Subjective Class Identification in Australia: Do Social Networks Matter?

ABSTRACT

This study investigates the factors that are associated with subjective class identification by using Australian national survey data. Results show that social networks play a significant role in respondents’ subjective evaluation of where they fit in the social hierarchy, with those individuals who have a reference group for social comparison being more likely to identify as middle class. Participating in social clubs, having a sense of neighborhood belonging and being able to access social resources from network members are positively associated with perceived social class. The hypothetical negative effect of social exclusion on subjective class identification is not statistically significant.

KEYWORDS:

• Subjective class identification
• reference group
• social networks
• Australia

Disclosure Statement

No potential conflict of interest was reported by the author.

Notes

1. The literature review shows that some terms are used interchangeably in existing studies, including class identification, class identity, subjective social status, and subjective social location. However, it should be noted that the nuances among these terms might reflect different research traditions across disciplines. For example, the concept of “subjective social status” (SSS) has been widely applied and measured by ladder rankings in medical and health sciences (e.g., Adler et al. 2008; Adler et al. 2000; Präg 2020). For clarity and consistency, this present study used the term “subjective class identification” throughout, based on its theoretical background in sociology as reviewed below.

2. The cases with “None” as the answer for subjective class identification was treated with caution by comparing a few options. They were initially coded as a separate category in the dependent variable, which was estimated by a multinomial logistic regression. Also, they were trialed to be combined with the “working” class and included in the ordinal logistic models or were excluded from the data analysis. As these options did not cause substantial changes to the assumed patterns between the independent variables and the dependent variable, the decision was made to exclude the cases without subjective class identification from the statistical analysis for conceptual and analytical clarity.

3. As an alternative approach, factor analysis was conducted to generate a factor score to represent the three items of neighborhood belonging, which was used in the relevant models and showed similar results (not presented). This contributed to reflecting the robustness of findings.

4. The 20 occupations and their status scores based on the AUSEI06 are as follows: General Practitioner (100), university lecturer (92.30), solicitor/barrister (90.70), engineer (86.10), primary/secondary school teacher (86.10), accountant (83.70), finance manager (81.50), nurse (80.70), journalist (77.40), police officer (64.00), secretary (44.80), clerk in public service (54.30), electrician (39.60), waiter/waitress (36.50), farmer (34.00), shop/sales assistant (30.80), bus/coach driver (29.90), chef (26.60), cleaner (20.40), and factory worker (12.10).

5. It is noted that a multidimensional approach has been adopted in specialized research projects of social exclusion that covered various domains and numerous indicators (Levitas et al. 2007; Scutella and Wilkins 2010). In AuSNet 2014, limited questions were asked about social exclusion as it was not the focus of the survey. While this data constraint should be recognized, the available measurement of social exclusion designed from a social networks perspective is useful.

6. When “subjective class identification” has been measured on a scale, scholars have adopted various methods to treat the variable and fit models (e.g., Chen and Williams 2018; Curtis 2016; Kikkawa 2000; Speer 2016). In this study, generalized ordered logit models and multinomial logistic regression models were also estimated for comparison purposes. These models showed similar results (not presented) to those obtained by the ordinal logistic regression models, and lent credence to the robustness of the findings.

7. Factor analysis offered a single factor score for the four variables of social resources (i.e., the number of contacts in listed occupations, the total status scores of generated occupations, the highest status score of generated occupations, and the range of status scores of generated occupations), which is entered in this model.

8. The results are briefly summarized as follows: (1) on average, respondents who have at least one club membership are 3.3 percentage points more likely than those who do not join any club to identify as upper and upper-middle class, and about 0.9 percentage points less likely to say they belong to middle and lower-middle class and 2.4 percentage points less likely to identify as working class. (2) Regarding neighborhood belonging, one unit increase in the score is associated with an increase of 1.4 percentage points in identifying as upper and upper-middle class. This is offset by a decrease of 0.4 percentage points in identifying as middle and lower-middle class and a decrease of 1 percentage point in identifying as working class. (3) As for social resources, one unit increase in the factor score is associated with an increase of 2.8 percentage points in identifying as upper and upper-middle class. This is offset by a decrease of 1 percentage point in identifying as middle and lower-middle class and a decrease of 1.8 percentage points in working-class identification. All these effects are significant at the 0.05 level. (4) Finally, the effect of social exclusion is not significant (p >0.05).

Additional information

Funding

This work was supported by the Australian Research Council [DP130100690].

Notes on contributors

Xianbi Huang

Xianbi Huang is Senior Lecturer in Sociology at La Trobe University, Australia. Her current research interests comprise social networks, social wellbeing, social stratification, and mobility. She has published articles in journals such as Sociology, Work, Employment and Society, Social Networks, Journal of Sociology, The China Quarterly, American Behavioral Scientist, and Research in the Sociology of Work.