ESRA 2017 Programme

Tuesday 18th July      Wednesday 19th July      Thursday 20th July      Friday 21th July     

     ESRA Conference App

Tuesday 18th July, 16:00 - 17:30 Room: F2 105

Analyses of social change with cross-sectional and longitudinal data 1

Chair Ms Malgorzata Mikucka (Université catholique de Louvain, Belgium )
Coordinator 1Mr Francesco Sarracino (Statec, Luxembourg)
Coordinator 2Ms Tatiana Karabchuk (United Arab Emirates University)

Session Details

The availability of repeated cross-sectional surveys and of panel data allows analyzing social change over time. This type of analyses became popular after the recent studies on the relationship between economic growth and the trends of subjective well-being. Since then, this approach has been applied in various domains. Currently, researchers are increasingly interested in combining longitudinal and cross-sectional approaches to study social change. However, this field of research is still in its infancy and consequences of various methodological choices are still not well understood.

This session invites papers discussing the conceptual and methodological problems of analyzing social change over time with data such as macro-level time series, cross-sectional, and longitudinal surveys. In particular we welcome substantive research which investigates social change over time, presents novel methodological approaches, as well as postulates “good practices” in analyzing such data.

The topics include, but are not restricted to:

1. Research which investigates short- and long-term trends over time, as well as discusses methods of estimating trends and their consequences;

2. Analyses of relationships between changes occurring in various domains of social life, performed both within time-series and comparative frameworks;

3. Papers that distinguish between the effects of cross-sectional differences and the effects of overtime changes of the same factors;

4. Studies analyzing social change with comparative panel data.

Paper Details

1. Life-Long Occupational Careers: A Dynamic Study Based on Data from the Polish Panel Survey, POLPAN 1988-2013
Professor Kazimierz Slomczynski (The Ohio State University, Polish Academy of Sciences and CONSIRT)
Dr Irina Tomescu-Dubrow (Polish Academy of Sciences and CONSIRT)
Dr Zbigniew Sawinski (Polish Academy of Sciences)

Original models of occupational careers using differential equations were based on two oversimplifying assumptions: (1) the value of education is constant, and (2) status does not decrease after reaching equilibrium (Sørensen1974, Tachibanaki 1979, Rosenfeld 1980). We test a model of career trajectories that overcomes these shortcomings. The model (Slomczynski, Krauze & Peradzyński 1986) is described by the equation: d(S – M)/dt = -α (S - M) + σ dM/dt where S denotes observed status (SES), M refers to the status imputed by its determinants, the parameter α is the rate of convergence of actual status S to an equilibrium enforced by M, and parameter σ is a correction factor, which increases or decreases the impact of changes in the recalibrated imputed status. Presence of σ dM/dt makes the model flexible, allowing for a decrease in status at the end of career. We test the model on the basis of data from the Polish Panel Survey, POLPAN 1988-2013, a survey conducted among a national sample representing Poland’s adult population (aged 21-65) in 1988, with face to face interviews administered every five-years. To ensure an adequate age balance, additional subsamples involving young cohorts have been supplemented since 1998. For analyzing these data we use CONVERTER, a special software that transforms records of jobs a particular person holds over the years into fixed time points of a full career. Results are in agreement with the econometric model of occupational careers proposed by Sicherman and Galor (1990), where highly educated individuals start their careers on a “higher step on the ladder” and face “longer ladders.”

2. Europe united or still apart? East-West Differences in Gender Role Attitudes Over Time
Mrs Jessica Walter (GESIS - Leibniz Institute for the Social Sciences)
Dr Malina Voicu (GESIS - Leibniz Institute for the Social Sciences)

In most countries gender role attitudes became more egalitarian in the last decades. Since the 1980s on support for the male bread-winner model decreased and more gender equal models of division of labor gained acceptance. Besides these general developments over time the question remains whether they lead to a convergence of gender role attitudes across Europe or whether cultural differences persist. Modernization and its associates, the increase in female labor force participation and changes in family structure, are assumed to be accompanied by more gender equality and egalitarian attitudes. However, path dependency theory assumes that cultural influences persist over time and previous research points out to the long lasting effects of factors such as religious traditions or family models on gender role attitudes. In Europe, the communist legacy acted as dividing factor between Central and Eastern Europe on one side and Western Europe on the other side. Due to the policies promoted by the communist regime, citizens of post-communist countries were highly supportive to female employment, but more reluctant with respect to gender equality within the family. In Western Europe, the general trend evolved more towards support for gender equality on both dimension simultaneously. Accordingly, East and West European countries offer a strong case for investigating how path dependency operates in case of gender attitudes. However, studies which compare East and West European countries over time are rare and not up to date. This paper focuses on overtime changes on gender role attitudes in Europe, investigating how support for gender equality evolved during the last decades under the impact of modernization, how persistent the effect of early socialization is and if the gap between post-communist countries and the rest of Europe reduced over time. The paper cumulates survey data coming from different large scale surveys (European Values Study, the World Values Study and the International Social Survey Programme), collected since 1990 on, and makes use of Cross-Classified Multilevel Models to answer to the research questions. Problems related to this procedure – such as different question formats and context differences – are discussed. First descriptive results confirm that gender role attitudes become more egalitarian over time. However, the differences between East and West still persist although the gap reduced especially since 2005.

3. The effect of nonresponse on long-term trends of self-rated health
Ms Oona Pentala (National Institute for Health and Welfare Finland)

Public health programmes have aimed at narrowing health inequalities in the population. Nevertheless, evidence on changes over longer time periods in health inequalities is lacking. Lahelma et al (2016, submitted) examined 36-year trends in educational differences in self-rated health among working aged Finnish women and men.

Data for the study were derived from the Finnish Adult Health Behaviour Survey (1979-2014) by the Finnish National Institute for Health and Welfare (THL). The data have been annually collected using mail surveys (n=2183 to 4089). The analyses included 25-64 years old respondents. Health status was measured by average or poorer self-rated health where 5-year moving averages and 95% confidence intervals was calculated for the age-adjusted prevalence of self-rated health. Socioeconomic position was measured by relative education, i.e. years of completed education were divided into quantiles.

Throughout the 36-year period the prevalence of average or poorer self-rated health improved somewhat, with women’s health being slightly better than that oHealth inequalities between educational groups remained large over the study period. Health in the lowest educational group was clearly worse than that in the two higher groups. f men. Health inequalities among women and men were broadly similar and narrowed somewhat from the 1980s to 1990s. Since the late 1990s, health inequalities have remained almost unchanged, and continue to be large and consistent in the mid-2010s.

However, these results may be biased since the effect of nonresponse was not accounted for. The response rate have been clearly declining and starting from the 1990s it has been less than 70% in nearly every age group ever more focusing on low-educated men and non-married persons (Tolonen et al, 2006). Unfortunately it is not possible to link this survey data with other registers yet but adjustments can be done based on population and respondent distributions.

Preliminary nonresponse adjusted results show that the health inequalities between high and low educated groups are even greater than expected and the trend of poor or worse self-rated health prevalence is not as quickly declining in the lower educational group as the unweighted results seem to suggest. These results should be more studied and better communicated to the decision-makers who use the results in monitoring health inequalities.

Narrowing health inequalities among working aged Finns has turned out to be a difficult task. Public health programmes, such as the Health 2015 by the Finnish government, have not been successful in narrowing health inequalities. Providing non-biased data for monitoring the success of these programmes should be also one of the top goals on every program. Promoting health among those in the lower socioeconomic position, in particular, would help improve the overall population health and it would be easier if we would be able to reduce the nonresponse in those groups or at least minimize the effect of nonresponse.

4. Gender inequality and the gender gap in depression
Mrs Malgorzata Mikucka (Université catholique de Louvain, Belgium)

Depression is the leading cause of chronic disability in developed countries, and it systematically strikes women more than men. Past research documented a range of socio-demographic characteristics associated with increased intensity of depressive symptoms. It also showed that the prevalence of depression and the gender gap differ across countries, even controlling for their socioeconomic positions and family situation. Also the gender gap in depression differs across Europe. However, previous research failed to establish a clear link between gender equality and gender gap in depression; studies suggested that men, or only specific categories of women, have lower depressive symptoms in more gender equal societies.

The discrepant results suggest some methodological problems. The first possible source of problems is the use of various measures of gender equality without discussing differences between results obtained with various measures of gender equality. Second, in majority of studies, the effect of gender equality has not been separated from plausible correlated, though separate phenomena: economic development, income equality, and traditional gender roles. Third, although past studies aimed to provide policy-relevant information, they did not control for the unobserved heterogeneity among countries.

This study methodologically improves the investigation of the link between depression in men and women and gender equality, as it accounts for potentially correlated marco factors (economic development, income inequality, traditional gender attitudes), accounts for four different measures of gender equality, and distinguishes cross country differences from overtime change. I use European Social Survey Data for 24 countries (waves 3, 6 and 7) and multilevel regression. The results show that gender gap in depression is smaller in more gender equal countries, but a decline of gender inequality does not correlate with an increase in gender gap in depression. However, gender gap in depression correlates with increasing income inequality. These results suggest that income inequality harms women particularly strongly, whereas both genders benefit from better economic conditions.