Studying fertility in Israel with structural equation modeling

This post by Evgenia Bystrov highlights the controversy around the ultimate causes of high fertility in Israel: religiosity or nationalism.

There is a long-running debate on the drivers of fertility among Jews in Israel. Two papers in particular stand out: Friedlander and Feldmann’s (1993) and Anson and Meir’s (1996). Surprisingly, the papers reach a different conclusion regarding the determinants of high fertility among Jews in Israel, although both of them used the same data sources, namely, 1984 national elections and 1983 census. Religiosity was measured as an aggregate of votes for religious parties in selected urban units for both studies, while the latter study added nationalism as a separate factor. Fertility was calculated from census data from birth numbers in the analyzed geographical units. To introduce a new dimension to the debate, current article goes beyond aggregate data by employing individual level data from the 2009-2010 political surveys of the Guttman Center of The Israel Democracy Institute. Current study investigates the individual level mechanisms behind childbearing behaviour. It does so by developing measures of religiosity and nationalism as latent constructs, and employing structural equation modeling (SEM).

Most social phenomena that are of interest for demographers have a time dimension. Reconstructing such phenomena from self-reported measures in a cross-sectional survey is not a simple task. A question arises, why would anyone analyze fertility from political survey data? To answer that, political surveys have an advantage of measuring social values and attitudes in addition to the standard socio-demographic variables. This is crucial for testing the theories at hand, since both religiosity and nationalism can be regarded as socio-cultural subjective measures. Population data are often lacking in subjective measures for individuals, although they follow accurately demographic behaviour. At the same time, political data are frequently missing an accurate description of respondents’ fertility. This is why some seemingly straightforward questions, as whether religiosity or nationalism cause high fertility, still puzzle demographers.

Cross-sectional research design is criticized widely for its many pitfalls: the lack of temporal perspective (where retrospective information is not provided), and therefore, inability to fully account for age-period-cohort effects, the issue of spurious relations between research variables, the problem of identifying causality, and selection bias. From cross-sectional data analysis, it is often difficult to address mechanisms that underlie social structures without falling into one of these traps. Indeed, authors are requested to show that the most important and highly valued goals in contemporary population research are met. They need to justify causal inferences based on statistical associations, exclude any possible selection bias, and convince the reader that the explanations are not confounded.

Since this research uses cross-sectional data, the choice of method for data analysis is a main methodological concern. Structural equation modeling used in current study captures the complexity of interconnections between the social, cultural, political and economic factors that affect fertility. In doing so, structural equation modeling combines two advantages: firstly, it reduces data by constructing latent variables from observed measurements. Secondly, it is able to depict recursive relations between research variables, similarly to simultaneous equations. In this way, endogeneity is tackled; it is modeled, instead of being regarded as a problem.

The competing theoretical explanations are shown in Figure 1. The restricted model follows Friedlander and Feldmann’s (1993) argument that religiosity explains fertility. The full mediation model represents Anson and Meir’s (1996) argument that spurious relations exist between religiosity and fertility, while nationalism is a common explanation for both of them. According to this argument, it is not religiosity but nationalism that affects fertility, and once nationalism is introduced in the equation, religiosity loses its explanatory power. In current empirical analysis, both factors are introduced and tested. Religiosity is constructed from two survey items and nationalism from four items. Fertility is measured by the respondent’s number of births. Results show that religiosity is the main determinant of fertility, while spurious relations exist between nationalism and fertility at the individual level.

Figure 1. Restricted model and full mediation model of fertility

Figure 1. Restricted model and full mediation model of fertility

The study also demonstrates that there are additional factors that affect both nationalism and religiosity (importantly, nationalism and religiosity are highly correlated). For instance, traditional attitudes towards gender roles are positively correlated with both nationalism and religiosity. Higher educational levels and socio-economic status are negatively associated with nationalism and religiosity. Being a migrant from the former USSR is positively correlated with nationalism but negatively with religiosity. Interestingly, immigrants from the former USSR are more nationalistically oriented, secular, highly educated but of a lower socio-economic status than the rest of Jewish Israelis.

The study concludes that socio-cultural factors – first and foremost religiosity – are the strongest predictors of individual level fertility among the Jewish population in contemporary Israel. This is in contrast to what happens in Western countries, where socio-economic factors, women’s education and labor market attainment explain fertility differentials. The high fertility of Israeli Jews is exceptional for a post-industrial society. The heterogeneity of fertility among population groups, which vary by the level of religiosity, explains these high fertility rates. Therefore, high fertility is likely to persist as long as the existing social structure and traditional cultural preferences persist.

To summarize, despite the built-in difficulty to account for temporal dynamics from cross-sectional survey data, this study found evidence in support of the theoretical argument of Friedlander and Feldmann (1993). It is important to emphasize that while these authors found religiosity being the ultimate determinant of Jewish fertility in Israel at the aggregate level, the current study analyzed individual level determinants of incomplete fertility among women aged 25-55. Therefore, this study affirms that the explanation of religiosity is valid at another level of analysis – the individual level. Furthermore, testing the structure of determinants of fertility at both levels, and further elucidating the mechanisms and complexity of this phenomenon, are promising directions for demographic research in the future. Without a doubt, introducing values, attitudes and worldviews as potential determinants of demographic behaviour is essential for a better understanding of societies.

Evgenia Bystrov holds a PhD in sociology (2013) from the University of Bremen and Jacobs University Bremen. This blog post is based on her article The case of religiosity, nationalism and fertility among Jews in Israel revisited. This article is a concluding part of her doctoral dissertation, entitled: ‘The Second Demographic Transition Theory in Practice: The Case of Israel’. Any feedback on this research spotlight is welcome via email


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: