1 / 32

The Effects of Media Forms on Social Capital across European Welfare Regimes: A Multiple Group Structural Equation Model

ESF Quantitative Methods in the Social Sciences (QMSS) Conference Prague, 20 th – 23 rd June 2007. The Effects of Media Forms on Social Capital across European Welfare Regimes: A Multiple Group Structural Equation Modelling Approach (MG-SEM).

dai
Télécharger la présentation

The Effects of Media Forms on Social Capital across European Welfare Regimes: A Multiple Group Structural Equation Model

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. ESF Quantitative Methods in the Social Sciences (QMSS) Conference Prague, 20th – 23rd June 2007 The Effects of Media Forms on Social Capital across European Welfare Regimes: A Multiple Group Structural Equation Modelling Approach (MG-SEM) Boris Kragelj, B. Comm.Léan McMahon, B.Soc.Sc., M.Soc.Sc. Faculty of Social Sciences School of Sociology University of Ljubljana, Slovenia University College Dublin, Ireland Geoffrey Hammond, B.Sc. Department of Psychiatry University of Cambridge, England Quantitative Methods . in the Social Sciences

  2. Background The ideas for this paper were originally conceived during the QMSS course, entitled ‘Theory and Practice in the Analysis of Cross-National Cross-Sectional Data: Structural Equation Modelling’ (August 2005, University of Oxford). The paper was developed and analysis conducted by the researchers in June 2006 in the University of Cambridge, funded by the QMSS Short Visit Competition. Quantitative Methods in the Social Sciences

  3. Structure of presentation BACKGROUND 1. THEORY and LITERATURE • Media effects / Social capital • Macro-structural approaches: Regime types and social capital 2. RESEARCH QUESTION and HYPOTHESES 3. DATA and VARIABLES 4. METHOD and ANALYSIS 5. RESULTS 6. CONCLUSIONS and DISCUSSION Quantitative Methods in the Social Sciences

  4. THEORY and LITERATURE Quantitative Methods in the Social Sciences

  5. Media effects • New media displaces consumption of traditional media forms • TV (when introduced) displaced radio listening and newspaper reading (Coffin 1955; Bogart 1956) • Is the Internet displacing TV? (Dimaggio et al 2001) Quantitative Methods in the Social Sciences

  6. Social capital definitions • What do we understand by ‘social capital’? • Contentious, multi-faceted and slippery term… • OECD definition: “Social capital is networks together with shared norms, values and understandings that facilitate cooperation within or among groups” (Cote & Healy 2001) • “The resources of information, norms and social relations embedded in communities that enable people to coordinate collective action and to achieve common goals” (Shah, McLeod et al 2001: 465) • Putnam (1993) stresses Durkheimian sense of socialisation of individuals into shared norms and cooperative societal action… social networks and norms of reciprocity Quantitative Methods in the Social Sciences

  7. Operationalisation of social capital • Our operationalisation of social capital uses 3 individual-level indicators: • Social participation (Dekker & Uslaner 2001; Putnam 1996; 2000) • Civic engagement [x2] (Brehm & Rahn 1997, 1990; Uslaner 1998) • Generalised trust [x2] (Uslaner 1999; Anheier & Kendall 2002; Sullivan & Transue 1999) Quantitative Methods in the Social Sciences

  8. Media effects and social capital • Television • Social/technological changes destroy community (Tönnies 1887) • TV leads to decline in civic involvement and social activism (Putnam 1996, 2000); dampens participation (Brehm & Rahn 1997) • TV incurs lack of trust (Uslaner 1998) • Critique of Putnam (Skocpol 2003) Quantitative Methods in the Social Sciences

  9. Media effects and social capital • Internet • Lack of longitudinal research and information • Erodes social connections (Kraut et al 1998; Nie & Erbring 2000); takes people away from communities and families (Slouka 1995; Stoll 1995); an “isolating medium” (Nie 2001) • Supports neighbouring and facilitates discussion/mobilisation around local issues(Hampton & Wellman 2003) • Provides information; sphere of political expression (Shah et al 2005) • Influencing changes in society “away from groups and towards networked individualism” (Wellman et al 2003) Quantitative Methods in the Social Sciences

  10. Macro-structural approaches • Structure: Fixed, largely beyond the power of agents to alter • Society-centered approaches: Society’s capacity to produce social capital is determined by long-term experiences of social organisation anchored in history and culture (Banfield 1958; Fukuyama 1999; Putnam 2000; Skocpol 2003) • Institution-centered: Social capital is embedded in and linked to formal political and legal institutions (Berman 1997; Hall 1999; Levi 1998; Rothstein & Kumlin 2001; Stolle 2002) Quantitative Methods in the Social Sciences

  11. Welfare regime theory Countries fit into a typology based on approaches to welfare spending and social rights (Esping-Andersen 1990, 1999; Janoski1998) • Liberal (most English-speaking countries, e.g. USA, Canada, Australia) • Social democratic (e.g. Scandinavian countries) • Corporatist (e.g. continental European countries)

  12. Regime types and social capital • Regime type influences social capital: Political structures shape the institutional means to pursue civic engagement and thereby influence the possibilities for individual action (Skocpol et al, 2000). • All indicators of social capital are reportedly high in social democratic regimes (Kumlin and Rothstein 2003) • Social capital, apart from trust (Inglehart 1999; Uslaner 2002),tends to be high in liberal regimes • Social capital is reported to be fluctuating in corporatist regimes, e.g. Italy (Putnam 1993), Germany (Offe and Fuchs 2002) Quantitative Methods in the Social Sciences

  13. Welfare and social capital: Rationale • Conservatives: Welfare crowds out ‘natural’ networks of association (e.g. families, communities) (Wolfe 1989); social structures erode; commitment and trust decline (Zijderveld, 1998); potentially leads toconsumerism, selfishness, hedonism. • Progressives: Welfare supports networks of association and enhances civil society (Kuhnle and Alestalo, 2000); encourages civic participation (Rothstein, 2001); welfare states alleviate worst excesses of market reform (Rothstein, 2003). • Dependency thesis  welfare induces dependency; erodes social capital • Division thesis  welfare cuts back on mutual obligation; socially divisive Quantitative Methods in the Social Sciences

  14. Past empirical studies • Most analysis of the social impact of the Internet uses cross-sectional data (e.g., Katz & Rice, 2002; LaRose et al., 2001). • Some studies use longitudinal analysis of change, e.g. hierarchical linear models (e.g. Bryk & Raudenbush, 1987, 1992; Singer & Willett, 2003); • Shklovski et al (2004) used cross-sectional analysis and hierarchical linear growth models to estimate effects of Internet use on social participation. Their cross-sectional results suggest that Internet use is associated with more social involvement; longitudinal analysis suggests greater use of the Internet caused respondents to decrease social contacts. • Shah et al (2005) used SEM to test the role of the Internet as a source of political information and a sphere for public expression. • Our study  the samemeasure of social capital and same method for testing all media effects across different welfare regimes Quantitative Methods in the Social Sciences

  15. 2. RESEARCH QUESTION and HYPOTHESES • Research question • To conduct multiple group structural equation modelling (MG-SEM) to test a casual relationship between media consumption and social capital, across five European welfare regimes. • Hypothesis 1 • That Internet usage displaces TV viewing. • Hypothesis 2 • That television viewing and Internet usage have different effects on different dimensions of social capital, as measured in terms of social participation, civic engagement and trust. • Hypothesis 3 • That television viewing and Internet usage are associated with people’s levels of social capital in different ways that depend on the particular political/welfare regime of the country in which they reside. Quantitative Methods in the Social Sciences

  16. WELFARE REGIMES Social participation Internet usage Interpersonal trust TV viewing Civic engagement Theoretical model: Multiple Group comparison with Structural Equation Model Quantitative Methods in the Social Sciences

  17. 4. DATA and VARIABLES • Data • European Social Survey, Round Two (2004) • 24 countries; 5 clusters/regimes • Total sample size=45,681 • Mean sample size per country=1,903 (ranging from 579 (Iceland) to 3,026 (Czech Republic)) Quantitative Methods in the Social Sciences

  18. Variables [Endogenous] Social Ptp. Meet friends, relatives / Social activities / Social organisations/ Helping people Per. TrustMost people… can be trusted vs. you can’t be too careful / try to take advantage vs. be fair / helpful vs. look out for t’selves Per. Civ. Eng. Voted / Wore campaign badge / Contact politician / Signed petition / Boycotted products / Took part in protest, demo Inst. TrustIn… Parliament / Legal system / Police / Politicians / Political parties / European Parliament / United Nations Inst. Civ. Eng.Member political party / Worked political party, action gp. ------------------------------------------------------------------------------------------------------- [Exogenous] TV viewingHours per dayInternet Use the Internet, WWW or Email ------------------------------------------------------------------------------------------------------- Demographics: Age Year of birth Education Years of full-time education completed Income Household’s total net income, all sources Quantitative Methods in the Social Sciences

  19. Clusters of European countries (N=24) SOCDEM LIB/ CORP ‘Mixed’ Finland UK Austria Denmark Ireland Belgium Norway France Sweden Germany Netherlands Luxembourg Iceland Switzerland (Esping-Andersen 1990; Janoski 1998; Olafsson 2003) MEDTERR Greece Portugal Spain (Saint-Arnaud & Bernard 2003; Flaquer 2000; Ferrera 1996) POSTCOMCzech Rep. Estonia Hungary Poland Slovakia Slovenia Ukraine (Deacon 1993, 2000) Preferable label: ‘Emerging democratic’? Quantitative Methods in the Social Sciences

  20. 3. METHOD and ANALYSIS • Multiple group structural equation modelling (MG-SEM) on data from the ESS (2004) using MPLUS (maximum likelihood estimation (MLE)); • Analysis was divided into 4 sections: • Construction of measurement model for social capital (5 latent correlated factors, single SEM model for whole sample) • Estimation of media effects on social capital factors, controlled for spurious effects of demographic variables (single SEM model for whole sample) • Comparison of social capital factors (means) across welfare regimes (Multiple group mesurement model) • Estimation of media effects across welfare regimes (full multiple group SEM model) Quantitative Methods in the Social Sciences

  21. 5. RESULTS Quantitative Methods in the Social Sciences

  22. Measurement model of social capital Quantitative Methods in the Social Sciences

  23. Social capital variable: CFA * Categorical variables Quantitative Methods in the Social Sciences

  24. Factor correlations Quantitative Methods in the Social Sciences

  25. Error terms (co-variances) between factor items POSTCOM SOCDEM Quantitative Methods in the Social Sciences

  26. Full SEM (MIMIC) one group model Tests of model fit: CFI 0.929 TLI 0.941 RMSEA 0.042 WRMR 5.869 Quantitative Methods in the Social Sciences

  27. Causal model: Estimates and model fit statistics * categorical Quantitative Methods in the Social Sciences

  28. Multiple group SEM model: Media effects across welfare regimes * Significance 5% Tests of model fit: CFI 0.906 TLI 0.948 RMSEA 0.043 WRMR 4.299 Quantitative Methods in the Social Sciences

  29. Means of social capital factors Quantitative Methods in the Social Sciences

  30. 6. CONCLUSIONS and DISCUSSION Quantitative Methods in the Social Sciences

  31. Conclusions • H1: • Yes. Internet usage displaces TV viewing. • H2: • No. Television viewing and Internet usage donot have different effects on different dimensions of social capital, as measured in terms of social participation, civic engagement and trust.Internet usage increases, while TV diminishes, all forms of social capital. • TV watching news/politics/current affairs programmes is more akin to Internet usage than TV watching in general, in terms of the effects on social capital. This may be due to the more interactive and less escapist nature of such programming. • H3: • For the most part, no. The welfare regime of the country in which people reside mostly does not affect the relationship between media consumption and social capital. • However, exceptions In Liberal/Mixed regimes, no media form has an effect on trust… • In post-Communist and Mediterranean countries, both TV and Internet have positive effects on trust • OTHER: • But level of (different forms of) social capital varies significantly across welfare regimes… Quantitative Methods in the Social Sciences

  32. Discussion • Methodological discussion: • Problems with running full SEM multiple group model (no convergence) • Problems of multi-colinearity with media consumption suggest better measurement… • Future research: • Level of (different forms of) social capital varies significantly across welfare regimes. To explain the differences in social capital across welfare regimes… applying multilevel analysis. Quantitative Methods in the Social Sciences

More Related