Universal Basic Income Universally Welcomed? – Relevance of Socio-Demographic and Psychological Variables for Acceptance in Germany

1.1 Gender, Age, and Education

Gender is much discussed in the context of a UBI (emancipation, gender pay gap, etc.). However, the empirical evidence on the association between gender and UBI acceptance is inconclusive: Vlandas (2021) finds slightly higher UBI support among men, which he did not expect based on the literature consulted, which rather suggests higher UBI support among women. Roosma and van Oorschot (2020) and Shin, Keppnanien, and Kuitto (2021) find no gender effect based on the same data (European Social Survey, 2016, 8th wave), neither do Adriaans, Liebig, and Schupp (2019) and Busemeyer, Rinscheid, and Schupp (2023) for the German resident population. However, even though gender may not be relevant for UBI support in resident populations, this could be different for the working population, which does not include pupils, students, and probably most importantly pensioners. Gender-specific differences in care work, income, and working hours are thus likely to be much more pronounced in the working population compared to the full resident population. As these are plausibly relevant factors shaping individuals’ expectation on how a UBI might influence their personal circumstances, gender may still be a relevant predictor in the working population. It is often argued that a UBI could provide women, who are still often earning less than men, with financial security, potentially strengthening emancipation as women’s financial independence increases (Schulz 2017). Additionally, it is likely that families expect a UBI to significantly increase their household income. Consequently, reducing regular working hours and using the UBI to compensate for the resulting loss of income may be an attractive option that allows for more flexibility in families. Assuming that women are still primarily responsible for childcare, it is plausible that they would particularly benefit from this gained flexibility. On the one hand, because more family time is created in which women are relieved of their care duties. On the other hand, because more care time can be covered by the partner, creating free space for women to use for their own purposes (professional or private). Taken together, we therefore find it plausible that women in the working population would expect greater personal benefits from a UBI than men. Accordingly, we assume that women are more likely to support the idea of a UBI and hypothesize for our data of the German working population:

In Germany, as a country with a relatively old and comparatively well-educated population, age and education play a significant role in the UBI discussion. A study conducted by Adriaans, Liebig, and Schupp (2019) revealed significantly higher UBI acceptance among people of younger age and higher education in Germany. Similar results have been found for Europe. Vlandas (2021) also found younger age and higher education to be significantly associated with higher UBI support, while Roosma and van Oorschot (2020) found the same age effect, but the opposite for education and Shin, Keppnanien, and Kuitto (2021) only found the same age effect. Findings by Busemeyer, Rinscheid, and Schupp (2023) point into a similar direction. Their analysis of two representative samples of the German resident population shows that compared to people above 65 years old, UBI acceptance was significantly higher among people between 25 and 45 years old, while there was no significant effect for people below 25 years and people between 46 and 65 years old. In tendency, these results suggest that UBI acceptance is higher among younger people in the working population. Thus, for our sample of the German working population, we hypothesize:

1.2 Perceived Economic Situation and Income

As a UBI would constitute an additional source of income, it is intuitively plausible that the current amount of income may have an influence on UBI acceptance. Vlandas (2021) and Shin, Keppnanien, and Kuitto (2021) found lower income to be associated with higher UBI support using the European Social Survey data. Results by Adriaans, Liebig, and Schupp (2019) and Busemeyer, Rinscheid, and Schupp (2023) for Germany also suggest higher UBI acceptance among those who would directly financially profit from a UBI.

However, a distinction must be made between the objective and subjective conceptualization and operationalization of people’s economic situation. It is possible that people are objectively in a privileged economic situation but perceive their economic situation as difficult. Accordingly, the objective and subjective assessment of people’s economic situation can diverge. The relevance of this differentiation is particularly important in cases of difficult economic situations, as Sinclair and Cheung (2016) show in their overview and Klug, Selenko, and Gerlitz (2021) in their current study for data from Germany. With regard to a UBI, research by Roosma and van Oorschot (2020) finds that unemployment (objective) but also a lower subjective income (subjective) correspond with higher UBI acceptance. Thus, we hypothesize:

1.3 Corona Perception

The COVID-19 pandemic was an “economic shock” (Hensher 2020, p. 1) and quickly became one of the largest global economic crises. This led to a large increase in macroeconomic (Carriero et al. 2021) and subjectively perceived economic uncertainty (Thum 2021). As times of financial insecurity often call for a change in policies, it is not surprising that the idea of a UBI has been discussed extensively among mainstream media since the onset of the pandemic (Hampel 2020). Both public opinion research insitutes (e.g., YouGov, WeMove Europe 2020) and academia (e.g., Nettle et al. 2021; Van Hootegem and Laenen 2023) have developed an interest in examining how support for a UBI changed during the COVID-19 pandemic and, if so, why.

Nettle and colleagues (2021) found that study participants who rated it as important to have a social security system that reduces stress and anxiety show higher UBI support. It seems likely that people who perceive the pandemic as threatening in that they expect negative consequences for themselves and people around them are more likely to experience anxiety and therefore more likely to value systems that reduce stress and anxiety. Thus, it seems plausible that these people also show higher UBI support. Busemeyer, Rinscheid, and Schupp (2023) analyzed two large representative samples of the German resident population and found that greater concern about one’s economic situation was significantly associated with higher UBI acceptance. Plausibly, a high perceived level of threat posed by the corona-pandemic includes concern about one’s economic situation and is thus also associated with higher UBI acceptance. Following this, we hypothesize:

1.4 Equality Orientation

Since a UBI would be paid unconditionally to all citizens, thus treating all citizens equally regardless of their characteristics or proof of indigency, it is reasonable to assume that people associate an increase in equality with the implementation of a UBI. Thus, it is plausible that people who inherently care about equality would also be in favor of a UBI. This is supported by Schwander and Vlandas (2020) in their analysis of the European Social Survey data from 2016 (8th wave) and Adriaans, Liebig, and Schupp (2019) who analyzed data from Germany and found that people who perceive the needs of the most vulnerable in society as not being met are more likely to support a UBI. Therefore, we assume:

1.5 Performance Orientation

A representative poll in Germany found that 80 % of participants perceive it as fair that people who perform well make a lot of money (Informationsdienst des Instituts der deutschen Wirtschaft 2019). “Performance must pay off” appears to be a central motive of performance-oriented societies such as Germany, in which we often also see a relatively high centrality of work in life (Lu et al. 2019). Thus, it is not surprising that in countries with higher average employment commitment, the acceptance of a UBI, which is need- and not performance-oriented (Segbers 2019), is lower (Kozák 2021). For Germany, Adriaans, Liebig, and Schupp (2019) found that people who are more likely to support the performance principle as a social distribution rule are less likely to support a UBI. Thus, we hypothesize:

1.6 Job Satisfaction

Low job satisfaction scores are associated with higher absenteeism rates (Scott and Taylor 1985; Čikeš, Maškarin Ribarić, and Črnjar 2018 for a review), and turnover intentions (Chen et al. 2011; Hom et al. 2017). However, people who are not satisfied with their professional situation often do not quit their jobs due to extrinsic restraints such as financial necessities (Herzberg et al. 1993; Suadicani et al. 2013). An unconditional regular payment could free these people from financial boundaries and their unsatisfying work situation. Following this, we hypothesize:

1.7 Further Variables

Apart from the abovementioned, further variables, i.e., people’s care duty for children, regional origin in Germany, employment status, unemployment frequency, leisure satisfaction, and loneliness might also be relevant for UBI acceptance. Since these variables were included in the full survey dataset but cannot profoundly be grounded in the literature, we investigate them exploratively.

The introduction of a UBI is one possible way to address economic insecurities. However, many unanswered questions, including the issue of acceptance, remain. Our research contributes to the ongoing UBI discussion by providing insights into socio-demographic and psychological variables associated with UBI acceptance among the German working population based on recent data from the COVID-19 pandemic.

2.1 Participants and Procedure

For the present study, we used a subsample of the 2020 data from a large representative multi-topic survey of the German-speaking resident population on physical and mental health that is conducted annually by the University of Leipzig. The 2020 survey was approved by the institutional ethics review board of the University of Leipzig (ref. 043/20-ek) and followed the ICH-GCP-guidelines. It was examined from ethical, medical-scientific, and legal points of view and was approved with no further restrictions.

The compilation of the survey sample and the data collection were carried out by USUMA GmbH, an independent institute for market and social research. The data were collected between May and June 2020. To ensure representativeness of the sample for the German-speaking resident population, the sample was randomly drawn using an area-based sample provided by the ADM (Arbeitskreis Deutscher Markt- und Sozialforschungsinstitute e.V. [Working Group of German Market and Social Research Institutes e.V.]) covering all inhabited areas in Germany. Within a total of 258 regional areas, 5418 households were randomly selected using a random-route-procedure and one target person was randomly identified in each household using a Kish-Selection-Grid. Of 5418 identified persons, 2521 could be reached and were interviewed.

All participants were informed about the study procedure, the voluntary nature, the right to withdraw, the strict confidentiality as well as the anonymization of all personal data. After giving informed consent, participants completed guided structured face-to-face interviews with trained interviewers at their homes. The interviews consisted of two parts. First, socio-demographic questions were read out loud and participant answers were noted down by the interviewer. Second, participants independently, and without any further moderation by the interviewer, filled in a questionnaire. A total of 18 interviews were unusable and directly excluded by USUMA so that the full survey dataset provided by USUMA comprised 2503 participants. Table S1 in the Supplemental Material provides details on the response rate.

For the present study, we used a subsample of the full survey dataset compiled by USUMA, which covers only the German working population (details on participant exclusion are provided in the analytic strategy section). The final sample size was N = 1986. The mean age was 44.4 years (SD = 13.64). Of the participants, 49.3 % were female, 15 % had a university degree, 32 % reported a personal net income of 2000 Euro or more, and 75.8 % reported being employed at the time of the study. Figure 1 provides an overview of the sampling procedure.

Figure 1:

Sampling and variable selection procedure.

2.2 Measures

The full survey covered a wide range of topics (e.g., politics, health, religion). For the present study, selected measures were used. Those are reported below. Note that the survey was conducted in German. Thus, an English translation is provided here.

2.3 Analytic Strategy

Data processing and all analyses were done in SPSS Version 27.0.1.0 (IBM 2020) and R Version 4.0.4 using the following packages: lavaan, Matrix, survey, spatstat, and weights. To create a subsample of the German working population, participants below 18 and over 67 years (n = 406), as well as university students and pupils (n = 190), were excluded. Furthermore, due to the small number of cases, participants who indicated a diverse gender (n = 1) were also excluded. Note that 80 of the excluded participants met several exclusion criteria. In total, data from 517 participants were excluded and the final sample comprised N = 1986. Participants with missing values in at least one of the variables of interest were identified (n = 462). A sensitivity analysis with G*Power Version 3.1.9.6 (Faul et al. 2009) indicated that the sample size provides sufficient power to detect effect sizes as small as f ² = .003 with a power of .80 and an alpha error probability of .05.

The representativeness of the full survey sample for the German-speaking resident population was ensured by USUMA through the sample compilation procedure (see Section 2.1). However, dropouts may have limited the representativeness of the sample. Comparing the data with the known distributions from the German microcensus on household and person level revealed slight deviations (the microcensus is the largest annual household survey in Germany conducted by the German Federal Statistical Office (Statistische Ämter des Bundes und der Länder 2022), participation is mandatory by law and thus self-selection bias is not an issue). To account for this, statistical weights were applied to the data to correct for these deviations and ensure representativeness of the data. Specifically, the statistical weighting corrected at household level for deviations in the distribution of household sizes and at the personal level for deviations in the joint distribution of age and gender as well as deviations in the distribution of the state of residence (USUMA 2020). All analyses were performed on the weighted dataset of N = 1986.

To answer the research question and hypotheses, a four-step hierarchical multiple linear regression was performed defining UBI agreement as outcome variable. The other variables were used as correlated predictors. The four models of the hierarchical multiple linear regression were specified and tested in lavaan as path models. Although the model specifications as path models were identical to the regression models, path modeling allows to use maximum likelihood estimation which has fewer assumptions than ordinary least square regression and allows to better address missing data via using a full information maximum likelihood estimator (Enders and Bandalos 2001).

The order of variable entry for the hierarchical regression was based on two criteria: the assumed sequence of effects between predictor variables (socio-demographic predictors before psychological predictors) and the importance of predictors in explaining the outcome (hypothesized effect before exploratory investigation). This means that, first, socio-demographic predictors (e.g., Age) entered the model before psychological predictors (e.g., Job satisfaction) and second, predictors for which hypotheses have been formulated (e.g., Age or Job satisfaction) entered the model before predictors that should be examined exploratively, i.e., without hypotheses (e.g., Regional origin or Loneliness). Thus, socio-demographic predictors were entered in steps 1 and 2, with “hypothesized-effect” predictors preceding “exploratory” predictors (step 1 included “hypothesized-effect” socio-demographic predictors: Gender, Age, Income, Education; step 2 included “exploratory” socio-demographic predictors: Care duty (children), Unemployment frequency, Regional origin, Employment status). In steps 3 and 4, psychological predictors were entered accordingly, i.e., with “hypothesized-effect” predictors preceding “exploratory” predictors (step 3 included “hypothesized-effect” psychological predictors: Job satisfaction, Perceived economic situation, Performance orientation, Corona perception, Equality orientation; step 4 included “exploratory” psychological predictors: Leisure satisfaction, Loneliness).

Multivariate normal distribution was tested. Potential suppression effects were evaluated and relative weights of predictors as a percentage of R ² were computed (Bhalla 2020; Johnson 2000). In order to estimate relative weights, a model implied correlation matrix was obtained for each path model. This variance is based on the sampling weights and full information maximum likelihood. To obtain the relative weights, the correlation matrix was entered into a Shiny App by Tonidandel and LeBreton (2015). Effect sizes were calculated and interpreted according to Gignac and Szodorai (2016) with r = .10 being a small, r = .20 a moderate, and r = .30 a large effect. Due to the large sample size, p-values should be interpreted with caution and only in the context of effect sizes. In line with this, it was refrained from halving p-values for one-tailed hypotheses.

To further test the robustness of the main effects and complement the main analysis, possible interaction effects were investigated. For this, an empirical-inductive (i.e., an exploratory) approach was adopted and all possible interactions between the significant variables in the main analysis were investigated. Dummy-coded variables were coded −1 and 1, variables with no meaningful zero-point were mean-centered, and all possible interactions were generated. First, a reduced regression model, specified and tested as a path model, was computed, including the significant predictors from the main model specification and all respective interactions. In a next step, the significant interactions from the first step were added to the main model specification to ensure the specificity of the interaction effects. Interaction plots were derived from this extended model using the interaction utility by Preacher, Curran, and Bauer (2006).

3.1 Descriptive Statistics

Frequencies of all dichotomous variables can be found in Table 1. Descriptive statistics for all metric variables are shown in Table 2. On average, participants rather agreed with a UBI (M = 3.34, SD = 1.38, CI _LL = 3.28, CI _UL = 3.41). Frequencies of participants’ UBI agreement rating can be found in Figure 2.

Figure 2:

Frequencies of participants’ UBI agreement rating. UBI agreement was measured with the item “I am convinced that everyone should receive a universal basic income (even people who are very well off financially).” on a scale from 1 = not correct at all to 5 = fully correct.

3.2 Model Specification: Hierarchical Regression

The correlations for all variables can be found in Table 3. The results of the hierarchical multiple regression are shown in Table 4. The models were evaluated using the amount of explained variance (R ²) and the increase in R ² across models (ΔR ²). As can be seen in Table 4, the R ² improves with each step, from .01 in step 1 to .07 in step 4 (ΔR ² = .06). Thus, the complete model containing all 15 predictors provides the best model information and can be considered the model of choice yielding a small to moderate amount of explained variance. The comparison of results using weighted and unweighted data did not reveal any substantial differences (see Table S3 in the Supplemental Material).

3.3 Main Analysis: Hypotheses

The basis for answering the hypotheses is the complete model of the hierarchical multiple linear regression (step 4 in Table 4). As can be seen here, Age (β = −.06, z = −2.34, p = .019, r _part = −.05), Regional origin (β = .10, z = 4.36, p < .001, r _part = .10), Perceived economic situation (β = −.13, z = −4.70, p < .001, r _part = −.09), Corona perception (β = .07, z = 3.04, p = .002, r _part = .07), Equality orientation (β = .16, z = 7.37, p < .001, r _part = .16), and Leisure satisfaction (β = .05, z = 2.12, p = .034, r _part = .04) had significantly unique relations with UBI agreement with small effect sizes except for Equality orientation, where the effect size was moderate.

Thus, H2 (Age), H5 (Perceived economic situation), H6 (Corona perception), and H7 (Equality orientation) could be confirmed. Hence, being comparably younger, perceiving one’s current economic situation to be worse and the threat of the corona-pandemic to be higher, as well as being more in favor of equal living standards were uniquely associated with higher UBI agreement. Additionally, being of East-German origin was also uniquely associated with higher UBI agreement. However, H1 (Gender, β = .01, z = 0.46, p = .648, r _part = .01), H3 (Education, β = .04, z = 1.74, p = .081, r _part = .04), H4 (Income, β = −.03, z = −1.14, p = .255, r _part = −.02), H8 (Performance orientation, β = −.03, z = −1.24, p = .217, r _part = −.03), and H9 (Job satisfaction, β = .04, z = 1.39, p = .166, r _part = .02) had to be rejected. Furthermore, there was no significant unique relationship between UBI agreement and Care duty (children), (β = −.01, z = −0.45, p = .653, r _part = −.01), Unemployment frequency (β = .04, z = 1.54, p = .125, r _part = .03), Employment status (β = −.02, z = −0.95, p = .344, r _part = −.02), and Loneliness (β > −.01, z = −0.10, p = .917, r _part < .01).

The final model explains about 7 % of variance in UBI agreement. To identify the contribution of each variable to R ², relative weights were computed. The largest contributions came from Equality orientation (39 %), Perceived economic situation (19 %), and Regional origin (14 %).

3.4 Further Analyses: Interactions

To explore interaction effects, an empirical-inductive approach was used. First, all possible interactions were generated based on the main effects (i.e., fifteen interactions for six main effects). Second, to identify significant interactions, a reduced regression model including the significant predictors from the complete model (i.e., the six main effects) and all respective interactions (i.e., fifteen interactions) was computed. Third, a second regression analysis was calculated, adding the significant interactions to the complete model (i.e., all 15 predictors and all significant interactions). The first analysis revealed three significant interactions between Perceived economic situation and Equality orientation (β = .06, z = 2.66, p = .008, r _part = .06), Equality orientation and Leisure Satisfaction (β = −.05, z = −2.13, p = .034, r _part = −.05), and Corona perception and Leisure satisfaction (β = .05, z = 2.01, p = .044, r _part = .04). Adding these three interactions to the complete model in the second analysis yielded significant coefficients with small effect sizes for the interactions between Perceived economic situation and Equality orientation (β = .07, z = 2.86, p = .004, r _part = .06) and Equality orientation and Leisure satisfaction (β = −.05, z = −2.07, p = .038, r _part = −.05) (see Table S4 in the Supplemental Material for the regression results). See Figures 3 and 4 for the corresponding interaction plots. As shown in Figure 3, for people with higher Equality orientation, the perception of their economic situation had a smaller relation with UBI agreement than for people with lower Equality orientation. Moreover, people with higher Equality orientation had a higher UBI agreement than people with lower Equality orientation regardless of their perceived economic situation and vice versa. As shown in Figure 4, while for people with higher Equality orientation there is almost no relation between UBI agreement and Leisure satisfaction, there is a strong positive relation between Leisure satisfaction and UBI agreement for people with lower Equality orientation. Again, Figure 4 also shows that people with higher Equality orientation had a higher UBI agreement than people with lower Equality orientation regardless of their Leisure satisfaction and vice versa.

Figure 3:

Interaction plot of Equality orientation and Perceived economic situation.

Figure 4:

Interaction plot of Leisure satisfaction and Equality orientation.

4.1 Summary of Results

To identify relevant socio-demographic and psychological predictors of UBI acceptance among the German working population, we analyzed representative data that were collected during the first peak of the COVID-19 pandemic between May and June 2020 in a comprehensive multivariate analysis. First of all, we found that, on average, members of the German working population supported the idea of a UBI. Further, we found that being more supportive of a UBI went along with being comparably younger (H2), of East-German origin, and more in favor of equal living standards (H7), as well as perceiving one’s economic situation to be worse (H5) and the threat of the corona-pandemic to be higher (H6). Moreover, for participants being more in favor of equal living standards, their perceived economic situation as well as their leisure satisfaction had a lower influence on UBI agreement and vice versa. Yet, we found no significant association between UBI agreement and Gender (H1), Education (H3), Income (H4), Performance orientation (H8), and Job satisfaction (H9).

With our research, we add to the literature on individual characteristics associated with UBI support by providing a perspective on the German working population using current data from the onset of the COVID-19 pandemic, and most importantly, by identifying not only relevant socio-demographic but also psychological predictors of UBI acceptance in a comprehensive multivariate analysis.

4.2 Discussion of Selected Results

4.3 Limitations and Implications

Our research is intended to provide insights into and guidance about socio-demographic as well as psychological variables relevant for UBI acceptance and thus lay a foundation for future research on this topic. First of all, our research intends to draw attention to the fact that there is an individual dimension to the UBI debate that needs to be considered. This dimension is unrepresented in existing UBI research and a psychological perspective is lacking. However, the consideration of psychological variables, both, at criterion and predictor level, is imminent when investigating individual characteristics associated with UBI support.

Second, we investigated a set of psychological and socio-demographic predictor variables. Of course, this set of variables is not exhaustive, particularly because the theory-based variable selection was limited to the variables available in the multi-topic survey that our data was based on. The amount of variance explained by these variables was moderate at best. Thus, there must be other variables, particularly other psychological variables, that might be relevant and should be investigated in future research. We made a conscious decision to include variables from different disciplines to examine their relevance in explaining UBI acceptance not only in isolation but also in relative comparison and would recommend this approach for future research. Besides predictor level, the amount of variance explained might be also partly due to the potentially constrained reliability and validity of the one-item outcome variable. There are no established UBI acceptance measures to date, so we developed a single-item measure for this purpose. As a UBI is still hypothetical, the development of such a measure necessarily depends on assumptions about the definition of a UBI and its design. Of course, these assumptions and the resulting measure may influence participants’ UBI acceptance and corresponding results must always be evaluated in light of the underlying UBI definition. In fact, research suggests that (U)BI support is fragile and highly dependent on the underlying basic income model (e.g., Busemeyer, Rinscheid, and Schupp 2023; Chrisp, Pulkka, and García 2020; Laenen, Van Hootegem, and Rossetti 2023). Thus, for future research examining individual characteristics associated with UBI acceptance, it may be interesting to investigate predictors of acceptance for different (U)BI models (see, for example, Chrisp, Pulkka, and García 2020, who examined socio-demographic predictors of support for different basic income models in the United Kingdom and Finland or Laenen, Van Hootegem, and Rossetti 2023, who examined support for a broad range of different basic income models in Belgium). Although there are some studies that have done this, to our knowledge this has not yet been done for Germany, and the predictors examined in the existing studies are again limited to socio-demographic predictors.

Third, we followed a “screening-approach” to identify variables that could possibly be relevant for UBI acceptance. For this, we had to use economic feasible measures, i.e., single item measures. Fortunately, the controversy on the use of single item measures, which has been around for decades, is easing, so that the use of single item measures can be considered an established procedure in psychology by now (see, e.g., Fisher, Matthews, and Gibbons 2016, or Allen, Iliescu, and Greiff 2022; Matthews, Pineault, and Hong 2022, for very recent perspectives on the topic). Thalmayer, Saucier, and Eigenhuis (2011) and Ziegler, Kemper, and Kruyen (2014) furthermore show that the criterion validity is not affected by the use of shortened scales. Regarding construct validity, the coverage of single item measures is of course smaller than it is when using scales. Nevertheless, in line with the literature, we are confident that single items can cover the core aspect of a construct and are therefore sufficient measures. Of course, in the context of further research, a reasonable next step is the investigation of a little number of constructs that have been identified as relevant (by us and others) in their full complexity, including facets, via well-established scales.

Fourth, we used cross-sectional data from Germany that was collected just after the first peak of the COVID-19 pandemic. There are three limitations to that. First, it is likely that the specific time-point had an impact on our results, but it remains unclear what the exact nature of this impact was. Second, due to the cross-sectional nature of the data, we did not gain any insights about dynamics and causal explanations are not feasible. Third, we adopted a German perspective and therefore cannot provide insights on cross-country comparisons. Our research aim was to investigate UBI acceptance and relevant correlates among the German working population during the COVID-19 pandemic. However, for future research with other research foci, it may be interesting to replicate our results with post-pandemic data, conduct longitudinal studies to investigate dynamics, and add a cross-country perspective to the country-specific investigation.

Last, to better understand why people with certain characteristics agree or disagree with a UBI, and thus to corroborate the quantitative results, we also recommend the use of qualitative research methods for future research. Although some qualitative studies have been conducted to examine individual attitudes towards a UBI or towards specific features of a UBI, for example, by Herke and Vicsek (2022) for Hungary, Rossetti and colleagues (2020) for the Dutch Tilburg, or Zimmermann and colleagues (2020) for Slovenia and Germany, to our knowledge there are no qualitative approaches focusing on individual characteristics associated with UBI acceptance. However, this qualitative research is central to bridge the gap between the scientific groundwork done in studies like ours and potential practical implications of the resulting research findings. As a UBI has not been implemented in Germany, decision makers would have to specify design principles to develop a concrete policy. For this, knowledge of socio-demographic and psychological variables corresponding with UBI acceptance can be useful when combined with the understanding of why people with certain characteristics oppose or favor a UBI. For instance, if there is evidence that UBI acceptance tends to be lower in older age groups, and the reasons behind this are assessed in further research, this knowledge can be used to specifically address the concerns of older age groups when designing and communicating a UBI.

4.4 Conclusion

Even though the UBI gained popularity in times of the COVID-19 pandemic, its relevance goes far beyond that. Technological unemployment calls for alternative welfare systems that provide people with social safety in times of economic insecurity. The UBI can be one important instrument to address this. However, there is a broader need for psychological research on UBI that complements the social and economic viewpoints often focused on in current debates. The insights about variables relevant for UBI acceptance and a stronger psychological focus, thus, constitute a first critical step on a long-term research agenda.