Social Capital & Trust |
|
| Chair | Dr Jan Mewes (Department of Sociology, Umeå University ) |
A core dimension of social capital is access to resources through one's social network. However, how to measure and operationalize social resources is still under debate. The first set of strategies focus on potential resource access: Who does the respondent know and which resources do these ties control? The second set of strategies focuses on actual resource mobilization: Which resources did respondents receive from their network? Both of these approaches have their merits and their shortcomings. Concentrating on potential resource access means that we rely on respondents' ability to correctly estimate which resources are within their reach. Actual resource mobilization, on the other hand, faces the problem of self-selection: we ignore the social capital that is potentially available to those who did not need it in reaching their goal.
The German National Educational Panel Study (NEPS) is the first nationally representative large-scale panel study that contains detailed information on the social capital of youths living in Germany who are facing the transition to vocational education and training (VET). The data contain prospective as well as retrospective information on resource access and mobilization for all secondary school graduates. Resources for the transition to VET include referrals, information about potential VET options, and help in writing an application. Prospective information is collected in the final year of schooling. Respondents are asked to estimate the chances of getting access to each of the social capital resources, as well as several follow-up questions about potential resource providers. Retrospective information on actual resource mobilization is collected in the fall after graduation. All respondents, irrespective of their transition status, are asked whether they actually received any of the resources, and are again asked follow-up questions about the resource providers. Both the prospective and retrospective questions are repeated yearly until respondents have made a successful transition to VET. Thus, the data offer a unique possibility to compare prospective and retrospective social capital information.
I focus on two research questions: Firstly, what is the relationship between prospective and retrospective social capital, i.e., between access and mobilization? Secondly, how are access to and mobilization of social capital related to the transition to VET? Preliminary results show that beliefs about access to resources and actual resource mobilization are only weakly correlated, but this correlation increases over time, indicating that respondents become more knowledgeable about their social network. There are strong signs of self-selection when it comes to mobilized social capital: Youths who struggle to find an apprenticeship are more likely to mobilize social resources. On the other hand, youths' estimations of their potential access to social capital show only a weak relation to their transition rates. This is most likely due to the fact that youths simply do not have a very good idea about their ability to utilize their social network for career-related resources. I will conclude by assessing what can be learned from the combined information of prospective and restrospective accounts of social capital.
Access Research Knowledge (ARK) undertakes a suite of cross-sectional public attitudes surveys, all of which are used in developing and evaluating key government policies in Northern Ireland. These comprise the Northern Ireland Life and Times Survey of adults, the Young Life and Times Survey of 16 year olds, and the Kids’ Life and Times Survey of 10-11 year olds.
Northern Ireland has taken an outcomes approach to evaluating policy and strategies, resulting in a set of indicators for each policy/strategy. Data from ARK’s surveys are integral to the sets of indicators associated with many key government strategies, not least being the proposed Programme for Government 2016-21. In particular, ARK time-series data has been used within several good relations policies, which address good relations between the two main ethno-religious communities in Northern Ireland (Catholic/nationalist and Protestant/loyalist).
In this paper, we will draw upon our experiences and challenges of using time-series data to explore social change in Northern Ireland. In particular, we will weigh up the usefulness of indicators approach with the limitations it places on survey research. Two key issues have emerged. First, the tensions involved in writing and fielding questions that are valid over a long period of time and in changing social and political landscapes. Second, the experience of undertaking academic survey research that is commissioned by government for policy purposes.
According to the seminal work of Bo Rothstein and Eric Uslaner, universal welfare states provide the ideal breeding ground for one of the most important parts of social capital: generalized trust. Means-tested welfare programs, by contrast, are believed to undermine people’s trust in others. We lack, however, longitudinal evidence showing that changes in those two types of social spending, means-tested vs. universal, are connected to changes in generalized trust. This study seeks to close this literature gap by making use of longitudinal time-series data from the cumulative cross-national European Social Survey study 2002 to 2014. On the country-level, I use detailed EUROSTAT data distinguishing between means-tested and universal spending during the period 2002 to 2014. To disentangle cross-sectional and longitudinal relationships between social spending and generalized trust, the social spending variables enter the equation twice, one variable capturing the respective country’s mean spending and the other one measuring changes within the respective country over time. Applying three-level MCMC hierarchical regression analysis (level 1: individuals, level 2: country-time, level 3: countries), I show that changes in means-tested welfare state spending are negatively correlated with changes in generalized trust, thus further corroborating the hypothesis that, when compared to welfare states making more generous use of universal social spending, residual welfare states fare worse in producing generalized trust. This result remains even robust when including an additional random slope for time. By contrast, I do not find a significant correlation between changes in universal social spending and changes in generalized trust. The latter result gives further support for the recently put forward ‘reverse causality’ hypothesis that it is generalized trust that precedes the development of universal welfare states, and not vice versa.
A large body of studies provides vast but mixed evidence on the determinants of generalized trust and their causal relations. The first and most obvious explanation for such inconsistency is that scholars often use different datasets and methods of analysis. The second and more profound explanation concerns the issue of how generalized trust is conceptualized and measured. The traditional trust question asking about trust in most people utilized in the greater part of existing surveys has been criticized over last decade. The most often raised critique concerns the fact that it has been taken for granted that the notion «most people» refers to strangers and members of out-groups. Yet, recent studies demonstrate that this is not necessarily the case (Sturgis & Smith, 2010; Reeskens 2012; Delhey, Newton, and Welzel 2011). While the trust radius is a widely pronounced limitation, there has been little discussion about another, similarly important aspect. Instead of capturing trust and distrust, the traditional question might rather evaluate the perceived trustworthiness of the general social environment of a person (Nannestad, 2008 Helliwell & Putnam, 2004) and caution (Miller & Mitamura, 2003). Country specific evidence on the US and Japan shows a gap between indicators of trust and perceived trustworthiness (Miller & Mitamura, 2003).
To this date, corresponding cross-cultural evidence on this gap, as well as, its detailed exploration in terms of micro and macro-foundations is inexistent. Utilizing the 5th wave of the WVS, we combine traditional trust question and question about trust in strangers differentiating four possible types: a) Trusting the trustworthy b) not trusting the trustworthy c) trusting the untrustworthy and d) not trusting the untrustworthy. Descriptive analyses thereby indicate that trust and trustworthiness do not necessarily mirror each other perfectly. In addition, CFA indeed indicates that these two questions hardly belong to the same dimension.
Such discrepancy is not only an issue of measurement. Thorough investigation may shed additional light on the concepts of «trust», «trustworthiness» and help to differentiate «rational» and «norm-driven» or «moral» trust (Ashraf, Bohnet, & Piankov, 2006; Mansbridge, 1999; Nannestad, 2008; Nooteboom, 2002; Stolle, 2002). “Rational” approach is based on the calculation of possible losses and gains and implies uncertainty, risk and caution. “Norm-driven” approach sees trust as a default attitude which lies in the commonality of moral values and norms. Following this idea trusting the untrustworthy reflects rational trust, while trusting the trustworthy - moral. To test this hypothesis we conducted multinomial regression analysis that demonstrated that these types have different foundations.
Using Schwartz` values and a set of important variables, we found that those who have higher level of universalism, protest participation and confidence in political institutions are more likely to have «norm-driven» trust. The evidence for trusting the untrustworthy is mixed. We detected no statistically significant difference between trusting the trustworthy and trusting the untrustworthy in relation to stimulation/risk taking. Therefore, the act of trusting implies a certain amount of risk no matter of the perceived of trustworthiness of possible partners.
Measuring how volunteering contributes to the UK Economy is a subject of interest to various Government departments and sports and volunteering charities in the UK. In light of this, the Office for National Statistics (ONS) is researching the possibility of including a question on volunteering on the 2021 England and Wales Census.
A question was designed for inclusion in the 2017 Census test, using the definition of formal volunteering from the UK Government which is giving unpaid help for official clubs, groups and organisations. Guidance was included telling respondents not to include paid activities or activities they are required to do. The answer options for respondents were broad, asking how many hours a week they spent volunteering over a reference period of 12 months.
The 2021 Census will be carried out predominantly online, so for cognitive/usability testing the volunteering question was embedded in a short on-line questionnaire about activities such as caring and paid employment. Purposive sampling was used to recruit 30 participants in England and Wales. Participants were volunteers for a range of public-facing and non-public-facing organisations. Ex-volunteers and non-volunteers were also represented in the sample. Participants who volunteered ranged from people who volunteered regularly, on an ad-hoc basis and people who only volunteered in the school term time. Cognitive testing was conducted to explore ability to answer, accuracy of answer given and willingness to answer. Recall over the 12 month reference period was also tested and differences between participants who volunteered regularly and on an ad-hoc basis were explored.
The findings of the above testing will be outlined regarding wording of the question stem, guidance and the response options. The acceptability of asking a question about volunteering on the England and Wales Census will also be presented.
Future iterations will test different approaches to the volunteering question and the results from these tests will also be presented and discussed. The pros and cons of asking volunteers to indicate how many hours they volunteer a week versus how frequently they volunteer a week over a reference period of 12 months will be discussed.