spain; and the impact of population replacement should have been smaller than individual change in Poland.

As we have been able to prove, since the dictatorship ended in Spain and the country transitioned to a new political regime, the generational effect on religious change has grown over time: the negative effects on levels of religious practice have been greater for each successive generation. Our interpretation of the powerful generational dynamics of religious change in this country is primarily based on the Catholic Church’s commitment to the symbolic construction and political legitimization of the Franco dictatorship, as well as on the negative effects that commitment has had on the religiosity of generations that grew up during the final stages of the dictatorship and that opposed it and drove the democratic transition. Note that the intensity of the distancing from the Catholic Church is already quite noticeable in Spanish cohorts born between 1940 and 1950, that is, those who were adolescents in the 1960s and who became adults in the 1970s. Not only were these cohorts witnesses to but also protagonists of the intense social and economic modernization that in those years led to the final political decomposition of the Franco dictatorship. Three characteristics distinguish these cohorts: in their early youth they acquired a very clear understanding of the Catholic Church’s support for the Franco regime, they matured in a social and cultural context that was highly critical of the Church’s commitment to the dictatorship, and they saw the arrival of democracy at the start of their adult life. This is why they noticeably distanced themselves from the institutional religion and have very consistently maintained that distance throughout their adult lives. Without exception, cohorts born in Spain after 1950 have followed the same pattern towards the Catholic Church that had begun with the cohort born in the 1940s: maintaining the same distance, lack of interest and a critical attitude to the Catholic Church, leading to decreasing aggregate levels of religious practice.

The results of our analysis show that in the Spanish case the generational dynamics of religious change clearly and by a wide margin surpass individual dynamics during the post-transition to democracy. Although period effects could have been important right around the time of political transition (Requena and Stanek, 2012), once the democratic regime was implemented and consolidated, on an individual level the Spanish have barely changed the level of religious commitment they acquired before adulthood. This is perfectly in line with the basic premise of the cohort analysis (Glenn, 2005; Mannheim, 1952; Ryder, 1965) which sees that the dissimilar socialization of cohorts born in different historical moments is decisive in creating the attitudes and behaviours that they will follow in their adult lives. In this regard, it must be emphasized that all cohorts born after 1950 have been socialized and grew up in a political context marked by a growing distance from institutional religion and in a sociocultural environment increasingly favourable to secularization, thanks to the advance of socioeconomic modernization and the well-known opposition to the Catholic Church involved in some of these contemporary processes of social change (Requena, 2008).

It is not surprising, therefore, that during the period analysed in this work the generational dynamics of intense religious change is more evident in the youngest cohorts – that is, among those who were born between 1961 and 1980. The world view of these generations who grew up and matured in democracy is largely incompatible with the doctrine of the Catholic Church, which has ultimately been unable to free itself from the legacy of its political commitment to the Franco dictatorship. In addition, these powerful dynamics of religious change between generations have allowed Spain to quickly join the general secularization trend characteristic of other European countries such as France, the United Kingdom or the Netherlands and diminished the practice of Catholicism in the country to unprecedented levels.

In accordance with our premises, the evolution of the religious phenomenon in Poland after the democratic transition follows a very different pattern than in Spain. As mentioned earlier, the establishment of the communist system in Poland converted the Catholic Church into the institution most capable of integrating and mobilizing large segments of society to oppose (passively and actively) the autocratic regime. Moreover, in the increasingly confrontational climate that existed before the Polish democratic transition, the communist regime saw the Church as the institution most capable of stabilizing the explosive political and social situation in the country and, therefore, it become the main political interlocutor and de facto representative of society when dealing with the authorities (Eberts, 1998). In terms of our interpretation of the political and religious change, the unifying role of the institutional religion in opposition to the communist regime created a specific context of socialization that profoundly influenced broad sectors of the population. It is precisely this specific socialization context, in which the Catholic Church embodied and concentrated the political resistance to communism, which explains the absence of significant cohort effects on religious change between generations who reached adulthood during the communist period. In contrast to what occurred in Spain, the Polish generations socialized under communism did not develop a political-ideological hostility towards the institutional religion, but rather tied their hopes and aspirations for political liberation and change to Catholicism. In addition, the relative homogeneity of religious behaviour and relatively high levels of religious practice in the cohorts socialized during communism allow other characteristics of religious change in Poland to be compared: the close tie between the Catholic Church and national identity and resistance to communism contributed to counteracting the negative effects of secularization on religious practice derived from the religious policy of the communist regime and the structural changes associated with intense industrialization and urbanization (Bruce, 2002; Varga, 1995).

However, the uniform religious behaviour of generations who shared the experience of growing up and maturing in a communist context has been broken by those born in the 1970s and 1980s, that is, the cohorts socialized during and after the collapse of communism in Poland. Our analysis has revealed important differences between the communist and post-communist generations, with the former being significantly more religious. Furthermore, this generation gap tends to become sharper with the succession of cohorts: Poles born in the 1970s have significantly higher levels of religious practice than those born in the following decade. Along these same lines, generational religious change has been reinforced by a decrease in religious practice on the individual level after 2003. The recent confluence of the cohort and period effects gives an idea of the extent of the change in religious values, attitudes and behaviour that Polish society has experienced since the fall of the communist regime.

More specifically, the social modernization and pluralism that have arisen after the fall of communism in Poland have created a process of social differentiation in which the Church has lost its social position as the main representative of society before political power (Mariański, 2003). At the same time, in the post-communist period liberal and secularizing currents have arisen or been revived that demand greater limits on the real and symbolic influence of the Catholic hierarchy. These currents are in conflict with the Church’s attempts to maintain its political influence, an influence that causes discontent in many sectors of public opinion (Burdziej, 2005; Mach, 2000). In addition, the involvement of the Church in public debate generates a new perception of this institution among some sectors of Polish society. Church is increasingly being perceived as just one of the many actors participating in the development of a vibrant new civil society in Poland.

Therefore, for younger generations of Poles the Catholic Church and its doctrinal moral message have become just one of many lifestyle options to choose from (Szlendak, 2004). In summary, the gradual transformation of religious patterns due to cohort replacement and individual change reveals that the fall of the communism regime has also entailed, perhaps ironically, the intensification of secularization in Poland.

Our findings confirm Martin’s secularization theory (1978), which claims that variation in patterns of religious change not only results from the divergent pace of urbanization, industrialization and expansion of scientific world view across Europe but also is related to the role of the institutional religion in political conflict. In addition, the comparison between Spain and Poland confirmed that the process of secularization in traditionally Catholic countries presents uneven patterns depending on whether the society suffers internal political conflict or faces an external threat. To conclude, our study showed that the secularization process is a non-linear phenomenon that is often interrupted by parallel processes and events. Therefore, any analysis of historical changes in religiousness should take into consideration the historical, political and sociocultural particularities of each country.

Footnotes

Acknowledgements

The authors would like to thank the anonymous reviewers for their helping comments and suggestions on the original manuscript.

Funding

Mikolaj Stanek would like acknowledge the support of research grant (25.102.462-M.481) awarded by the Spanish Centre for Sociological Research in 2011.

Notes

Author biographies

Miguel Requena is Full Professor of Sociology in the Department of Social Structure (Universidad Nacional de Educación a Distancia, Madrid, Spain) and member of the ‘Population and Society Research Group’ (GEPS). He has been Senior Associate Member at St Antony’s College (Oxford University) and Visiting Professor at Princeton University. His research interests are demography and sociology of family, social stratification, religion and social change.

Mikolaj Stanek is a postdoctoral researcher at the Institute of Economy, Geography and Demography of the Spanish Council for Scientific Research (Madrid, Spain) and a member of the ‘Population and Society Research Group’ (GEPS). His current research interests include the study of religious changes in post-authoritarian countries, migrants’ labour market integration and international mobility patterns of post-communist societies.

References

Berger

(1999) The desecularization of the world: A global overview. In: Berger

(ed.) The Desecularization of the World: Resurgent Religion and World Politics. Washington, DC: Washington Ethics and Public Policy Center.

Berger

Davie

Fokas

(2008) Religious America, Secular Europe? A Theme and Variation. Farnham: Asghate.

Box

(2010) España, año cero. La construcción simbólica del franquismo. Madrid: Alianza Editorial.

Bruce

(1999) Modernisation, religious diversity and rational choice in Eastern Europe. Religion, State and Society 27(3/4): 265–275.

Bruce

(2002) God is Dead: Secularisation in the West. Oxford: Blackwell.

Bruce

(2011) Secularization: In Defence of an Unfashionable Theory. Oxford: Oxford University Press.

Burdziej

(2005) Religion and politics: Religious values in the Polish public square since 1989. Religion, State and Society 33(2): 165–174.

Caplow

(1985) Contrasting trends in European and American religion. Sociological Analysis 46(2): 101–108.

Casanova

(1994) Public Religions in the Modern World. Chicago: University of Chicago Press.

10.

Chaves

(1994) Secularization as declining religious authority. Social Forces 72(3): 749–774.

11.

Davie

(2000) Religion in Modern Europe: A Memory Mutates. New York: Oxford University Press.

12.

Davie

(2007) Sociology of Religion. London: Sage.

13.

De Graaf

Te Grotenhuis

(2008) Traditional Christian belief and belief in the supernatural: Diverging trends in the Netherlands between 1979 and 2005? Journal for the Scientific Study of Religion 47(4): 585–598.

14.

Dobbelaere

(2002) Secularization: An Analysis at Three Levels. Brussels: P.I.E. – Peter Lang.

15.

Eberts

(1998) The Roman Catholic Church and democracy in Poland. Europe-Asia Studies 50(5): 817–842.

16.

Filali-Ansary

(1996) Islam and liberal democracy: The challenge of the secularization. Journal of Democracy 24(4): 76–80.

17.

Firebaugh

(1997) Analysing Repeated Surveys. Thousand Oaks, CA: Sage.

18.

Firebaugh

Harley

(1991) Trends in U.S. Church attendance: Secularization and revival, or merely lifecycle effects? Journal for the Scientific Study of Religion 30(4): 487–500.

19.

Glenn

(2005) Cohort Analysis. Thousand Oaks, CA: Sage.

20.

Hagenaars

(1990) Categorical Longitudinal Data: Log-Linear Panel, Trend, and Cohort Data. Newbury Park, CA: Sage.

21.

Halman

Draulans

(2006) How secular is Europe? The British Journal of Sociology 57(2): 263–288.

22.

Kaufmann

Goujon

Skirbekk

(2012) The end of secularization in Europe?: A socio-demographic perspective. Sociology of Religion 73(1): 69–91.

23.

Linz

(1993) Religión y política en España. In: Díaz Salazar

Giner

(eds) Religión y sociedad en España. Madrid: CIS, pp. 1–50.

24.

Lois

(2011) Church membership and church attendance across time: A trend analysis considering differences between East and West Germany. Comparative Population Studies – Zeitschrift für Bevölkerungswissenschaft 36(1): 161–192.

25.

Mach

(2000) The Roman Catholic Church in Poland and the dynamics of social identity in Polish society. In: Inglis

Mach

Mazanek

(eds) Religion and Politics: East–West Contrasts from Contemporary Europe. Dublin: University College Dublin Press, pp. 113–128.

26.

Mannheim

(1952) The problem of generations. In: Kecskemeti

(ed.) Essays on the Sociology of Knowledge. London: Routledge and Kegan Paul, pp. 276–322.

27.

Mariański

(2003) Kościół katolicki a rozwój społeczełstwa obywatelskiego. Rocznik Lubuski XXIX(I): 113–134.

28.

Mariański

(2011) Katolicyzm Polski – ciągłośc i zmiana. Studium socjologiczne. Kraków: Wydawnictwo WAM.

29.

Martin

(1978) A General Theory of Secularisation. Oxford: Basil Blackwell.

30.

Martin

(2011) The Future of Christianity: Reflections on Violence and Democracy, Religion and Secularization. Farnham: Ashgate.

31.

Müller

(2009) Religiosity in Central and Eastern Europe: Results from the PCE 2000 Survey in comparison. In: Pickel

Müller

(eds) Church and Religion in Contemporary Europe: Results from Empirical and Comparative Research. Wiesbaden: VS Verlag fur Sozialwissenschaften, pp. 65–87.

32.

Müller

(2011) Secularization, individualization, or (re)vitalization? The state and development of churchliness and religiosity in post-communist Central and Eastern Europe. Religion and Society in Central and Eastern Europe 4(1): 21–37.

33.

Norris

Inglehart

(2004) Sacred and Secular: Religion and Politics Worldwide. Cambridge: Cambridge University Press.

34.

Pérez-Agote

(2010) Religious change in Spain. Social Compass 57(2): 224–234.

35.

Pérez-Agote

Calvo

Fernández

. (1993) Religión, política y sociedad en el País Vasco. In: Díaz Salazar

Giner

(eds) Religión y sociedad en España. Madrid: Centro de Investigaciones Sociológicas, pp. 243–282.

36.

Pickel

(2009) Secularization as a European fate? Results from the church and religion in an enlarged Europe project 2006. In: Pickel

Müller

(eds) Church and Religion in Contemporary Europe: Results from Empirical and Comparative Research. Wiesbaden: VS Verlag fur Sozialwissenschaften, pp. 89–122.

37.

Pickel

(2011) Contextual secularization: Theoretical thoughts and empirical implications. Religion and Society in Central and Eastern Europe 4(1): 3–20.

38.

Requena

(2005) The secularization of Spanish society: Change in religious practice. South European Society and Politics 10(3): 369–390.

39.

Requena

(2008) Religión y sociedad: la secularización de la sociedad española. In: González

Requena

(eds) Tres décadas de cambio social en España. Madrid: Alianza Editorial, pp. 317–343.

40.

Requena

Stanek

(2012) Impact of democratic transitions on religiosity in Spain and Poland: A period effect analysis. Documentos de Trabajo del Grupo de Estudios Población y Sociedad IV(1), Madrid. Available at: www.geps.es.

41.

Ryder

(1965) The cohort as a concept in the study of social change. American Sociological Review 30(6): 843–861.

42.

Sasaki

Suzuki

(1987) Changes in religious commitment in the United States, Holland, and Japan. American Journal of Sociology 92(5): 1055–1076.

43.

Szlendak

(2004) Supermarketyzacja. Religia i obyczaje seksualne młodzieży w kulturze konsumpcyjnej. Wrocław: Wydawnictwo Uniwersytetu Wrocławskiego.

44.

Tilley

(2003) Secularization and aging in Britain: Does family formation cause greater religiosity? Journal for the Scientific Study of Religion 42(2): 269–278.

45.

Tomka

(2011) Expanding Religion: Religious Revival in Post-Communist Central and Eastern Europe. Berlin and New York: De Gruyter.

46.

Varga

(1995) Modernity or pseudo-modernity? Secularization or pseudo-secularization? Reflections on East-Central Europe. In: Roberts

(ed.) Religion and the Transformation of Capitalism: Comparative Approaches. London: Routledge, pp. 231–247.

47.

Voas

(2009) The rise and fall of fuzzy fidelity in Europe. European Sociological Review 25(2): 155–168.

48.

Voas

(2010) Explaining change over time in religious involvement. In: Collins-Mayo

Dandelion

(eds) Religion and Youth. Farnham: Ashgate, pp. 25–32.

49.

Voas

Doebler

(2011) Secularization in Europe. Religious change between and within Birth cohorts. Religion and Society in Central and Eastern Europe 4(1): 39–62.

50.

Voicu

Constantin

(2012) Religious revival in Romania: Between cohort replacement and contextual changes. In: Pickel

Sammet

(eds) Transformations of Religiosity: Religion and Religiosity in Eastern Europe 1989–2010. Wiesbaden: Springer, pp. 101–119.

51.

Wallis

Bruce

(1998) Secularization: The orthodox model. In: Bruce

(ed.) Religion and Modernization: Sociologists and Historians Debate the Secularization Thesis. Oxford: Oxford University Press, pp. 8–30.

52.

Wohlrab-Sahr

(2011) ‘Forced’ secularity? On the appropriation of repressive secularization. Religion and Society in Central and Eastern Europe 4(1): 63–77.

53.

Wolf

(2008) How secularized is Germany? Cohort and comparative perspectives. Social Compass 55(2): 111–126.