Gene Glass – 1940 Present

Biography

The trajectory of twentieth-century behavioral science, psychometrics, and educational evaluation cannot be understood without reckoning with the intellectual legacy of Gene V. Glass (born June 19, 1940). Operating at the vital intersection of advanced mathematical statistics, empirical social research, and critical policy analysis, Glass fundamentally transformed how scientific fields aggregate, interpret, and mobilize empirical evidence. Prior to his pioneering conceptual and methodological interventions in the mid-1970s, social scientific research synthesis was predominantly an informal, subjective, and literary exercise. Scholars confronted by dozens or hundreds of conflicting empirical studies routinely relied on selective narrative reviews, idiosyncratic vote-counting procedures, or ideological heuristics to declare a given intervention effective or inert. Glass disrupted this epistemological inertia by demonstrating that research findings themselves constitute a rich domain of quantitative data susceptible to rigorous, systematic, and empirical analysis.

Most renowned as the originator of the term and foundational methodology of meta-analysis—a concept he formally introduced during his 1976 presidential address to the American Educational Research Association—Glass provided the scientific community with a reproducible metric for harmonizing disparate research scales: Glass’s Delta. Far from remaining an insular innovation within psychometrics, this quantitative synthesis framework diffused rapidly into clinical psychology, psychiatric epidemiology, pharmacology, education, and ultimately modern evidence-based medicine. Glass utilized his newfound methodology to resolve long-standing deadlocks in social policy and psychotherapy outcome research, demonstrating empirically that psychological therapies exerted substantial positive effects and that class size reductions produced marked, non-linear improvements in student scholastic achievement. In doing so, he showed that quantitative techniques could dismantle longstanding clinical and institutional orthodoxies that had disenfranchised vulnerable populations.

Yet, to characterize Gene Glass purely as an abstract statistician or quantitative methodologist is to miss the radical ethical core of his intellectual vocation. Throughout a career spanning professorships at the University of Colorado Boulder and Arizona State University, Glass consistently mobilized rigorous statistical inquiry in defense of public education and democratic egalitarianism. He authored incisive critiques of the arbitrary cut-scores embedded within minimum-competency standardized testing, exposed the statistical instability and methodological fraudulence of value-added teacher evaluation models, and founded the Education Policy Analysis Archives (EPAA) in 1993 as one of the world’s first open-access digital peer-reviewed scholarly journals. Glass’s career represents a profound synthesis of statistical sophistication, methodological rebellion, and unabashed public advocacy against the neoliberal commercialization of public education.

1. Early Life, Educational Background, and Formative Academic Influences (1940–1965)

1.1 Childhood and Formative Years in the American Midwest

Gene V. Glass was born on June 19, 1940, in Lincoln, Nebraska, arriving into an American Midwest that was negotiating the immediate aftermath of the Great Depression and the mobilization for World War II. The regional socio-economic fabric of Nebraska during the 1940s and early 1950s was characterized by a pragmatic, agrarian ethos interwoven with the burgeoning industrial optimism of the post-war era. Growing up within this midwestern landscape, Glass was exposed early on to the values of practical utility, communal responsibility, and direct inquiry—cultural dispositions that would later infuse his empirical approach to social problems. His family life provided an environment wherein discipline and academic curiosity were encouraged, positioning education not merely as a route toward upward social mobility, but as an essential civic engagement.

Throughout his early schooling in the Lincoln public school system, Glass exhibited an exceptional aptitude for mathematics, formal logic, and linguistic structures. Rather than viewing mathematical abstraction as an isolated puzzle, he displayed an instinctive inclination toward understanding how quantitative structures corresponded to observable phenomena in the physical and social worlds. The expansion of post-war public infrastructure in the late 1940s and 1950s—fueled by federal investments in scientific education in the wake of technological competition—fostered a cultural climate that valorized scientific and mathematical mastery. Glass capitalized on this environment, mastering advanced secondary-school algebra, calculus, and natural sciences with an innate ease that presaged his later quantitative career.

Upon completing his secondary education, Glass matriculated at the University of Nebraska-Lincoln, where he pursued an undergraduate program centered around mathematics, physical science, and the humanities, particularly German. This dual focus on mathematical precision and humanities-based hermeneutics provided him with a unique epistemological posture; he became acutely aware of how language and conceptual definitions frame quantitative inquiries. His undergraduate coursework in mathematics deepened his appreciation for probability theory, matrix algebra, and formal proof, while his encounters with mid-century sociological and psychological texts revealed a conspicuous gap: behavioral and social disciplines frequently collected vast quantities of observational data yet lacked the rigorous, unified statistical architectures that characterized the physical sciences. This realization became the initial catalyst for his decision to redirect his analytical talents toward educational psychology and quantitative methodology.

1.2 Doctoral Training at the University of Wisconsin-Madison

Following the completion of his bachelor’s degree at the University of Nebraska, Glass gained admission to the doctoral program in educational psychology at the University of Wisconsin-Madison, arriving on campus in the autumn of 1962. At that time, Wisconsin was universally recognized as an international powerhouse for behavioral measurement, psychometrics, and quantitative experimental design. The department was animated by a cadre of preeminent scholars, most notably Julian C. Stanley, a brilliant and exacting statistician who had studied under the legendary experimental design pioneer Helen M. Walker. Stanley, who would become Glass’s primary doctoral mentor, was dedicated to instilling uncompromising methodological rigor into educational experimentation, co-authoring what would become canonical works on experimental and quasi-experimental designs with Donald T. Campbell.

Immersed in this intellectually demanding environment, Glass underwent exhaustive training in modern psychometric theory, advanced analysis of variance (ANOVA), multivariate statistics, and statistical inference. He was deeply influenced by the conceptual tensions between classical test theory, factor analysis, and the emerging paradigms of criterion-referenced measurement championed by figures such as Chester W. Harris. The mid-1960s at Wisconsin was a watershed moment where the assumptions underlying classical null-hypothesis significance testing (NHST) were being vigorously scrutinized. Glass distinguished himself among his peers not only by mastering the complex mathematical derivations of general linear models, but by probing the limits, robustness, and mathematical failure modes of these models when applied to messy, non-normal social science data.

Glass’s doctoral dissertation, completed in 1965 under Stanley’s supervision, was an intricate investigation into statistical distribution theory and inferential validity. Specifically, Glass explored the sampling distributions of various dependent metrics, testing how violations of core parametric assumptions—such as homogeneity of variance and distributional normality—distorted inferential error rates and degrees of freedom in analysis of variance models. His dissertation revealed a scholar preoccupied not with dogmatic mathematical formalism for its own sake, but with the practical resilience of statistical tools when confronted with real-world empirical violations. Graduating with his Ph.D. at the young age of twenty-five, Glass emerged from Madison possessing an extraordinary foundation in statistical mechanics and a burgeoning skepticism regarding unexamined methodological conventions.

1.3 Emergence of Epistemological Commitments in Social Science

The conclusion of his formal training at Wisconsin crystallized Glass’s overarching epistemological worldview, which can best be characterized as a fusion of empirical pragmatism, scientific realism, and radical methodological accountability. Influenced by the logical positivist traditions of the early twentieth century yet deeply attuned to the post-positivist critiques of Karl Popper and Donald T. Campbell, Glass maintained that social scientific claims to knowledge must be tethered to transparent, reproducible, and mathematically coherent evidence. He rejected the prevailing hermeneutic complacency that permitted educational scholars to issue authoritative policy prescriptions based on subjective clinical intuitions, uncontrolled observational anecdotes, or rhetorical elegance.

A central pillar of Glass’s emerging epistemological critique was aimed directly at the traditional, narrative literature review that dominated mid-century psychology and education. In these standard reviews, an individual scholar read a self-selected collection of published papers, applied unstated criteria of validity, and constructed a qualitative narrative asserting that a specific treatment was either effective or ineffective. Glass recognized that this methodology was fundamentally unscientific. It lacked quantitative systematization, harbored massive susceptibility to confirmation bias, treated small exploratory studies identically to massive randomized trials, and provided no objective mechanism to reconcile diametrically opposed empirical outcomes across independent laboratories. Narrative reviews, in Glass’s eyes, were little more than literary exercises masquerading as cumulative science.

Consequently, Glass began formulating a scientific philosophy premised on metric harmonization, comprehensive operational transparency, and the systematic democratization of data. He argued that the empirical literature of a scientific discipline represented an archival population of empirical measurements that had to be analyzed with the identical statistical rigor that researchers routinely applied to individual participants within a single laboratory experiment. This philosophical commitment required researchers to abandon the myth of the definitive, single-study breakthrough. Instead, scientific progress had to be understood as cumulative, distributed, and fundamentally metric-dependent. To realize this vision, the behavioral and social sciences required a new set of measurement tools capable of bridging disparate experimental scales, a quest that would preoccupy Glass over the ensuing decade.

2. Early Academic Career and Methodological Foundations at the University of Colorado

2.1 Faculty Appointment at the University of Colorado Boulder

In 1965, immediately following the completion of his doctorate, Gene Glass accepted an academic appointment at the School of Education at the University of Colorado Boulder. Within this setting, he quickly established himself within the Laboratory of Educational Research (LER), an innovative academic center designed to inject sophisticated quantitative methodologies into the evaluation of educational programs, psychometric assessments, and social reforms. The Boulder campus in the late 1960s was an invigorating, highly collaborative academic crossroads. Free from the deeply entrenched orthodoxies of older Ivy League institutions, Colorado provided the young assistant professor with the intellectual autonomy necessary to challenge established paradigms in educational psychology and measurement.

At Boulder, Glass assumed the responsibility of teaching advanced doctoral-level coursework in experimental design, multivariate statistical methods, psychometric theory, and program evaluation. Rather than relying solely on abstract mathematical proofs, Glass emphasized the practical application of general linear models, repeated-measures analysis of variance, and multiple regression to complex institutional datasets. His pedagogy was marked by a relentless insistence on questioning the underlying mathematical assumptions of common statistical software packages, which were then beginning to proliferate across academic mainframe computing centers. His classes quickly developed a reputation among graduate students as demanding, philosophically rich, and methodologically transformative.

During this formative period, Glass established enduring collaborative relationships with prominent social scientists, methodologists, and cognitive researchers at Boulder, including Kenneth R. Hammond, who was pioneering social judgment theory, and later, clinical researchers such as Mary Lee Smith. The intellectual synergy within the Laboratory of Educational Research catalyzed a sustained period of empirical and theoretical productivity. Surrounded by colleagues grappling with the challenges of naturalistic observation, curriculum development, and cognitive testing, Glass was pushed to extend his doctoral inquiries beyond pristine laboratory distributions into the turbulent, confounded realities of actual educational classrooms, public clinics, and community programs.

2.2 Early Publications in Psychometrics and Experimental Design

Between 1966 and 1974, Glass produced a sequence of influential peer-reviewed articles and monographs that solidified his standing as one of the premier young quantitative methodologists in the United States. A central focus of his early scholarship was the measurement of change—a notoriously treacherous domain within classical psychometrics. Educational researchers had long relied on crude “gain scores” (subtracting pre-test scores from post-test scores) to evaluate the success of instructional interventions. Glass published sharp mathematical expositions demonstrating how gain scores were plagued by unreliability, regression to the mean, and the mathematical artifact known as Lord’s Paradox, wherein the direction and magnitude of an intervention effect can arbitrarily invert depending on whether one utilizes analysis of covariance (ANCOVA) or repeated-measures ANOVA on gain scores.

Beyond his psychometric critiques, Glass contributed substantially to the literature on experimental design, particularly regarding the analysis of time-series experiments. In collaboration with colleagues such as Victor L. Willson and John M. Gottman, Glass explored how interrupted time-series designs could be rigorously analyzed using integrated moving average models (ARIMA), which had been popularized in econometrics by Box and Jenkins. By adapting these complex stochastic models to educational and behavioral data, Glass provided researchers with a powerful quasi-experimental tool to assess the impact of legislative policies, school-wide curricula, or behavioral interventions where classical randomized controlled trials were logistically or ethically impossible.

The culmination of his early methodological scholarship was the publication of what would become a seminal textbook in behavioral statistics: Statistical Methods in Education and Psychology, co-authored with his former mentor Julian C. Stanley in 1970. The text was widely praised for its exceptional clarity, pedagogical rigor, and integration of modern general linear model perspectives. Unlike traditional introductory manuals that presented statistical tests as an arbitrary cookbook of recipes (e.g., separate, disconnected procedures for t-tests, one-way ANOVA, and correlation), Glass and Stanley synthesized these disparate metrics within a unified mathematical architecture. The textbook was swiftly adopted in top-tier graduate programs across North America, establishing Glass as a central authority on educational measurement while still in his early thirties.

2.3 Involvement in Federal Educational Evaluation Programs

The mid-to-late 1960s coincided with the monumental expansion of the American welfare state under President Lyndon B. Johnson’s Great Society initiatives. Chief among these was the Elementary and Secondary Education Act (ESEA) of 1965, specifically its massive Title I provision, which allocated unprecedented federal resources toward compensating for the educational deficits of socioeconomically disadvantaged children. Crucially, the legislation mandated that local educational agencies evaluate the effectiveness of these federally funded programs using objective, quantitative measures. This mandate precipitated a financial and logistical explosion in the field of educational evaluation, drawing academic researchers like Glass directly into the assessment of large-scale social engineering programs, including Project Head Start and Project Follow Through.

Glass was frequently retained as a statistical consultant, methodologist, and evaluation auditor for these sprawling federal evaluation initiatives. As he examined the reams of evaluation reports crossing his desk, he was confronted by a deeply demoralizing reality. Despite millions of federal dollars expended and thousands of independent program evaluations conducted, the empirical literature was an unintelligible morass of contradictory findings. One school district would report that a diagnostic-prescriptive reading intervention dramatically improved reading scores, while a neighboring district evaluating the identical curriculum reported null or negative effects. The reports varied wildly in their sample sizes, outcome measures (ranging from idiosyncratic teacher ratings to standardized norm-referenced achievement tests), and statistical methodologies.

Traditional evaluation methodology proved wholly incapable of synthesizing this fragmented knowledge base. Federal policy makers, desperate for unambiguous answers regarding whether compensatory education “worked,” were treated to sprawling narrative summaries that selectively emphasized particular program models based on the ideological predispositions of the evaluator. Glass realized that the problem did not reside solely in the quality of the individual interventions, but in the catastrophic absence of an aggregate scientific methodology. The traditional paradigm possessed no common currency to translate an outcome measured on a local criterion-referenced test into the metric of an Iowa Test of Basic Skills, nor did it possess a method to weight the precision of massive studies against micro-pilots. This chronic epistemological failure across Title I evaluations laid the direct empirical groundwork for Glass’s conceptual breakthrough in 1976.

3. The Conceptual Genesis and Coining of Meta-Analysis (1976)

3.1 The 1976 AERA Presidential Address

In April 1976, Gene Glass delivered his presidential address to the American Educational Research Association at its annual meeting in San Francisco, California. At age thirty-five, Glass was one of the youngest individuals ever elected to lead this preeminent organization of educational scholars. Rather than delivering a conventional, platitudinous overview of institutional accomplishments, Glass used the prestigious platform to launch a frontal assault on the epistemological foundations of educational research synthesis. His paper, titled “Primary, Secondary, and Meta-Analysis of Research,” and subsequently published in the Educational Researcher, marked an irreversible historical inflection point in the philosophy and mechanics of scientific literature review.

In this address, Glass formally coined and defined the term meta-analysis. He defined it succinctly and elegantly as “the statistical analysis of a large collection of analysis results from individual studies for the purpose of integrating the findings.” To establish conceptual clarity, Glass delineated three distinct tiers of research analysis:

• Primary Analysis: The original, direct analysis of empirical data collected during a specific investigation (the traditional laboratory experiment or survey administration).
• Secondary Analysis: The re-analysis of primary data using better statistical techniques or to answer new questions directly from the original raw data files.
• Meta-Analysis: The analysis of analyses; an approach wherein the individual published study is treated as a single observation, its summary statistics are extracted as data points, and statistical models are applied to these extracted statistics to detect macro-level patterns, main effects, and structural moderators across the cumulative literature.

Glass argued that the contemporary scientific community was suffering from a pathology of hyper-specialization and information overload. Scientific journals were proliferating exponentially, each individual study presenting a narrow, probabilistic glimpse of reality through the lens of local sampling error. Because individual studies routinely suffered from low statistical power, minor variations in implementation, or local measurement idiosyncrasies, reliance on single studies or narrative summaries inevitably generated an illusion of pervasive scientific conflict. Glass proclaimed that the ultimate scientific enterprise was not the continuous, atomized generation of isolated experiments, but the systematic, quantitative aggregation of the global scientific archive.

3.2 Theoretical Framework of Quantitative Synthesis

The mathematical and theoretical architecture of meta-analysis proposed by Glass was predicated on two foundational pillars: metric standardization and comprehensive exhaustive aggregation. To overcome the long-standing obstacle that independent researchers measured identical underlying constructs using radically different scales—such as varying psychological batteries, disparate grade-point averages, or differing clinical rating matrices—Glass established a standardized, scale-free metric of treatment impact, which would become globally known as Glass’s Delta ($\Delta$).

Glass’s Delta was derived by calculating the raw difference between the mean of the experimental group ($\bar{X}_E$) and the mean of the control group ($\bar{X}_C$), and dividing that difference by the standard deviation of the control group ($S_C$):

$$\Delta = \frac{\bar{X}_E – \bar{X}_C}{S_C}$$

This formulation expressed the magnitude of an experimental effect in standard deviation units. An effect size of $\Delta = +1.00$ meant that the average individual in the experimental condition scored one full standard deviation above the average individual in the untreated control group. By converting disparate test scores, behavioral counts, and clinical assessments into this universal metric of standard deviation units, Glass provided social science with an objective, scale-free currency. Researchers could now directly aggregate, compare, and mathematically average findings from studies conducted decades apart across entirely different laboratory settings.

Crucially, Glass’s theoretical framework demanded an absolute rejection of subjective narrative reviews. In traditional reviews, scholars routinely employed the “vote-counting” method, categorizing studies into positive, negative, or non-significant based purely on statistical significance tests ($p < .05$). Glass demonstrated the catastrophic mathematical flaws of vote-counting: a series of small, underpowered studies with true positive effects might all fail to reach nominal statistical significance due to small sample sizes ($N$), leading a vote-counting reviewer to falsely conclude that an intervention had no effect, whereas a quantitative aggregation of their effect sizes would reveal overwhelming, incontrovertible evidence of a powerful, statistically significant impact. Meta-analysis converted the literature review from an idiosyncratic artistic appraisal into an empirical, reproducible branch of mathematical science.

3.3 Early Intellectual Reception and Methodological Controversies

The introduction of meta-analysis was met with immediate, ferocious resistance from prominent traditional quantitative researchers, psychometricians, and experimental purists. Chief among the early critics was the renowned British-German psychologist Hans Eysenck, who published a scathing critique famously dismissing Glass’s meta-analytic work as “an exercise in mega-silliness.” Eysenck and his intellectual allies argued that applying advanced statistics to aggregates of published literature violated the fundamental canon of experimental science: the primacy of methodological quality. They formulated the enduring “Garbage In, Garbage Out” (GIGO) critique, asserting that averaging hundreds of poorly designed, uncontrolled, or methodologically weak studies alongside pristine, randomized experiments served only to produce an enormous, highly precise mountain of statistical garbage.

A second major counter-assault was termed the “Apples and Oranges” problem. Critics argued that aggregating studies that utilized different operational definitions of treatments, different participant populations, and different measurement instruments was fundamentally illegitimate—akin to mathematically averaging the weight of apples and oranges to describe a nonexistent universal fruit. How, critics asked, could a scholar defensibly average an effect size derived from an eight-week psychodynamic therapy protocol administered to hospitalized neurotic adults with an effect size derived from a two-week behavioral desensitization program administered to snake-phobic undergraduate students? To traditionalists, the resulting composite effect size was a meaningless abstraction that obscured the essential idiosyncrasies of specific clinical and educational practices.

Glass met these fierce criticisms with unwavering empirical rejoinders that fundamentally redefined the philosophy of scientific synthesis. In response to the “Garbage In, Garbage Out” objection, Glass asserted that a priori exclusions of studies based on a reviewer’s subjective perception of “methodological quality” introduced lethal confirmation bias into reviews. Reviewers routinely used “quality criteria” as a pretext to discard empirical studies whose findings contradicted their theoretical orthodoxies. Glass argued that methodological quality should not be an arbitrary exclusionary gate; rather, study quality should be empirically coded as a study-level moderator variable. Researchers could then run statistical models to test whether poorly designed studies systematically produced larger or smaller effect sizes than rigorous randomized trials. If the data showed no statistical difference between high- and low-quality designs, the exclusion of the latter was unjustified. Regarding the “Apples and Oranges” critique, Glass famously quipped that synthesizing apples and oranges is entirely appropriate if one is interested in studying the broader class of “fruit.” Quantitative synthesis did not seek to eliminate diversity, but to model diversity systematically through empirical moderator analyses.

4. Landmark Meta-Analytic Investigations: Psychotherapy Outcome Research

4.1 The Eysenck Debate and the Efficacy of Psychotherapy

To demonstrate the transformative empirical power of his new methodology, Gene Glass chose to intervene in the most intractable, ideologically hostile controversy in post-war clinical psychology: the efficacy of psychotherapy. In 1952, Hans Eysenck had dropped an intellectual bomb on the psychological establishment by publishing a paper asserting that psychotherapy had no demonstrably beneficial effect on neurotic patients. Drawing upon a selective collection of insurance records, institutional discharge rates, and uncontrolled clinical reports, Eysenck claimed that roughly two-thirds (66%) of neurotic patients recovered spontaneously within two years without any formal psychological intervention—a phenomenon he termed “spontaneous remission.” Because the recovery rates reported in the psychotherapy literature hover around the same two-thirds mark, Eysenck concluded that clinical psychology and psychoanalysis were clinically useless, cost-ineffective placebo rituals.

Eysenck’s provocative thesis precipitated a quarter-century stalemate characterized by vicious, inconclusive academic warfare. Psychoanalysts, humanistic counselors, and behavioral therapists published hundreds of rejoinders, empirical trials, and narrative literature reviews. However, these narrative summaries were universally fractured along theoretical allegiances. Freudian analysts published reviews claiming psychodynamic methods were uniquely profound; behavioral therapists published reviews claiming behavioral conditioning was uniquely scientific; and skeptical experimental psychologists continued to echo Eysenck’s claim that psychological interventions were fundamentally unproven. The field was paralyzed because each camp possessed an extensive arsenal of individual studies to cite in support of its preferred narrative, while systematically dismissing opposing empirical evidence as methodologically flawed.

Recognizing that narrative methodology could never break this structural paralysis, Glass teamed up with his colleague and doctoral student at the University of Colorado, Mary Lee Smith. Together, they resolved to execute the world’s first comprehensive, exhaustive meta-analysis of the global psychotherapy outcome literature. Rather than picking a side in the ideological war between behaviorists and psychoanalysts, Glass and Smith treated the vast, sprawling archive of clinical trial literature as a unified dataset waiting for systematic empirical excavation. Their objective was to ascertain, once and for all, whether empirical evidence supported the claim that individuals receiving formal psychotherapy exhibited systematically superior clinical outcomes compared to untreated control groups.

4.2 Methodological Execution of the Smith and Glass (1977) Meta-Analysis

The execution of the Smith and Glass meta-analysis was an unprecedented feat of empirical literature extraction and computational labor. Working without modern digital databases, the internet, or automated citation aggregators, Smith and Glass manually searched library shelves, psychological abstracts, dissertations, and medical journals. They established rigorous, objective inclusion criteria: a study had to evaluate at least one psychological therapy condition against an untreated control group (such as a waiting-list control or attention-placebo control) and report sufficient statistical information—means, standard deviations, t-statistics, or F-ratios—to permit the calculation of an effect size metric.

Their final analytical sample comprised nearly 400 controlled evaluations of psychotherapy (specifically, 375 studies yielding 833 distinct effect sizes). For every single study, Glass and Smith meticulously extracted study-level characteristics: the theoretical orientation of the therapy (psychodynamic, Adlerian, client-centered, rational-emotive, systematic desensitization, behavior modification), the duration and dosage of treatment, the clinical diagnosis of the patient population, the setting of the intervention, the nature of the outcome measures, and whether the primary investigator had developed the outcome assessment tool. Every study outcome was then mathematically converted into Glass’s Delta effect size metric.

The results, published in the September 1977 issue of the American Psychologist, sent shockwaves through the scientific world. Across all 833 comparisons, the mean effect size of psychotherapy was $\Delta = +0.68$ standard deviations (which subsequent expanded analyses refined upward to 0.85). In plain statistical terms, this demonstrated that the average client who received psychotherapy scored nearly seven-tenths of a standard deviation better on psychological health, symptom reduction, and social functioning than the average individual in the untreated control group. Translated into percentiles, the findings demonstrated that the average treated client was functioning better than 75 percent of the untreated control population. Eysenck’s twenty-five-year claim of complete therapeutic futility was comprehensively and empirically demolished.

4.3 Publication of ‘The Benefits of Psychotherapy’ (1980)

Following the monumental impact of their 1977 journal article, Glass, Smith, and methodologist Thomas I. Miller expanded their investigations into a comprehensive, book-length masterwork: The Benefits of Psychotherapy, published in 1980 by Johns Hopkins University Press. For this treatise, the researchers expanded their corpus to 475 controlled clinical trials, extracting over 1,700 individual effect sizes. Beyond confirming their initial finding that psychotherapy was overwhelmingly effective compared to no treatment, the book delved deep into complex moderator analyses that tackled another legendary controversy in clinical science: the comparative efficacy of distinct therapeutic modalities.

For decades, behavioral psychologists had claimed that behavioral techniques (such as systemic desensitization and operant conditioning) were vastly superior to “unscientific” talk therapies like psychoanalysis or client-centered counseling. Glass, Smith, and Miller conducted rigorous multiple regression models and ANOVA models comparing effect sizes across differing theoretical schools. Their findings delivered a second massive shock to the clinical establishment: when confounding variables—such as the reactive nature of outcome measures tailored specifically to behavioral tasks—were statistically controlled, there were virtually no meaningful, statistically significant differences in clinical effectiveness between major therapeutic orientations. Behavioral, cognitive, psychodynamic, and humanistic therapies all produced remarkably similar, robustly positive effect sizes.

This empirical confirmation became universally known as the empirical validation of the “Dodo Bird Verdict”—a literary reference to Lewis Carroll’s Alice in Wonderland (“Everybody has won, and all must have prizes”), originally postulated by Saul Rosenzweig in 1936. Glass and his colleagues demonstrated that the specific technical ingredients championed by rival psychological sects were far less important than common therapeutic factors: the therapeutic alliance, client expectancy, emotional catharsis, and the systematic provision of a structured rationale for healing. The Benefits of Psychotherapy fundamentally altered the trajectory of clinical psychology, laying the structural blueprint for modern evidence-based practice and serving as the gold standard for clinical literature syntheses in medicine and mental healthcare for decades to come.

5. The Class Size Debate: The Glass-Smith Meta-Analyses and Policy Repercussions

5.1 Historical Context of Class Size Research in American Schooling

Having conquered the empirical frontier of clinical psychology, Gene Glass turned his meta-analytic engine toward one of the most volatile, expensive, and politically contentious domains of American public education: the relationship between class size (pupil-teacher ratios) and academic achievement. For nearly a century, educational practitioners, teachers’ unions, and parents had intuitively maintained that smaller classes facilitated superior student learning, individualized instruction, and diminished behavioral disruptions. Conversely, school board administrators, municipal budget directors, and conservative economists consistently argued that the staggering fiscal expense of reducing class sizes was not justified by measurable academic returns.

The academic literature on class size was, predictably, a swamp of paralyzing contradiction. Landmark sociometric studies, such as James S. Coleman’s monumental 1966 report Equality of Educational Opportunity (the Coleman Report), had suggested that school structural resources—including class sizes, laboratory equipment, and teacher expenditures—had virtually no independent correlation with student achievement once family socioeconomic background was controlled. Prominent educational economists, most notably Eric Hanushek, published numerous narrative syntheses of the observational literature arguing that there was no consistent, statistically significant positive relationship between expenditure-driven class size reductions and student academic performance. Because observational studies suffered from profound confounding—such as low-achieving or behavioral-intervention students deliberately being placed in smaller remedial classes—unadjusted field data routinely showed small classes performing no better, or even worse, than large classes.

Faced with escalating municipal strikes, ballooning educational budgets, and an total absence of empirical clarity, the National Institute of Education (NIE) commissioned Gene Glass and Mary Lee Smith in the late 1970s to undertake a definitive, exhaustive meta-analysis of the global literature on class size and its relationship to both cognitive achievement and non-cognitive affective outcomes. The objective was to eliminate ideological speculation and discover whether empirical science could identify a mathematically precise curve relating pupil-teacher ratios to student learning.

5.2 The 1978–1979 Class Size Meta-Analyses

In 1978 and 1979, Glass and Smith released their landmark reports through the Far West Laboratory for Educational Research and Development, later publishing their findings in the Educational Evaluation and Policy Analysis journal. Their methodological strategy was uniquely sophisticated: to eliminate the massive confounding variables inherent in naturalistic observational studies, Glass and Smith concentrated on isolating within-study comparisons. They identified nearly 80 studies spanning decades of educational research that contained over 700 comparative comparisons of class sizes where the pupil assignments were either strictly randomized or carefully matched across identical curricula and instructional settings.

Using these rigorously extracted effect sizes, Glass and Smith plotted student achievement as a mathematical function of class size. Rather than assuming a simple, flat linear relationship, they fitted non-linear, polynomial regression models to the empirical data points. The resulting mathematical curve revealed a striking, elegant relationship:

• The relationship between class size and academic achievement was distinctly non-linear and logarithmic.
• Reducing class sizes incrementally from large configurations (e.g., from 40 students down to 30 students) produced only minor, modest improvements in standardized test performance.
• However, the regression curve exhibited a dramatic, unmistakable inflection point as class sizes dropped below 20 pupils, with achievement gains accelerating exponentially as class sizes moved toward 15, 10, and individual tutoring (1:1 instruction).

Glass and Smith demonstrated that a student taught in a class of 15 pupils would typically perform substantially higher on standardized achievement batteries than an equivalent student taught in a class of 30 pupils—an advantage equivalent to jumping multiple deciles on a national achievement distribution. Furthermore, their secondary meta-analyses on affective outcomes revealed that smaller classes generated monumental improvements in classroom operational climate, teacher morale, pupil engagement, and individualized diagnostic feedback. The Glass-Smith meta-analysis provided the first definitive, mathematically modeled empirical proof that class size reduction possessed genuine, powerful instructional efficacy, provided the reduction was deep enough to cross the sub-twenty threshold.

5.3 Statewide Policy Initiatives and Subsequent Controversies

The political and legislative consequences of the Glass-Smith class size meta-analyses were seismic. For the first time in American educational history, state legislatures possessed an authoritative, empirical benchmark with which to justify capital investments in educational staffing. The findings became the intellectual foundation for major state-sponsored empirical experiments, most notably the state of Tennessee’s renowned Project STAR (Student/Teacher Achievement Ratio) in the mid-1980s. STAR was a massive, longitudinal randomized controlled trial involving thousands of kindergarten through third-grade students. When the Project STAR results were published, they directly corroborated the exact non-linear predictions Glass and Smith had derived a decade earlier: young children placed in classes of 13 to 17 pupils outperformed peers in standard classes of 22 to 25 pupils by substantial, lasting margins.

The apex of this research-to-policy pipeline occurred in 1996, when California Governor Pete Wilson launched the California Class Size Reduction (CSR) initiative—a multi-billion-dollar statewide legislative mandate that provided financial subsidies to every public school district that capped kindergarten through third-grade classrooms at a strict maximum of 20 students. Within months, California school districts scrambled to build thousands of new classrooms and hired over 25,000 new elementary teachers to satisfy the policy requirements derived directly from the threshold discovered by Glass and Smith.

However, the rapid statewide implementation of California’s CSR initiative revealed severe institutional and macroeconomic trade-offs that triggered intense academic controversy. Critics, including economists such as Eric Hanushek, argued that while the internal validity of the Glass-Smith meta-analysis was technically sound, the real-world macro-policy implementation was economically inefficient and educationally disastrous. The sheer velocity of California’s mandate created an acute teacher shortage. Wealthy suburban districts recruited experienced, certified teachers from low-income, predominantly minority urban schools, forcing inner-city districts to fill classrooms with uncertified, emergency-credentialed novice instructors. As a consequence, the academic benefits of smaller classes in low-income schools were frequently neutralized by the dramatic collapse in average teacher quality. In subsequent retrospective analyses, Glass acknowledged these systemic challenges, reiterating that class size reduction could never function as a silver-bullet technocratic fix when severed from comprehensive investments in teacher preparation, facility infrastructure, and socio-economic equity.

6. Statistical Innovations: Effect Size Metrics and Quantitative Synthesis Frameworks

6.1 Mathematical Formulation of Glass’s Delta

The core statistical innovation that unlocked modern meta-analysis was the formulation of Glass’s Delta ($\Delta$). While standard inferential statistics prioritize the null-hypothesis significance test—yielding the ubiquitous $p$-value—Glass recognized that the $p$-value is an epistemologically deficient metric for research synthesis. Because $p$-values are a joint mathematical function of effect size and sample size ($N$), an astronomically small, trivial experimental effect can achieve statistical significance in a massive study ($N = 10,000$), while a massive, life-altering intervention can fail to reach significance in a pilot study ($N = 10$). Glass’s Delta circumvented this fatal limitation by extracting a pure, scale-free parameter of substantive magnitude that remained independent of study-level sample size.

The mathematical specification of Glass’s Delta relies on a critical theoretical decision regarding the denominator:

$$\Delta = \frac{\bar{X}_E – \bar{X}_C}{S_C}$$

Where $\bar{X}_E$ is the mean of the post-intervention experimental group, $\bar{X}_C$ is the mean of the control group, and $S_C$ is the unbiased sample standard deviation of the control group:

$$S_C = \sqrt{\frac{1}{N_C – 1} \sum_{i=1}^{N_C} (X_{iC} – \bar{X}_C)^2}$$

Glass explicitly championed the use of the control group standard deviation rather than a pooled standard deviation ($S_{pooled}$). His theoretical rationale was rigorous: in complex behavioral, psychological, and educational experiments, the experimental intervention itself frequently introduces variance heterogeneity. An intervention may help some participants dramatically while having no effect on, or even harming, others—thereby artificially expanding or contracting the experimental group variance ($S_E^2$). Under such conditions, pooling the experimental and control variances violates the assumption of homoscedasticity and distorts the natural scale of the baseline population. The control group variance, having remained unmanipulated by the intervention, represents the only pure, uncontaminated metric of baseline population variability.

In the years following Glass’s formulation, alternative effect size formulations emerged in the psychometric literature. Jacob Cohen popularized Cohen’s $d$, which utilized a pooled standard deviation assuming strict homogeneity of variance across conditions:

$$S_{pooled} = \sqrt{\frac{(N_E – 1)S_E^2 + (N_C – 1)S_C^2}{N_E + N_C – 2}}$$

Subsequently, Larry Hedges derived Hedges’ $g$, which applied a small-sample degrees-of-freedom correction factor ($J \approx 1 – \frac{3}{4(N_E + N_C) – 9}$) to eliminate the slight upward positive bias inherent in both Glass’s Delta and Cohen’s $d$ when applied to small sample sizes ($N < 20$). Furthermore, John E. Hunter and Frank L. Schmidt pioneered the “Hunter-Schmidt” meta-analytic framework in industrial-organizational psychology, which centered around the aggregation of correlation coefficients ($r$) corrected for psychometric measurement unreliability and range restriction. Despite these mathematical refinements, Glass’s Delta remains the foundational prototype from which all standardized mean difference metrics directly evolved.

6.2 Meta-Regression and Moderator Analysis

A second monumental statistical innovation introduced by Glass was the formalization of meta-regression and quantitative moderator analysis. Traditional reviewers operated under a binary, univariate mindset: they sought to answer monolithic questions such as “Does educational technology work?” or “Is psychodynamic therapy effective?” Glass recognized that these questions were fundamentally simplistic. In reality, treatment effects are complexly conditioned by the structural context of the intervention, the demographic makeup of the participants, the dosage and duration of exposure, and the methodological attributes of the experimental evaluation.

Glass demonstrated that the meta-analyst could transform individual studies into distinct data points within a general linear regression model. In this framework, the standardized effect size of study $j$ ($ES_j$) serves as the continuous dependent variable:

$$ES_j = \beta_0 + \beta_1 X_{1j} + \beta_2 X_{2j} + dots + \beta_k X_{kj} + \epsilon_j$$

Where $X_{1j}, X_{2j}, dots, X_{kj}$ represent coded, study-level explanatory variables (moderators), such as:

• The average age or grade level of the study sample;
• The duration of the intervention in hours, weeks, or semesters;
• The publication format (peer-reviewed journal article versus unpublished doctoral dissertation);
• The degree of alignment between the taught curriculum and the outcome testing battery (reactive versus non-reactive instruments);
• The presence or absence of random assignment versus quasi-experimental matching.

This formulation marked an epistemological leap from simple point estimation to the structured exploration of scientific heterogeneity. Rather than lamenting the presence of conflicting study outcomes, Glass utilized meta-regression to interrogate the sources of that variance mathematically. If an educational intervention yielded an average effect size of $+0.80$ in high-poverty urban schools but only $+0.20$ in affluent suburban schools, the regression coefficient on school socioeconomic status captured that critical interaction. Glass demonstrated that systematic heterogeneity was not noise to be swept away, but the very terrain where theoretical insight and policy discovery were located.

6.3 Addressing the File Drawer Problem and Publication Bias

From the inception of meta-analysis, Gene Glass was acutely conscious of the structural vulnerability that would come to be known as publication bias. In behavioral and medical sciences, journals routinely prioritized statistically significant, novel findings ($p < .05$), while studies yielding null, ambiguous, or counter-intuitive results were systematically rejected by peer reviewers or discarded by researchers into desk drawers—a phenomenon famously termed the “File Drawer Problem” by psychologist Robert Rosenthal in 1979. Glass realized that if a meta-analyst only synthesized peer-reviewed, published journal articles, the aggregate effect size would be systematically, artificially inflated by the structural censorship of null results.

To combat this existential threat to cumulative empirical science, Glass established an uncompromising methodological mandate: meta-analysts were ethically and scientifically obligated to conduct exhaustive literature searches that extended far beyond indexed commercial journals. Glass demanded the aggressive inclusion of “grey literature,” encompassing unpublished doctoral dissertations, institutional research reports, government evaluation monographs, conference proceedings, and unpublished technical working papers. In his own meta-analyses on psychotherapy and class size, Glass systematically extracted hundreds of unpublished dissertations, demonstrating empirically that doctoral dissertations—which were rarely subject to journal editorial biases favoring significance—frequently yielded slightly smaller, yet methodologically robust, effect sizes compared to published literature.

Glass’s uncompromising insistence on gray-literature inclusion catalyzed an entire sub-discipline of statistical methods dedicated to detecting, modeling, and correcting for publication bias. His early conceptual insights directly inspired subsequent mathematical developments, including Rosenthal’s “Fail-Safe N” (calculating the number of unpublished null studies required to bring a meta-analytic effect size down to statistical insignificance), Light and Pillemer’s funnel plots (identifying asymmetry in sample size versus effect magnitude distributions), and modern non-parametric techniques such as the “trim-and-fill” algorithm and selection models. Glass established the enduring scientific principle that a quantitative synthesis is only as valid as the comprehensiveness and integrity of its retrieval architecture.

7. Educational Evaluation Methodologies and Critique of Standardized Testing

7.1 Contributions to Standards-Setting Methodologies

In the late 1970s, parallel to his meta-analytic revolutions, Gene Glass directed his analytical focus toward the explosive national proliferation of the “minimum-competency testing” movement. Facing public panics over perceived declines in educational standards, state legislatures across the United States enacted statutes mandating that high school students pass standardized competency examinations to receive high school diplomas, and that educators achieve minimum passing scores on licensure tests. Central to this political movement was the necessity of establishing a strict “cut-score”—a numerical dividing line separating “competence” from “incompetence,” or “mastery” from “failure.” Psychometricians rushed to develop mathematical models—such as the Nedelsky, Angoff, and Ebel methods—claiming they could establish scientifically objective passing standards.

In 1978, Glass published a devastating, landmark critique in the Journal of Educational Measurement titled “Standards and Criteria.” In this masterwork of psychometric skepticism, Glass systematically dismantled the claim that objective, scientifically valid cut-scores could be established on continuous achievement distributions. He demonstrated that all standards-setting methodologies, regardless of their statistical ornamentation or algorithmic complexity, were fundamentally grounded in subjective human value judgments. Whether a committee of “subject matter experts” guessed how many minimally competent students would answer a multiple-choice item correctly (the Angoff method), or arbitrarily decided that passing required answering 70 percent of items correctly, the resulting cut-score was a social and political decision, not a discovery of natural law.

Glass exposed the pernicious social consequences of dressing up political value judgments in the clothes of psychometric pseudo-science. A cut-score set at an arbitrary threshold could instantly double the failure rate of socioeconomically marginalized or racial minority students, denying them diplomas and economic opportunities, all while state officials disclaimed political responsibility by claiming the test results were “purely objective statistical science.” Glass championed the vital distinction between criterion-referenced measurement (describing what a learner knows and can do in concrete, operational terms) and norm-referenced sorting (arbitrarily deciding who passes and who fails). His work in “Standards and Criteria” fundamentally reshaped educational measurement philosophy, forcing psychometricians to acknowledge the inescapable ethical, political, and subjective dimensions of standard setting.

7.2 Analysis of High-Stakes Accountability and Test-Driven Reform

As the minimum-competency movement of the 1970s metastasized into the high-stakes accountability regime of the 1990s and 2000s—crystallizing in the federal No Child Left Behind (NCLB) Act of 2001—Gene Glass emerged as one of the nation’s most formidable, empirically armed critics of test-driven school reform. NCLB mandated that every state implement annual standardized testing in grades three through eight, establishing arbitrary escalating benchmarks of “Adequate Yearly Progress” (AYP) with the absurd, mathematically impossible goal of achieving 100 percent student proficiency by the year 2014. Schools that failed to hit these escalating benchmarks faced severe punitive sanctions, including loss of Title I funding, state reconstitution, forced conversion to charter status, or mass firings of instructional staff.

Glass subjected the NCLB accountability framework to relentless quantitative scrutiny. He documented how high-stakes testing paradigms functioned as a textbook validation of Campbell’s Law—the sociological principle formulated by his mentor’s collaborator, Donald T. Campbell, which states: “The more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor.” Glass provided empirical evidence showing that attaching existential high-stakes sanctions to standardized test scores did not improve genuine cognitive learning; instead, it induced widespread systemic corruption across the public education ecosystem.

Through empirical analyses, Glass and his collaborators documented how high-stakes accountability pressures resulted in:

• Severe curriculum narrowing, wherein schools radically eliminated non-tested subjects such as history, music, fine arts, physical education, and recess in order to drill students endlessly on low-level, test-aligned mathematics and reading items;
• Widespread manipulation of testing pools, including the strategic suspension of low-performing students during testing weeks, the misclassification of students into special education tracks to exclude their scores, and artificial increases in grade retention;
• A massive illusion of academic progress, wherein states lowered their proficiency cut-scores or engaged in blatant test-score inflation on high-stakes state assessments, while flatlining or declining on the low-stakes, non-corrupted National Assessment of Educational Progress (NAEP).

Glass showed that test-driven accountability was not a neutral diagnostic tool, but an aggressive institutional mechanism that widened educational inequality, demoralized the teaching profession, and alienated students from authentic intellectual inquiry.

7.3 Value-Added Models (VAM) and Teacher Evaluation Critiques

In the late 2000s and 2010s, high-stakes accountability shifted its primary target from the school building directly onto the individual classroom instructor through the proliferation of Value-Added Models (VAM). Supported by the federal Race to the Top initiative and massive corporate philanthropic financing, state departments of education adopted econometric growth models designed to isolate the unique “value added” by a single teacher to a student’s longitudinal standardized test trajectory. Teachers whose students failed to achieve predicted test-score gains were systematically labeled “ineffective,” leading to public rankings in metropolitan newspapers, termination of employment, and denial of tenure.

Glass launched an eviscerating methodological critique of VAM, publishing technical analyses that exposed the profound statistical instability and conceptual invalidity of these econometric models. He demonstrated that teacher VAM scores exhibited shocking, unacceptable year-to-year volatility. A teacher ranked in the top 10 percent of effectiveness in a given school year had a nearly equal probability of being ranked in the bottom quartile the following year simply due to random variations in student classroom composition, localized testing interruptions, or measurement error. The test instruments themselves—designed for broad school-level or regional norming—lacked the psychometric precision and curricular sensitivity required to serve as individual performance metrics.

More fundamentally, Glass attacked the core econometric assumption of VAM: the assertion that statistical control variables (such as free-and-reduced lunch status or prior test scores) could successfully control for the deeply entrenched, non-random sorting of children into classrooms and schools. Classrooms are not randomized clinical trials; children are systematically sorted based on special educational needs, parental advocacy, behavioral accommodations, and socioeconomic micro-geographies. Glass argued that attributing complex, multi-year, socially embedded cognitive developmental trajectories solely to the isolated “treatment” of a single adult standing at the front of a classroom for nine months was an act of gross statistical hubris. He advocated for the immediate cessation of high-stakes VAM applications, calling instead for multifaceted, qualitative, peer-assisted, and institutional systems of clinical evaluation.

8. Pioneering Open Access: Founding Education Policy Analysis Archives (EPAA)

8.1 Genesis of Digital Scholarly Communication in 1993

In 1993, long before the terms “Open Access,” “Creative Commons,” or “digital scholarship” had entered the general academic lexicon, Gene Glass executed an audacious, revolutionary institutional initiative: he founded the Education Policy Analysis Archives (EPAA). At that historical juncture, academic publishing was completely dominated by a small, predatory cartel of commercial publishing houses (such as Elsevier, Springer, and Taylor & Francis) and conservative professional associations. These entities locked peer-reviewed research findings behind exorbitantly priced institutional subscriptions and print paywalls. Educational policy research—which was largely financed by public taxpayer dollars—was effectively hidden from the teachers, administrators, school board members, and citizens who desperately needed it to make informed democratic decisions.

Glass recognized that the nascent architecture of the global Internet, specifically the World Wide Web, university mainframe servers, and early electronic networks (such as BITNET and early HTML protocols), offered a historic opportunity to bypass the commercial publishing monopoly entirely. Operating from his office at Arizona State University, Glass single-handedly built, coded, and edited the first issues of EPAA. There were no submission fees, no page processing charges, no subscription paywalls, and no corporate advertising. Every article published was instantly, freely accessible to anyone with an Internet connection anywhere on the planet.

EPAA was not an informal blog or an unvetted repository; it was an uncompromising, peer-reviewed academic journal designed to adhere to the most rigorous standards of methodological and theoretical scholarship. Glass leveraged his immense international stature as a former AERA president and quantitative titan to attract submissions from the world’s foremost educational researchers, sociologists, and policy analysts. By demonstrating that a world-class, peer-reviewed journal could be produced and distributed globally for a fraction of the cost of traditional print journals, Glass became one of the foundational architects of the international open-access movement, anticipating developments such as the Public Knowledge Project (PKP) and the Budapest Open Access Initiative by nearly a decade.

8.2 Editorial Philosophy and Policy Impact of EPAA

The editorial philosophy that Glass instilled within EPAA was characterized by rapid intellectual turnaround, methodological pluralism, and radical civic engagement. In traditional commercial journals, the peer-review and publication pipeline routinely dragged on for eighteen months to three years—a sluggish timeline that rendered published research utterly useless for ongoing legislative debates and educational policy crises. Glass streamlined the editorial process, utilizing digital peer-review networks to deliver decisions and publish accepted manuscripts within weeks, directly injecting rigorous empirical scholarship into contemporaneous policy struggles.

Under Glass’s pioneering editorial leadership, EPAA published some of the most influential and widely cited educational policy investigations of the late twentieth and early twenty-first centuries. The journal became the primary publishing venue for critical empirical analyses on:

• The real-world outcomes of school voucher systems and educational privatization experiments in cities like Milwaukee and Cleveland;
• The demographic and economic consequences of racial desegregation rollbacks and court-mandated unitary status decisions;
• State school finance litigation and the structural inequities of property-tax-based educational funding systems;
• The empirical flaws of bilingual education bans and the academic impacts of English-only immersion policies across the American Southwest.

Furthermore, Glass pioneered cross-cultural and international scholarly dissemination by transforming EPAA into a multilingual publication platform. Recognizing that linguistic barriers reinforced North American and Eurocentric hegemony in educational science, Glass expanded the journal to accept, peer-review, and publish articles in Spanish and Portuguese alongside English. This strategic internationalization established EPAA as a preeminent academic conduit for educational researchers across Latin America, Brazil, and the Iberian Peninsula, bridging intellectual divides and fostering a truly global dialogue on educational equity and democratic governance.

8.3 Long-Term Significance for Academic Publishing

The long-term institutional significance of Gene Glass’s work with the Education Policy Analysis Archives cannot be overstated. By successfully sustaining EPAA through its critical formative decades, Glass demonstrated the viability and resilience of independent, diamond open-access publishing models (free for both authors and readers). He proved that scholarly communities could maintain the highest echelons of academic rigor, indexing prestige, and citation impact without surrendering copyright and public intellectual capital to commercial rent-seeking publishers.

As the journal expanded, Glass ensured its long-term sustainability by transitioning its institutional stewardship to successive editorial leadership teams based at Arizona State University’s Mary Lou Fulton Teachers College. Today, EPAA remains one of the longest-running, continuously published, fully open-access academic journals in human history, having published thousands of research articles that have been downloaded millions of times by researchers, grassroots educators, and policymakers across the globe.

Throughout his career, Glass remained an outspoken, caustic critic of the academic publishing industry’s commercial exploitation. He routinely castigated university administrations for forcing faculty into publishing behind commercial paywalls to achieve promotion and tenure. Glass argued that public university scholars had a fundamental ethical obligation to ensure that knowledge produced through public support was returned unconditionally to the public commons. His visionary work with EPAA established a durable institutional paradigm that continues to inspire the global movement toward open science, transparent research data, and epistemic democratization.

9. Critical Perspectives on Modern Education Reform, Charters, and Marketization

9.1 The Neo-Liberal Restructuring of Public Education

From the mid-1990s onward, Gene Glass directed his intellectual and analytical energies toward critiquing the structural reorganization of American public schooling under the hegemony of neoliberalism. Beginning with the publication of the Reagan administration’s alarmist 1983 report A Nation at Risk, a bipartisan political consensus emerged asserting that American public schools were systematically failing, and that this failure could only be cured through the importation of free-market mechanisms: school choice, competitive vouchers, tuition tax credits, high-stakes merit pay, and the rapid expansion of privately managed charter schools. Glass viewed this movement not as a benign search for educational improvement, but as a coordinated, ideological campaign aimed at the privatization and dismantling of a vital public institution.

Glass subjected the central claims of the market-reform movement to rigorous empirical refutation. Conservative think-tanks and free-market economists argued that introducing voucher competition would force traditional public schools to “compete or close,” naturally lifting academic achievement across the entire educational system. Glass marshaled longitudinal demographic and testing data to demonstrate that market competition produced no such systemic improvement. Instead, school choice mechanisms operated as powerful demographic sorting machines, driving racial and socioeconomic resegregation.

Through incisive scholarly critiques, Glass documented how charter schools and voucher programs functioned to siphon off the most motivated families and highest-performing students, leaving under-resourced traditional public neighborhood schools to educate disproportionate concentrations of children living in deep poverty, students with severe physical and cognitive disabilities, and English-language learners. Furthermore, Glass exposed the pervasive influence of corporate philanthropy—spearheaded by multi-billion-dollar private foundations such as the Bill & Melinda Gates Foundation, the Broad Foundation, and the Walton Family Foundation—in aggressively underwriting and manufacturing political consensus for unproven, market-driven school reforms without democratic transparency or public accountability.

9.2 The Book ‘Fertilizers, Pills, and Magnetic Strips’ (2008)

In 2008, Glass published one of his most conceptually audacious and interdisciplinary masterworks: Fertilizers, Pills, and Magnetic Strips: The Fate of Public Education in America (published by Information Age Publishing). In this provocative treatise, Glass stepped back from narrow psychometric technicalities to construct a sweeping socioeconomic and historical critique of the forces distorting contemporary education policy. The title itself encapsulated an extended, brilliant historical analogy drawn between the transformation of agriculture, medicine, and commerce, and the impending corporate capture of public education.

Glass examined how modern industrial agriculture had been transformed by chemical fertilizers, how clinical medicine had been dominated by pharmaceutical pills, and how everyday commerce had been reorganized around magnetic credit card strips. In each of these domains, complex, human, and ecological processes were systematically reduced to standardized, commodified, and technocratic transactions driven by corporate profit extraction. Glass argued that the identical corporate logic was now descending upon public education. Tech companies, testing conglomerates, and education management organizations (EMOs) were peddling reductive “techno-fixes”—automated instructional software, standardized diagnostic scripts, and corporate data platforms—as technological substitutes for authentic, relational, and well-funded human instruction.

The central thesis of Fertilizers, Pills, and Magnetic Strips was unyielding: the profound academic achievement gaps that plagued American schooling were not the result of defective teachers, incompetent administrators, or a lack of market competition; they were the direct, mathematical reflection of endemic societal poverty. Glass demonstrated that when one statistically controls for childhood poverty, family income insecurity, lack of access to healthcare, and community distress, American public schools perform at the very top of international achievement distributions. Glass argued that seeking to cure educational disparities through technocratic curriculum gimmicks or market-based choice mechanisms was an act of deliberate political bad faith. Real, lasting educational improvement, he asserted, could never be achieved without fundamental structural investments in the social safety net: universal healthcare, affordable housing, living wages, and direct poverty eradication.

9.3 Collaborative Scholarship on School Commercialism and Online Schooling

Throughout the late 2000s and 2010s, Glass deepened his collaborative research initiatives with scholars at the National Education Policy Center (NEPC), located at his former institutional home, the University of Colorado Boulder. Working alongside critical methodologists, including Alex Molnar and Kevin Welner, Glass turned his investigative attention toward the rapid encroachment of corporate commercialism, targeted marketing, and private profit models inside American classrooms.

A major focus of this investigative scholarship centered on the burgeoning industry of for-profit virtual (online) charter schools. Operated by publicly traded educational conglomerates such as K12 Inc. and Connections Academy, virtual charter schools recruited tens of thousands of vulnerable students into fully online, asynchronous instructional environments, collecting billions of public tax dollars while incurring minimal operational costs for physical facilities and instructional labor. Glass and his NEPC colleagues produced comprehensive, multi-year empirical audits evaluating the operational and academic outcomes of these virtual academies.

The findings of these empirical investigations were devastating. Glass documented that online charter schools exhibited:

• Catastrophically low graduation rates, frequently hovering below 50 percent;
• Staggering student attrition and mobility rates, where up to half the student body rotated out of the school within a single academic year;
• Cumulative academic achievement deficits of historic proportions, with students falling months or years behind their traditional public school peers in basic mathematics and reading;
• Rampant corporate profiteering, where public educational funds were systematically siphoned off into executive compensation, aggressive digital marketing campaigns, and corporate shareholder dividends while student-teacher ratios exceeded hundreds of pupils per remote instructor.

Glass’s analytical work with the NEPC provided state legislators, regulatory agencies, and investigative journalists with the rigorous, unassailable empirical ammunition needed to demand audits, block expansion proposals, and enact statutory oversight over predatory educational corporations.

10. Academic Leadership, Editorial Influence, and Institutional Roles

10.1 Regents’ Professorship at Arizona State University

In 1986, after more than two decades of distinguished service at the University of Colorado Boulder, Gene Glass transitioned his academic home to the College of Education at Arizona State University in Tempe, Arizona. At ASU, Glass found an expansive, dynamic institution committed to reimagining public higher education. He was appointed to the prestigious rank of Regents’ Professor—the highest academic faculty honor bestowed by the Arizona Board of Regents—spanning both the Division of Educational Leadership and Policy Studies and the Department of Psychology in Education.

At Arizona State, Glass played a transformative role in building the university’s doctoral infrastructure in quantitative research methods, measurement, and education policy analysis. He mentored multiple generations of doctoral students and post-doctoral fellows who would go on to assume prominent faculty chairs, directorships of institutional research, and leadership posts in state and federal agencies across the globe. Glass’s mentorship was legendary for its intellectual generosity; he rejected the hierarchical, exploitative dynamics common in high-stakes academic laboratories, actively encouraging his graduate students to co-author foundational papers as equal intellectual partners and to pursue unorthodox, paradigm-shifting research trajectories.

Beyond his direct mentorship, Glass was instrumental in shaping the institutional ethos of ASU’s Mary Lou Fulton Teachers College. He championed an interdisciplinary research paradigm that fused statistical rigor with qualitative sociology, critical theory, and public policy advocacy. Even within a rapidly expanding university system, Glass remained a voice of academic integrity, consistently fighting against institutional administrative bloat, demanding transparency in university governance, and insisting that higher education must serve the democratic interests of the local community and the state.

10.2 Professional Governance and Institutional Service

Gene Glass’s influence extended across the highest echelons of national and international scientific governance. His election as President of the American Educational Research Association (AERA) for the 1975–1976 term placed him at the center of national educational research policy at a critical historical moment. Within AERA, Glass utilized his leadership platform to modernize the organization’s publication portfolios, demand higher standards of empirical transparency across association journals, and create institutional space for emerging quantitative and qualitative methodologies.

In recognition of his monumental contributions to behavioral measurement and research methodology, Glass was elected to the National Academy of Education (NAEd)—an elite, congressionally chartered body of the nation’s most distinguished educational scholars. Within the National Academy, Glass served on expert study panels addressing high-stakes testing, standard setting, program evaluation methodologies, and the ethical use of student data. His contributions were instrumental in producing national advisory reports that cautioned federal and state agencies against using reductive standardized testing scores as the primary determinants of school closure, student retention, or teacher dismissals.

Furthermore, Glass served on numerous national advisory committees, scientific review boards, and technical panels for organizations including the National Research Council, the National Science Foundation, and the American Psychological Association. In these institutional roles, Glass functioned as a critical methodological bulwark against statistical overreach. Colleagues frequently noted that when commercial testing corporations or technocratic agencies proposed dubious statistical algorithms to measure complex human learning, Glass was the singular voice in the room possessing the mathematical brilliance to dismantle the algorithm’s flaws on the blackboard and the rhetorical eloquence to explain its real-world human dangers to policymakers.

10.3 Emeritus Transitions and Continued Scholarly Engagements

Upon achieving emeritus status at Arizona State University, Gene Glass did not retreat into quiet academic retirement. Instead, he maintained dual institutional affiliations, serving as an active research scholar and visiting professor at the Connie L. Lurie College of Education at San José State University in California, while remaining intimately connected to intellectual life at ASU. His transition to emeritus status provided him with complete institutional independence, freeing him from university committee obligations and allowing him to dedicate himself entirely to public scholarship, political writing, and contemporary methodological debates.

Glass swiftly adapted to the digital transformations of the twenty-first century, becoming a pioneer of public academic blogging. Through platforms such as his personal blog, EdFacts, and collaborative educational policy forums, Glass published frequent, sharp, and accessible empirical commentaries on ongoing educational crises. When state departments of education issued press releases trumpeting miraculous test-score gains under new charter initiatives, Glass would download the raw state datasets, run secondary regression analyses overnight, and publish devastating blog posts revealing that the gains were statistical artifacts caused by selective student attrition or changed exclusion rules.

In his emeritus writings, Glass has continued to offer profound, self-reflective critiques regarding the contemporary trajectory of quantitative social science. While celebrating the global ubiquity of meta-analysis and empirical synthesis, he has consistently cautioned younger researchers against falling into the trap of mindless algorithmic technocracy. To Glass, statistical technique must always remain the humble servant of humanistic and democratic values. His continued interventions into debates surrounding algorithmic bias, artificial intelligence in education, and the ongoing privatization of public assets demonstrate a scholar whose critical intellectual acuity remains as fierce and uncompromising in his ninth decade as it was when he introduced meta-analysis in 1976.

11. Epistemological and Methodological Debates: Critiques and Refinements of Meta-Analysis

11.1 The Quality Sorting Debate: Inclusion Criteria vs. Methodological Screening

Throughout the evolution of quantitative research synthesis, the most persistent, philosophically deep methodological debate has centered on the tension between Gene Glass’s exhaustive inclusion philosophy and the selective screening approach championed by prominent scholars, most notably Robert E. Slavin. In the mid-1980s, Slavin launched a sustained critique of Glassian meta-analysis, arguing that including methodologically flawed studies (such as non-randomized quasi-experiments or studies with massive participant attrition) alongside pristine, highly controlled trials produced distorted and misleading aggregate conclusions. Slavin proposed an alternative paradigm he coined “Best-Evidence Synthesis.”

Slavin’s best-evidence approach asserted that the meta-analyst should establish rigorous, a priori methodological criteria—such as minimum duration requirements, true random assignment, and standardized non-reactive outcome measures. Studies failing to meet these strict gold standards would be systematically discarded from the synthesis before any effect sizes were calculated. Slavin argued that a synthesis based on five flawless randomized trials provided a vastly more accurate representation of truth than a synthesis based on five hundred messy, uncontrolled classroom studies.

Glass mounted an unyielding theoretical and empirical defense of his exhaustive inclusion model. He pointed out a fatal epistemological flaw in Slavin’s paradigm: the determination of what constitutes “best evidence” is itself inherently subjective, idiosyncratic, and susceptible to the reviewer’s personal theoretical biases. One reviewer might decide that only studies lasting a full academic year are valid, while another insists that only studies with blinded evaluators qualify. In practice, selective screening criteria routinely served as an intellectual filter used to purge studies that contradicted an author’s favored conclusions.

Glass demonstrated that the scientifically rigorous solution was not to discard studies a priori, but to operationalize study quality as an empirical moderator variable:

• Every study in the literature should be included in the analytical corpus;
• Methodological features—such as randomization status, sample attrition, blinding, and instrument reactivity—must be systematically coded as independent categorical or continuous variables;
• Statistical models (ANOVA or meta-regression) must be computed to directly test whether effect sizes correlate with methodological quality.

If the data show that methodologically weak studies systematically yield inflated effect sizes, the researcher can mathematically discount or separate them in the final reporting. If, however, the data show no empirical difference in effect magnitude between weak and strong studies, discarding the weaker studies serves only to throw away valuable statistical power and sample diversity. Glass’s empirical approach transformed “study quality” from a subjective philosophical dispute into an empirical, testable scientific question.

11.2 Statistical Refinements by Hedges, Olkin, and Rosenthal

While Gene Glass laid the conceptual and practical foundations of meta-analysis, the statistical mechanics of the methodology were rapidly expanded, formalized, and mathematically refined during the late 1970s and 1980s by a brilliant group of mathematical statisticians, led by Larry V. Hedges, Ingram Olkin, and Robert Rosenthal. In Glass’s original formulations, effect sizes were typically treated as simple observations that were averaged together using unweighted arithmetic means, or analyzed using ordinary least squares (OLS) regression models.

Hedges and Olkin pointed out a fundamental statistical problem with unweighted effect size averaging: it treated a small study with a sample size of ten participants ($N = 10$) with the identical statistical weight as a massive, multi-site clinical trial with ten thousand participants ($N = 10,000$). Because the sampling variance of an effect size ($v_i$) is a direct inverse function of sample size:

$$v_i = \text{Var}(d_i) \approx \frac{N_E + N_C}{N_E N_C} + \frac{d_i^2}{2(N_E + N_C)}$$

Hedges and Olkin established the modern system of inverse-variance weighting. In this framework, each effect size is multiplied by a precision weight ($w_i$), which is mathematically equal to the inverse of its sampling variance ($w_i = \frac{1}{v_i}$). Larger, more precise studies contribute proportionally more weight to the pooled effect size, while tiny, volatile studies contribute minimal weight, maximizing the statistical efficiency and minimizing the standard error of the composite estimate.

Furthermore, Hedges and Olkin pioneered the critical distinction between fixed-effects models (which assume that all studies share a single, identical true population effect size, and that all observed variance is pure within-study sampling error) and random-effects models (which assume that the true effect size varies across studies due to differences in populations and implementations, requiring the estimation of between-study variance, $\tau^2$). While Glass recognized and embraced these advanced mathematical developments, he maintained a healthy, pragmatic skepticism toward excessive statistical hyper-specialization. He repeatedly warned that mathematical over-formalization risked creating an insular, technocratic elite of statistical modelers who lost touch with the practical, substantive, and political realities of the educational and psychological phenomena they were ostensibly synthesizing.

11.3 Meta-Analysis in Medicine and the Cochrane Collaboration

The most consequential historical trajectory of Gene Glass’s methodological innovation was its dramatic migration from educational psychology into clinical medicine and global public health. In the late 1970s and early 1980s, British epidemiologist Archie Cochrane issued a stinging indictment of the medical profession, observing that clinical medicine had failed to create a systematic, continuously updated critical summary of all randomized controlled clinical trials. Millions of patients were suffering or dying because life-saving medical discoveries—or, conversely, dangerous pharmaceutical side-effects—were buried in scattered, conflicting medical journals across the world.

Medical researchers, led by clinical epidemiologists such as Sir Iain Chalmers, directly imported Gene Glass’s meta-analytic methodology to revolutionize healthcare synthesis. In 1993, Chalmers and an international consortium of physicians founded the Cochrane Collaboration (now known simply as Cochrane), an international independent non-profit organization dedicated to producing systematic reviews and meta-analyses of all healthcare interventions. Meta-analysis became the absolute methodological gold standard of the international Evidence-Based Medicine (EBM) movement, positioning systematic quantitative reviews at the very pinnacle of the medical evidence hierarchy—above randomized controlled trials, cohort studies, and expert clinical opinion.

Today, the quantitative synthesis of odds ratios, relative risks, and standardized mean differences directly determines international treatment protocols for oncology, cardiology, psychiatric pharmacology, and pandemic response. However, interesting cultural and methodological divergences exist between the social science origins of meta-analysis and its medical manifestation:

• Medical meta-analyses typically enforce far narrower inclusion criteria than Glass originally advocated, frequently restricting synthesis strictly to double-blind randomized controlled trials (RCTs);
• Medical reviews rely heavily on structured risk-of-bias tools (such as the Cochrane Risk of Bias Tool) to evaluate methodological internal validity;
• Conversely, educational and social scientific meta-analyses, continuing in the Glassian tradition, routinely confront massive, messy non-experimental datasets and prioritize complex meta-regression and moderator analyses to capture context-specific structural factors.

Despite these disciplinary adaptations, the underlying mathematical architecture of modern evidence-based healthcare is the direct intellectual offspring of Gene Glass’s 1976 presidential address.

12. Enduring Legacy and Contemporary Relevance of Gene Glass’s Scholarship

12.1 The Ubiquity of Meta-Analysis Across Modern Scientific Disciplines

The intellectual footprint of Gene Glass’s core innovation has achieved a level of universal, cross-disciplinary ubiquity matched by few methodological advances in the history of science. What originated in 1976 as a controversial proposal by a young educational psychologist has evolved into an indispensable, foundational pillar of global quantitative science. Meta-analysis is now universally practiced across clinical medicine, psychiatric epidemiology, agricultural research, evolutionary biology, behavioral economics, environmental toxicology, and criminology (institutionalized through the Campbell Collaboration, the social-science sister organization to Cochrane).

Meta-analysis has fundamentally altered the scientific understanding of cumulative progress. Prior to Glass, the history of behavioral and educational science was plagued by cynical cycles of excitement and disillusionment. A researcher would publish a groundbreaking experiment; a decade later, another laboratory would fail to replicate it; a third laboratory would report mixed results; and the scientific community would abandon the inquiry as hopelessly intractable. Meta-analysis broke this destructive cycle by demonstrating that apparent empirical contradictions are frequently the natural, mathematical consequence of sampling error operating within underpowered individual studies.

By providing a mathematical framework to aggregate thousands of independent research outcomes, meta-analysis has made it possible to map the genuine topography of human knowledge. It has allowed humanity to quantify, with breathtaking mathematical precision, the efficacy of vaccines, the impact of smoking cessation programs, the cognitive benefits of early childhood education, and the systemic effects of micro-lending programs in the developing world. The term coined by Glass in San Francisco nearly fifty years ago has become an indispensable epistemic mechanism by which science accumulates reliable truths.

12.2 The Integration of Methodological Rigor and Social Justice

Perhaps the most extraordinary and enduring dimension of Gene Glass’s legacy is his seamless, uncompromising integration of advanced mathematical sophistication with radical social justice advocacy. In the dominant culture of modern academic science, researchers are frequently conditioned to accept a false dichotomy: one must either be a detached, apolitical quantitative methodologist who views human suffering through cold, abstract matrices, or a qualitative social activist who views statistical metrics with deep ideological suspicion.

Gene Glass shattered this false binary. Throughout his career, he demonstrated that uncompromising mathematical brilliance could serve as a powerful, devastating weapon against institutional injustice, technocratic oppression, and social inequality. Rather than retreating into an ivory-tower cloister of abstract formulaic derivations, Glass repeatedly brought his statistical expertise into direct combat with the most powerful institutions of American society. He utilized quantitative metrics to prove the humanity and clinical recoverability of mental health patients against Eysenckian nihilism; he proved the tangible, structural benefits of smaller classes for vulnerable children against corporate budget-cutters; and he exposed the statistical fraudulence of using standardized tests to sort poor and minority children into educational oblivion.

Glass refused to decouple empirical statistics from ethical and humanitarian obligations. He consistently cautioned against the technocratic hubris that assumes social problems can be solved by algorithms severed from human empathy and democratic accountability. In doing so, he provided an enduring model for what the public intellectual must be: a scholar possessing the technical mastery to expose the mathematical fallacies of the powerful and the moral courage to speak truth on behalf of the vulnerable.

12.3 Gene Glass’s Enduring Monument in Intellectual History

As the scientific community reflects upon the expansive trajectory of twentieth- and twenty-first-century scholarship, Gene V. Glass stands as a towering, monumental figure whose contributions have fundamentally reshaped how humanity evaluates empirical truth. His major lifetime achievements—the invention and theoretical formalization of meta-analysis, the empirical resolution of the psychotherapy and class-size debates, the pioneering launch of the global open-access movement through EPAA, and his incisive critiques of test-driven accountability and educational marketization—represent an intellectual corpus of historic proportions.

In an era increasingly dominated by corporate algorithms, automated artificial intelligence models, commercialized academic publishing monopolies, and relentless efforts to dismantle public democratic institutions, Glass’s scholarly warnings resonate with prophetic urgency. His writings continue to remind us that education is not an assembly line for producing standardized human capital, that social science is not a corporate consulting service, and that public institutions belong to the public commons, not to private markets.

Gene Glass provided the modern world with the mathematical and conceptual tools necessary to synthesize the scattered fragments of scientific inquiry into coherent, cumulative human knowledge. Simultaneously, he instilled within the scientific vocation an uncompromising moral conscience, proving that the highest purpose of quantitative analysis is not to sort, control, and commodify human beings, but to liberate them. His life and scholarship remain an enduring monument to the profound unity of empirical rigor, scientific transparency, and democratic egalitarianism.

Conclusion

The lifetime intellectual journey of Gene V. Glass, spanning from the post-war agrarian landscapes of Nebraska to the forefront of international scientific methodology, reflects a singular, unified quest: the pursuit of methodological transparency in service of democratic equality. Confronted with a mid-twentieth-century behavioral science paralyzed by subjective literature reviews, ideological entrenchments, and fragmented data, Glass possessed both the mathematical acumen to invent a new empirical methodology and the epistemological daring to redefine how cumulative science operates. Meta-analysis fundamentally dissolved the myth of the single, definitive study, providing modern disciplines with an objective, scale-free architecture capable of systematically mining the global scientific archive.

Yet, the defining genius of Glass resides in the fact that he never permitted his quantitative innovations to degenerate into cold, apolitical technocracy. Whether resolving the bitter post-war debates regarding the efficacy of psychotherapy, deriving the profound non-linear achievement curves of class size reductions, exposing the subjective arbitrariness of high-stakes testing cut-scores, or single-handedly launching the diamond open-access movement with the Education Policy Analysis Archives, Glass mobilized empirical statistics as an engine of human liberation and public institutional defense. His career stands as an unyielding refutation of the neoliberal marketization of education and a timeless demonstration that advanced mathematical rigor, far from being the antithesis of social justice, can function as its most formidable empirical vanguard.

References

• Carroll, L. (1865). Alice’s Adventures in Wonderland. Macmillan.
• Cochrane, A. L. (1972). Effectiveness and Efficiency: Random Reflections on Health Services. Nuffield Provincial Hospitals Trust.
• Cohen, J. (1977). Statistical Power Analysis for the Behavioral Sciences (Rev. ed.). Academic Press.
• Coleman, J. S., Campbell, E. Q., Hobson, C. J., McPartland, J., Mood, A. M., Weinfeld, F. D., & York, R. L. (1966). Equality of Educational Opportunity. U.S. Government Printing Office.
• Eysenck, H. J. (1952). The effects of psychotherapy: An evaluation. Journal of Consulting Psychology, 16(5), 319–324. https://doi.org/10.1037/h0063633
• Eysenck, H. J. (1978). An exercise in mega-silliness. American Psychologist, 33(5), 517. https://doi.org/10.1037/0003-066X.33.5.517.a
• Glass, G. V. (1976). Primary, secondary, and meta-analysis of research. Educational Researcher, 5(10), 3–8. https://doi.org/10.3102/0013189X005010003
• Glass, G. V. (1978). Standards and criteria. Journal of Educational Measurement, 15(4), 237–261. https://doi.org/10.1111/j.1745-3984.1978.tb00069.x
• Glass, G. V. (2008). Fertilizers, Pills, and Magnetic Strips: The Fate of Public Education in America. Information Age Publishing.
• Glass, G. V., McGaw, B., & Smith, M. L. (1981). Meta-Analysis in Social Research. SAGE Publications.
• Glass, G. V., & Smith, M. L. (1979). Meta-analysis of research on class size and achievement. Educational Evaluation and Policy Analysis, 1(1), 2–16. https://doi.org/10.3102/01623737001001002
• Glass, G. V., & Stanley, J. C. (1970). Statistical Methods in Education and Psychology. Prentice-Hall.
• Glass, G. V., Willson, V. L., & Gottman, J. M. (1975). Design and Analysis of Time-Series Experiments. Colorado Associated University Press.
• Hedges, L. V. (1981). Distribution theory for Glass’s estimator of effect size and related estimators. Journal of Educational Statistics, 6(2), 107–128. https://doi.org/10.3102/10769986006002107
• Hedges, L. V., & Olkin, I. (1985). Statistical Methods for Meta-Analysis. Academic Press.
• Hunter, J. E., & Schmidt, F. L. (1990). Methods of Meta-Analysis: Correcting Error and Bias in Research Findings. SAGE Publications.
• Molnar, A., Miron, G., Elgeberi, N., Barbour, M. K., Huerta, L., Pratt, C. Q., & Charbonneau, J. K. (2019). Virtual Schools in the U.S. 2019. National Education Policy Center. https://nepc.colorado.edu/publication/virtual-schools-annual-2019
• Nichols, S. L., Glass, G. V., & Berliner, D. C. (2006). High-stakes testing and student achievement: Does accountability pressure increase student learning? Education Policy Analysis Archives, 14(1), 1–175. https://doi.org/10.14507/epaa.v14n1.2006
• Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86(3), 638–641. https://doi.org/10.1037/0033-2909.86.3.638
• Rosenzweig, S. (1936). Some implicit common factors in diverse methods of psychotherapy. American Journal of Orthopsychiatry, 6(3), 412–415. https://doi.org/10.1111/j.1939-0025.1936.tb05248.x
• Slavin, R. E. (1986). Best-evidence synthesis: An alternative to meta-analytic and traditional reviews. Educational Researcher, 15(9), 5–11. https://doi.org/10.3102/0013189X015009005
• Smith, M. L., & Glass, G. V. (1977). Meta-analysis of psychotherapy outcome studies. American Psychologist, 32(9), 752–760. https://doi.org/10.1037/0003-066X.32.9.752
• Smith, M. L., Glass, G. V., & Miller, T. I. (1980). The Benefits of Psychotherapy. Johns Hopkins University Press.