Gene V. Glass

Meta-Analysis of Psychotherapy Outcome Studies

Results of nearly 400 controlled evalua- tions of psychotherapy and counseling were coded and integrated statistically. The findings provide convincing evidence of the efficacy of psychotherapy. On the average, the typical therapy client is better off than 75% of untreated individuals. Few important differ- ences in effectiveness could be established among many quite different types of psychotherapy . More generally, virtually no difference in effectiveness was observed be- tween the class of all behavioral therapies (systematic desensitization, behavior modification) and the nonbe- havioral therapies (Rogerian, psychodynamic, rational- emotive, transactional analysis, etc.).

9: Integrating Findings: The Meta-Analysis of Research

Maccoby and Jacklins (1974) review of research on the psychology of sex differences encompassed 1,600 works published before 1973. Considering the literature on that topic since 1973, and the fact that many other studies that are not focused specifically on sex differences may contain data on the question, a population of over 5,000 studies on sex differences can be imagined. There are dozens of educational problems on which the research literature is comprised of several hundred articles: ability grouping, reading instruction, programmed learning, instructional television, integration, and so on. Educational research and evaluation is a large and widely scattered enterprise. On problems of importance, it produces literally hundreds of studies in less than five years. The research techniques used, the measurements taken, the types of person studied—all may vary from one st;udy to the next, even though the topic is the same. The research enterprise in education and the social sciences is a rough-hewn, variegated undertaking of huge proportions. Determining what knowledge this enterprise has produced on some question is itself a genuinely important scholarly endeavor. The evolution of style of research integration has been shaped by the size of the research literature. In the 1940s and 1950s, a reviewer contributing to the Review of Educational Research or Psychological Bulletin might find one or two dozen studies on a topic. A narrative, rhetorical integration of so few studies was probably satisfactory. By the late 1960s, the research

Meta-Analysis of Research on Class Size and Achievement

THERE IS No POINT IN RECORDING THE obvious about class size: that teachers worry about it more than nearly anything else, that administrators want to increase it, that it is economically important, and the like. The problem with class size is the research. It is unclear. It has variously been read as supporting larger classes, supporting smaller classes, and supporting nothing but the need for better research. Review after review of the topic has dissolved into cynical despair or epistemological confusion. The notion is wide-spread among educators and researchers that class size bears no relationship to achievement. It is a dead issue in the minds of most instructional researchers. To return to the class-size literature in search of defensible interpretations and conclusions strikes many as fruitless. The endeavor is surrounded by a faint aroma of Chippendale, which it resembles in other respects: unwieldy and antique. One could document the confusion in previous reviews of research on the class-

The effect of psychological intervention on recovery from surgery and heart attacks: an analysis of the literature.

A quantitative review of 34 controlled studies demonstrates that, on the average, surgical or coronary patients who are provided information or emotional support to help them master the medical crisis do better than patients who receive only ordinary care. A review of 13 studies that used hospital days post-surgery or post-heart attack as outcome indicators showed that on the average psychological intervention reduced hospitalization approximately two days below the control groups average of 9.92 days. Most of the interventions were modest and, in most studies, were not matched in any way to the needs of particular patients or their coping styles. Beyond the intrinsic value of offering humane and considerate care, the evidence is that psychological care can be cost-effective.

The Big Picture: A Meta-Analysis of Program Effectiveness Research on English Language Learners:

This article presents a meta-analysis of program effectiveness research on English language learners. The study includes a corpus of 17 studies conducted since Willigs earlier meta-analysis and uses Glass, McGaw, and Smiths strategy of including as many studies as possible in the analysis rather than excluding some on the basis of a priori “study quality” criteria. It is shown that bilingual education is consistently superior to all-English approaches, and that developmental bilingual education programs are superior to transitional bilingual education programs. The meta-analysis of studies controlling for English-language-learner status indicates a positive effect for bilingual education of .23 standard deviations, with outcome measures in the native language showing a positive effect of .86 standard deviations. It is concluded that bilingual education programs are effective in promoting academic achievement, and that sound educational policy should permit and even encourage the development and implementation of bilingual education programs.

The effects of exercise on blood lipids and lipoproteins : a meta-analysis of studies

The results of 66 training studies involving the measurement of human blood lipid and lipoprotein changes over time, conducted over the last 26 yr, and representing 2925 subjects (2086 experimental and 839 control) were collected and statistically aggregated using the meta-analysis technique. Across all types of subjects, treatments, sources, and research designs, the average exercising subject was found to have a reduction in total cholesterol of 10 mg X dl-1 (P les than 0.01), total triglyceride decreased by 15.8 mg X dl-1 (P less than 0.01), DHL-C increased by 1.2 mg X dl-1 (NS), LDL-C decreased by 5.1 mg X dl-1 (P less than 0.05), and total/HDL-C ratio showed a large decrease of 0.48 (P less than 0.01). None of the changes for the control groups were significant. Initial levels of total cholesterol, total triglyceride, HDL-C, and total/HDL-C ratio were strongly correlated with their respective changes as a result of training, regardless of the data partitioning. Higher initial levels of total cholesterol, total triglyceride, and total/HDL-C ratio resulted in greater decreases post-exercise (r = 0.48, 0.76, and 0.75, respectively; P less than 0.01), and lower initial levels of HDL-C resulted in greater post-exercise increases (r = 0.50; P less than 0.01). Overall, physical training seemed to produce beneficial changes in blood lipids and lipoproteins. However, researchers must be careful when examining the relationship between physical training and serum lipids and lipoproteins because initial levels, age, length of training, intensity, VO2max, body weight, and percent body fat have been shown in this meta-analysis to interact with exercise and serum lipid and lipoprotein changes.

Meta-analysis of Research on Class Size and Its Relationship to Attitudes and Instruction

Features of 59 studies of this relationship were coded and quantified and 371 findings were transformed into a common metric for statistical integration. Analysis, based on a logarithmic model, revealed a substantial relationship between class size and teacher and pupil attitudes as well as instruction. Favorable teacher effects (workload, morale, attitudes toward students) are associated with smaller classes as are favorable effects on students (self-concept, interest in school, participation). Smaller classes are associated with greater attempts to individualize instruction and better classroom climate. The results complement those of a previous meta-analysis that showed positive effects of class size on achievement.

Mental health treatment and medical care utilization in a fee-for-service system: outpatient mental health treatment following the onset of a chronic disease.

Charges for medical services of persons covered by the Blue Cross/Blue Shield Federal Employees Program from 1974 through 1978 who were first diagnosed as having one of four chronic diseases in 1975 and within one year began mental health treatment (MHT) were compared with persons who also were first diagnosed as having one of these diseases in 1975 but had no subsequent MHT. In the third year following the diagnosis, those having seven to 20 MHT visits had medical charges

Choice of the Metric for Effect Size in Meta-analysis

309 lower and those having over 21 MHT visits had medical charges

The Effect of Experimental Mortality on the Internal and External Validity of the Randomized Comparative Experiment

284 lower than the comparison group. The savings in medical charges over three years of the group having seven to 20 MHT visits were a function of lower use of inpatient services and roughly equaled the cost of 20 MHT visits. Outpatient mental health treatment can be included in a fee-for-service medical care system to improve the quality and appropriateness of care and, if not extensive, may also serve to lower medical care costs.