"Effective Policies" Alone Won't Reduce Inequality: The Need for a Perspective on "Reach" to Reduce Social Inequality

When considering policies and support measures to eliminate social inequality, we tend to focus solely on whether the measure is effective. Does academic counseling encourage university enrollment among low-income youth? Are women and immigrants disadvantaged in recruitment processes? Do community activities and civic participation foster trust and mutual aid among people from different backgrounds?

To answer such questions, experimental methods have become increasingly important in recent sociology. Experiments are suited to investigating whether a certain measure or condition has a causal impact on outcomes. Randomized controlled trials, hiring experiments using fictitious resumes, and trust games have been used as powerful methods to measure invisible discrimination and the effects of support within society.

However, new research by Irena Pietrzyk and Marita Jacob from the University of Cologne points out a significant blind spot here. Even if an experiment shows that a measure is "effective," it doesn't necessarily mean that inequality across society will be reduced. This is because, in the real world, not everyone is exposed to the measure or condition in the same way.

The concept that researchers emphasize is "treatment prevalence," or the proportion of people who actually receive a certain measure or condition. In Japanese, this can be translated as "prevalence rate of treatment," "contact rate with the measure," or "reach rate of support." In short, no matter how effective a support measure is, if it doesn't reach the people who need it, it won't lead to a reduction in inequality.

For example, suppose an academic counseling program increases the university enrollment rate of students from low-income families. The experiment confirmed a clear effect. At first glance, it seems that expanding the policy could reduce educational inequality. However, what if the counseling is primarily received by students from families already well-endowed with information and resources? Despite the support measure being "effective," it might not reduce inequality but rather further bolster already advantaged groups.

Conversely, even if the effect itself is not large, if the support reaches those who truly need it, it can have the power to reduce inequality across society. In other words, to reduce social inequality, one must not separate the "effectiveness of the measure" from the "reach of the measure." The core question of inequality countermeasures is who receives effective support, to what extent, and under what conditions.

The research introduces three case studies.

The first is an experiment on trust games conducted in Italy. This experiment did not show a clear tendency for people with immigrant backgrounds to be less trusted than those without. Superficially, it seems that ethnic background does not significantly disadvantage individuals. However, in reality, opportunities to participate in civic groups, volunteer activities, and community organizations may differ among groups. If people with immigrant backgrounds find it difficult to participate in such venues, they will have fewer opportunities to benefit from the mutual aid and trust that arise there. Even if discrimination is not visible in experiments, if the structure of social participation differs, inequality remains as a result.

The second is an experiment on hiring for professorships conducted in Germany and Italy. A survey using fictitious applicant profiles did not necessarily show that female candidates were rated lower than male candidates. At first glance, there seems to be no clear gender discrimination in the evaluation itself. However, in academia, men are more likely to become lead authors of papers, which may lead to differences in how achievements are perceived before applying. Even if men and women are treated equally in the evaluation stage, if the prerequisites for that evaluation differ, the structure that makes it easier for men to secure professorships is maintained.

The third is a large-scale field study on academic counseling conducted in North Rhine-Westphalia. The study found that intensive guidance counseling had the effect of promoting higher education enrollment among students from disadvantaged families. This is a hopeful result. However, if this program is actually used more by students from advantaged backgrounds, the effect of reducing inequality weakens. In some cases, the support measure could further expand the options for advantaged groups.

The important point here is not that "experiments are wrong." Rather, experiments are very effective in investigating whether a certain measure has a causal effect on individuals. The problem lies in directly linking these results to interpretations of inequality across society. What is effective at the individual level is not the same as reducing inequality at the group level.

This is easy to understand when compared to medicine. Even if a drug is proven effective in clinical trials, if it doesn't reach the patients who need it, the health inequality across society won't shrink. If it is expensive, accessible only to a few, and only those with information can access it, the drug, while effective, might exacerbate inequality. The same can happen with policies on education, employment, welfare, and community participation.

This perspective offers significant insights for Japanese society as well. For instance, scholarship programs, reskilling support, job counseling, childcare support, and consultation services for those in financial distress may exist as systems, but they don't necessarily reach those who truly need them. Complex application processes, lack of information, lack of time to seek counseling, psychological barriers, and the absence of experienced users in one's surroundings—these small barriers can significantly affect the reach rate of support.

Particularly, those in socially disadvantaged positions may have less capacity to access support systems. Time, transportation costs, digital environment, language skills, document preparation ability, and trust in the system—these are invisible but essential conditions for making policy effects a reality. The people assumed to be the target of support measures are often the ones who find it hardest to access the support. This is a common issue across many systems.

In that sense, the questions raised by this research are extremely practical. Policymakers, educational institutions, NPOs, and corporate HR departments must look not only at "whether this measure is effective" but also at "who is using this measure," "who is not using it," and "why it is not reaching them."

Judging by reactions on social media, interest in this research is primarily spreading among the university, researcher, and expert communities. The University of Cologne has introduced the research content on LinkedIn and Facebook, conveying the message that "even if effective in experiments, who actually benefits in reality is crucial." Additionally, Marita Jacob, one of the authors, explained in a previous LinkedIn post that when inferring social inequality from experimental results, it is necessary to consider how commonly a specific treatment or support is received within a group.

On the other hand, there is limited evidence of widespread discussion in general social media spaces. The Phys.org article itself did not show a high number of comments or shares at the time of publication. This does not mean the research is unimportant. Rather, because the theme involves policy evaluation and the methodology of experimental sociology, the first to react are often researchers, university affiliates, and experts interested in social policy.

However, the message of this research should ideally reach a broader audience. This is because we encounter terms like "effective educational support," "proven employment support," and "scientifically validated policies" daily. Of course, effectiveness verification is important. However, the outcome a measure brings about in society as a whole is not determined solely by the magnitude of its effect. Who participates, who cannot participate, and who is left out of the benefits—without looking at these, one might misjudge the success of a policy.

The visualization tool developed in this research is also based on this awareness. By combining the effects of measures with the reach rate in each group, it can simulate how inequality changes. This is useful not only for researchers but also for practitioners. For example, before rolling out a program nationwide, one can anticipate which groups it will easily reach and which will be left behind. Alternatively, one can check how changing participation rate biases affects the inequality reduction effect. Such considerations are important for using limited budgets more equitably.

What should be most avoided in inequality countermeasures is the mindset of "we've created a good system, so that's enough." A system does not function merely by existing. Support does not reach its target simply by being designed. Only when those in need can access it with confidence, actually use it, and continue to benefit from it, does a policy have the power to change society.

This research brings a sober sense of reality to discussions on social inequality. Inequality does not arise solely from differences in individual abilities or efforts. It also stems from invisible structures such as how opportunities are connected, access to support, pre-evaluation conditions, and ease of entering social participation venues. Therefore, policies to reduce inequality must do more than broadly define the target audience. They must design the very pathways to ensure they reach those in truly disadvantaged positions.

From "is it effective" to "who is it reaching." The perspective shown by this research will become indispensable for future policy evaluation. To genuinely reduce social inequality, it is necessary to look not only at the content of the measures but also at how they flow through society, where they stop, and who they pass by.

What truly matters is not just putting up a sign for support measures. It is ensuring they indeed reach those who need them.

Source URL

Phys.org: An article introducing research from the University of Cologne. It highlights that in measures to reduce social inequality, it is important not only to consider effectiveness but also who actually benefits.
https://phys.org/news/2026-05-social-inequality-scope-crucial.html

University of Cologne Official News: University official announcement explaining the research content, author comments, three case studies, and visualization tools.
https://uni-koeln.de/en/university/news/news/news-detail/reducing-social-inequality-why-the-scope-of-measures-is-crucial

Original Paper Published in Springer: "Why Treatment Prevalence Matters: Overcoming a Blind Spot in Experimental Inequality Research" by Irena Pietrzyk and Marita Jacob.
https://link.springer.com/article/10.1007/s11577-026-01068-7

Related Research Published in SAGE: A randomized controlled trial examining the effect of academic counseling on reducing university enrollment inequality in Germany.
https://journals.sagepub.com/doi/10.1177/00380407251323888

University of Cologne LinkedIn Post: Official university social media post about the introduction of this research.
https://www.linkedin.com/posts/university-of-cologne_unik%C3%B6ln-soziologie-ungleichheit-activity-7457001426364592128-ieqE

Marita Jacob's LinkedIn Post: A post explaining the need to consider treatment prevalence when inferring social inequality from experimental research.
https://www.linkedin.com/posts/marita-jacob-ba0861281_analyticalsociology-socialinequality-causalanalysis-activity-7399487836430815232-EXy7