Evidence-based social science: why, what, and future implications

The paper addresses longstanding doubts about the effectiveness of social science practice relative to natural sciences, highlighted by funding withdrawals, policy critiques, and institutional shifts in several countries. It asks why social science struggles to achieve reliable, accumulative practice outcomes and how to remedy the research–practice gap that leaves practitioners reliant on personal, tacit "practical wisdom." The author situates the problem within a broader classification of knowledge—distinguishing knowledge-of-knowing-the-world from knowledge-of-transforming-the-world—and argues that social science has insufficiently developed the latter compared to natural sciences and engineering. The purpose is to analyze root causes of low practical effectiveness and to propose evidence-based social science as a framework to systematize, accumulate, and transmit effective practice knowledge, thereby enabling "social engineering" analogous to natural engineering.

The paper reviews the evolution from evidence-based medicine (EBM) to evidence-based practice across domains. In response to rising healthcare costs and uneven outcomes, U.S. reforms (HMO Act, 1973; DRGs, 1983) catalyzed managed care and the EBM movement. Sackett’s definitions (1996, 2000) positioned EBM as integrating clinical expertise with the best external evidence via systematic research, supported by global databases (Cochrane, Campbell). The evidence-based movement spread within medicine (e.g., cardiology, nursing) and outward to psychology, social work, education, management, and policy, culminating in evidence-based social sciences (EBSS). The review also contrasts traditional reliance on personal practical wisdom with the EBM/EBSS model that emphasizes research evidence, guidelines, valuation of client/patient preferences, and contextual constraints. It summarizes the recognized hierarchy of evidence (systematic reviews/meta-analyses and randomized controlled trials at the top; quasi-experimental, single-case designs, correlational studies; down to expert opinion and professional consensus).

This is a theoretical and conceptual analysis. The author: (1) develops a knowledge classification framework distinguishing knowing vs transforming the world and positions natural science, social science, natural technology/engineering, and social technology/engineering within it; (2) analyzes the limitations of "practical wisdom" (situated, tacit, intuitive, personal) as a basis for practice and argues for its systematization; (3) synthesizes literature on EBM and the evidence-based practice movement to propose EBSS as the operationalization of "social engineering"; (4) articulates a five-step EBSS implementation model—asking, accessing, appraising, translating, and evaluating—linking researchers, practitioners, service recipients, and managers; (5) illustrates with an example from evidence-based reading instruction (National Reading Panel; What Works Clearinghouse guidelines); and (6) identifies present challenges and outlines future development strategies (evidence generation, localization, institutions, and training). No empirical data collection or statistical analysis is conducted; arguments are supported by references and exemplars.

- Social science’s practice ineffectiveness stems from a deficit in accumulative, transferable "knowledge-of-transforming-the-world" for people and society; reliance on individualized practical wisdom hinders progress. - Evidence-based social science (EBSS) functions as the "social engineering" of practical wisdom, translating research evidence into standardized, supervised, and improvable practice. - EBSS integrates four stakeholders—researchers, practitioners, service objects (clients), and managers—through a five-step process: (1) formulate specific questions; (2) access best available evidence; (3) appraise evidence quality; (4) translate/use evidence; (5) evaluate effectiveness, safety, and applicability. - EBSS embodies five unities: freedom and nature; truth-seeking and goodness-pursuing; individual and collective wisdom; autonomy and supervision; the general and the special. Meta-analyses/systematic reviews derive generalizable guidance while allowing contextual adaptation. - Evidence hierarchy prioritizes systematic reviews and randomized controlled trials, followed by multi-site and individual RCTs, quasi-experimental and single-case designs, correlational studies, narrative case studies, expert opinion, credible theory, and professional consensus. - Illustrative case: evidence-based reading instruction (e.g., National Reading Panel synthesis and What Works Clearinghouse guidelines) identifies core components (phonemic awareness, phonics, fluency, vocabulary, comprehension) and practical recommendations that, when applied, reduce reading difficulties for most children. - Key current challenges: scarcity of high-quality original studies suitable for meta-analysis and guideline development; limited understanding and adoption of EBSS among practitioners and the public; weak institutional support; evidence predominantly Western and not yet localized for diverse cultures; gaps in standards for evidence translation and fit in practice.

By reframing practice in social sciences as evidence-based "social engineering," the paper bridges the research–practice divide that historically left practitioners dependent on tacit, non-transferable wisdom. The proposed EBSS framework operationalizes how to transform people and society effectively: it systematizes procedural knowledge, externalizes and standardizes practice steps, and embeds supervision and evaluation. This addresses criticisms about low predictability and inconsistent guidance from social science by linking truth-seeking (research) to goodness-pursuing (effective, equitable services). The five unities clarify how EBSS respects human agency and context while adhering to empirically established regularities, enabling both generalizability (via meta-analytic synthesis) and local adaptation. The example of reading instruction illustrates how large-scale evidence syntheses can guide special practices and improve outcomes. The broader relevance is a pathway for social sciences to achieve cumulative, transparent, and equitable practical impact analogous to natural engineering in the physical world.

The paper advances evidence-based social science as the means to convert practical wisdom into cumulative, transmissible, and supervised practice—"social engineering." It calls for removing randomness in practice, implementing standardization, lowering costs, and improving efficiency via rigorous evidence grading, guidelines, manuals, and open databases. Stakeholders should collaborate to expand high-quality, culturally localized evidence; establish institutions to coordinate and fund EBSS; and train practitioners using standardized interventions and case-based instruction. With widespread adoption, the author envisions an "evidence-based society" where evidence-guided decision-making permeates roles such as teachers, clinicians, social workers, managers, police, and judges, and where individuals both consume and create practice-relevant evidence. Future research should prioritize generating rigorous, context-sensitive studies; conducting meta-analyses; refining evidence taxonomies; improving evidence translation and fit in practice; and building global yet localized evidence infrastructures.

The study is conceptual and does not present original empirical data or formal evaluations of EBSS implementations, limiting empirical validation of claims. Examples (e.g., reading instruction) are illustrative rather than tested within this paper. The current evidence base in many social domains is acknowledged to be sparse, uneven in quality, and culturally skewed toward Western contexts, constraining generalizability. Adoption barriers—limited practitioner awareness, motivation, and institutional support—may impede implementation and bias assessments of effectiveness. Consequently, proposed frameworks and benefits require empirical testing across diverse settings and populations.