Class Imagery and Subjective Social Location During Iceland’s Economic Crisis, 2008–2010

ABSTRACT

It has long been argued that perceptions of the class structure shape how individuals see their position in society. However, researchers using nationally representative data to study subjective social location rarely include class imagery in their models. This study uses OLS regression of two surveys, conducted 18 months apart following Iceland’s 2008 economic collapse, to analyze the effects of class imagery on subjective social location. Results show that respondents who perceive the class structure as inegalitarian report lower subjective social location as compared to those with egalitarian class imagery. The influence of objective class position on subjective social location depends, partly, on class imagery. The study finds that Icelanders saw the class structure become more egalitarian and their own subjective social location climb following the collapse, even as the economic crisis deepened between the two surveys. These findings reflect that income inequality in Iceland dropped sharply after the collapse.

Notes

¹ As Evans and Kelley (2004) discussed, describing self-placement in a social hierarchy is a contentious issue among social scientists. In this study, subjective social location is measured with an abstractly worded top-bottom self-identification question asking respondents to place themselves in a social hierarchy with ten rungs (Smith 1986). Some researchers refer to placements such as this as “subjective class” (Evans, Kelley, and Kolosi 1992). Other researchers have reservations about measuring subjective class with a question that does not explicitly evoke the word “class” and/or does not mention ownership of the means of the production (cf., Marx 1977), market situation (cf., Weber 1978), and so forth. I do not engage this debate here. Instead, I use the ten-rung question in this study to examine patterns of subjective social location (which I see as a broad measure, tapping elements of both class and status) and the social factors that shape such patterns. For clarity, I talk primarily of “subjective social location” when referring to my analysis.

² The former survey is referred to as the “2008 Survey” and the latter as the “2010 Survey.”

³ Multicollinearity is a potential problem when including interaction variables in OLS regression models. Even though the regression estimators are still the best linear unbiased estimators, multicollinearity can make these estimators have large variances and become sensitive to even small changes in the data (Gujarati 2003). To get around this potential problem, I centered the household income variable before including it in the regression models. More specifically, I shifted the values of the variable around its mean, that is, I set the mean to zero in the new variable, before creating the interaction variable. The centering method is very effective in reducing multicollinearity (Robinson and Schumacker 2009). Collinearity diagnostics, such as variance inflation factor, were used to detect possible multicollinearity issues in the models. None of the regression models have multicollinearity issues, as the largest variance inflation factor is 2.095 (household income in Model 5), which is well below common benchmarks such as 10, 5 or even 2.5 (Gujarati 2003:362). In addition, the distribution of regression residuals was examined by mapping a histogram and normal P-P plot in order to detect possible violations of the normality assumption. No violations were found.

⁴ I also ran regression models using interaction variables of class imagery and economic class. However, these effects proved insignificant and did not improve the fit of the models. These variables are not included in the models, but the results are available from the author upon request.

Additional information

Notes on contributors

Guðmundur Oddsson

Guðmundur Oddsson is Assistant Professor of Sociology at Northern Michigan University. His research focuses on social control, deviance, and class inequality, particularly the subjective dimensions of class. His work has appeared in The Sociological Quarterly, Acta Sociologica, International Journal of Sociology, and Social Policy.