Answering these research questions extends previous findings by considering different operationalizations of SES, study field and data collection methods.<\/p>\n
Methods<\/h1>\nInformation sources, search strategy and study selection<\/h2>\n
A systematic literature search was carried out using the databases Web of Science, PubMed, PsycINFO, Psyndex, IBZ-Online and OLC Sozialwissenschaften up to July 2021. The literature search was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Moher et al., 2009). The search strategy was developed collaboratively, combining synonyms and related terms of objective and subjective socioeconomic status (detailed search strategy see S1). Title\/Abstract and full text screening were performed independently by one of the authors and a research assistant based on the predefined eligibility criteria. All dissents were resolved by consensus.<\/p>\n
Eligibility criteria and data extraction<\/h2>\n
Studies in English and German that measured both objective and subjective status and reported a bivariate correlation between the two indicators were included. The population of interest was the general population aged 18 years or older. Studies that only interviewed students were excluded since data collection on OSS is limited for this group, as students are homogeneous with respect to the OSS indicators. Meta-analyses, Literature reviews, experimental studies manipulating indicators of SES, clinical trials or case-studies, and proxy reports were also excluded. Data extraction was divided between one of the authors and a research assistant, both cross-checking each other’s work. Dissents were resolved by consensus.<\/p>\n
Quality appraisal<\/h2>\n
The quality of the included studies was determined using the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies with 14 quality criteria (National Institute of Health, National Heart, Lung and Blood Institute). This tool comprises a subjective quality rating (good, fair or poor) of the individual papers, which was quantified based on the numbers of positively assessed quality criteria (good: met all applicable categories; fair: met most categories and low: met minority of categories). The quality of the eligible studies was assessed independently by two researchers. Disagreement and indecisions were resolved by discussion. Most of the included studies (n=63) were rated good on overall quality; 5 of the studies were rated as fair quality (for a more detailed overview of the quality of the included studies see table S2).<\/p>\n
PRISMA flow chart of study selection<\/h2>\n
A total of 2001 records were identified up to July 2021. After deleting 801 duplicates, 1200 articles remained, of which 832 articles had to be excluded after abstract screening. One reason for excluding these articles was that participants did not belong to the target population, the target population was not clarified, or the studies addressed a single homogenous group (n=349). For other excluded abstracts, the SES was not considered (n= 157), or either OSS or SSS was addressed (n=282). After screening the remaining 368 full-texts, 300 articles had to be excluded because no bivariate correlations between OSS and SSS were reported (n=250), the study included only one measure of status (n=32), or a full-text was not available (n=18) (see Figure 1). Of the eligible publications, 68 matched the pre-defined inclusion criteria and were included. These publications contained a total of 89 studies with 253 bivariate linear correlations between OSS and SSS. However, six of these correlations were excluded from subsequent analyses as duplicates, resulting in 249 bivariate linear correlations being included in the analysis.<\/p>\n
Figure 1 PRISMA flow chart of study selection<\/em><\/p>\n Different operationalizations of OSS (Income, Education, OSS Index) and Income (Household Income, Personal Income) as well as Education (Highest Level of education, Years of education) and an OSS Index (three indicators: Income, Education, Occupation; two indicators\u00a0: Income with Occupation or Income with Education) were reported in the studies and therefore included in the analysis. It is worth noting that different operationalizations for occupation were not investigated due to the predominant use of occupational prestige in the evaluated studies. Similarly, different SSS operationalizations were not investigated as the majority of studies used the MacArthur Scale.<\/p>\n Several additional moderators potentially influencing the correlation between OSS and SSS were included in the analysis. In terms of study context, the country setting (USA, Europe, Asia and Other) and research field (Health\/Neurosciences, Psychology, Social Sciences) were considered. Regarding study design and methodology, the study type (cross-sectional, longitudinal), sampling strategy (non-random and random sampling), and data collection mode (Face to Face, Telephone, Paper and Pencil Self-administered, Online) were included.<\/p>\n Data analysis was conducted using R version 4.3.1 and the metafor package, implementing multi-level random effects model, to account for several studies clustered in single publications. With regard to the first research question, the cumulative effect size of the correlations between objective and subjective socioeconomic status was estimated to be small with a coefficient of r=0.30 (SE = 0.02; Z = 19.07, p<0.001). The methods used were Pearson\u2019s correlation (n = 17), Spearman\u2019s correlation (n = 8) and different kinds of correlation dependent on the scale of the variables (n = 3). Other studies did not report the kind of correlations they employed, but all use linear correlations as descriptive statistics for linear regressions and Structural Equation Modeling (SEM). The Q-statistic test for heterogeneity resulted in a significant outcome (Q (df = 247) = 11911.42, p < .001), suggesting substantial variability in the effect sizes across the studies. The observed heterogeneity suggests that there may be underlying factors that contribute to the variation in the correlation between OSS and SSS across different studies.<\/p>\n The majority of the 249 investigated correlations were derived from studies conducted in the USA (n=146), followed by Europe (n=65) and Asia (n=26). In terms of research field, the highest number of reported correlations came from Health and Neurosciences (n=122), followed by Psychology (n=109) and Social Sciences (n=18). Of the total correlations reported, the vast majority (n=243) stemmed from cross-sectional studies, with only a small number (n=6) from longitudinal studies. Non-random sampling was the most common approach (n=165), followed by random sampling (n=50). Most correlations were derived from face-to-face surveys (n=68), followed by online (n=57) and self-administered surveys (n=34). OSS was mainly operationalized using income (n=103), with the majority using net household income (n=75), and education (n=101), particularly the highest level of education attained (n=73). (See table 1 for a detailed overview of predictors included in the analysis). SSS was primarily operationalized via the MacArthur society ladder (n=160) or MacArthur community ladder (n=42). Our study thus primarily investigates SSS as measured with the MacArthur scale: \u201ca summative judgment of one\u2019s socioeconomic position across the dimensions education, occupation and income\u201d (Adler et al. 2008).<\/p>\n Moderating effects were analyzed using single multi-level meta-regressions including a moderating variable. Table 1 presents the results of the investigated moderators that potentially influence the size of the correlation between OSS and SSS.<\/p>\n Examining the operationalization of OSS (research question 2), the highest correlations with SSS of the size of .49 were observed when OSS was measured using an index comprising two or three objective status components, which was higher than measurement by income, whereas the lowest effect size was obtained for education. In addition, occupation and education are less strongly correlated with SSS than is income. Notably, the correlation between OSS and SSS was particularly strong when income was measured by household income as opposed to individual income. Regarding education, the highest level of education showed a stronger correlation with SSS than the number of years of education. With regard to the third research question, an analysis of the country setting revealed a significant effect, indicating a lower correlation between OSS and SSS in Asia compared to the USA, Europe, and other regions. Furthermore, in terms of the research field, the correlation between OSS and SSS was found to be significantly higher in the field of psychology than in health\/neurosciences and social sciences. No significant effects were found for study type (cross-sectional vs. longitudinal), sampling strategy (random sampling vs. non-random sampling), or mode of data collection (face-to-face, telephone, self-administered, online).<\/p>\n Table 1 Potential factors influencing the correlation (Mean Estimate) and Confidence Intervals (CI) between OSS and SSS<\/em><\/p>\n![\"\"](\"https:\/\/surveyinsights.org\/wp-content\/uploads\/2024\/04\/Figure_1.png\")
<\/a><\/p>\nModerators<\/h2>\n
Results<\/h1>\n
Correlation between subjective and objective socioeconomic status<\/h2>\n
Factors potentially influencing the correlation between OSS and SSS<\/h2>\n