Random assignment

Abstract— A computational parallel approach for 3D ray-tracing techniques in the radio propagation prediction is presented. This approach considers that the main tasks in a 3D ray-tracing technique can be evaluated in an independent and/or parallel way. The workload distribution ...

Corporations operate in political, economic, and technological environments that can change rapidly. Competitive advantage can be the reward for those that anticipate, plan, and prepare for change. This exercise demonstrates the essence and value of scenario planning and enables students to experience the process of planning in volatile environments. Teams are formed, an organization is selected, and a list of strategy questions is prepared and narrowed to each team's overriding "Big Question." Students brainstorm a range of critical events in the competitive and general environments. Surprise is injected into the exercise by random assignment of trajectories to critical events that combine to serve as the basis for structuring a scenario. Students summarize their scenarios and prepare a strategy to capitalize on the opportunity or minimize the threat. Class presentations and discussions follow.

Given 2𝑛 points, 𝑛 ``red'' and 𝑛 ``blue'', in a Euclidean space, solving the associated Euclidean Assignment Problem consists in finding the bijection between red and blue points that minimizes a functional of the point positions. In the stochastic version of this problem, the points are a Poisson Point Process, and some interest has developed over the years on the typical and average properties of the solution in the large 𝑛 limit. This PhD thesis investigates this problem in a number of cases (many exact results in 𝑑=1, the derivation of some fine properties in 𝑑=2, in part still conjectural, an investigation on self-similar fractals with intermediate dimensions,...). In particular, we present new contributions on the asymptotic behavior of the cost of the solution, motivated (among others) by the physics of Disordered Systems, and by results in Functional Analysis on the Monge--Kantorovich problem in the continuum.

Background: Hand hygiene is the most effective measure for preventing infections related to healthcare, and its impact on the reduction of these infections is estimated at 50%. Non-compliance has been highlighted in several studies in hospitals, although none have been carried out in primary healthcare. Main objective: To evaluated the effect of a "Hand Hygiene for the reduction of healthcareassociated infections" training program for primary healthcare workers, measured by variation from correct hand hygiene compliance, according to regulatory and specific criteria, 6 months after the baseline, in the intervention group (group receiving a training program) and in the control group (a usual clinical practice). Secondary objectives:-To describe knowledges, attitudes and behaviors as regards hand hygiene among the professionals, and their possible association with "professional burnout", stratifying the results by type of group (intervention and usual clinical practice).-To estimate the logistic regression model that best explains hand hygiene compliance. Methods/Design: Experimental study of parallel groups, with a control group, and random assignment by Health Center. Area of study.-Health centers in northeastern Madrid (Spain). Sample studied.-Healthcare workers (physicians, odontostomatologists, pediatricians, nurses, dental hygienists, midwife and nursing auxiliaries).

Study selection Study designs of evaluations included in the review Studies were included if hormone therapy was used, and if outcome data on at least one quantitative measure of depression or depressed mood was reported. There were no restrictions on the study designs included in the review. Studies utilised placebo control groups, notreatment control groups, treatment versus placebo crossover design, and no control group. The length of treatment ranged from 1 month to 2 years. Specific interventions included in the review HRT. The treatments included oestrogen alone, progesterone alone and in combination with oestrogen, and androgen alone and in combination with oestrogen. The most common oestrogen treatment was premarin, taken orally at a dosage of either 0.625 or 1.25 mg/day. The most common type of progesterone used was norethisterone. Participants included in the review The mean age of the participants was 49.5 years (standard deviation 3.02). Menopausal status: 16 studies (61.5%) utilised only postmenopausal participants, 2 studies (7.7%) used only perimenopausal participants, and 6 studies (23.1%) used both postmenopausal and perimenopausal participants. The remaining 2 studies (7.7%) did not report the menopausal status of their participants.

The aim of this study was to determine the effectiveness of psychodrama on the attachment styles of young adults and to assess the qualitative gains of psychodrama participants. The subjects of the study were 21 master's level students. Data were gathered by using the Experiences in Close Relationships-Revised (ECR-R) form and a personal information form. The study used the mixed method. The quantitative aspect of the study involved the use of pretest-posttest experimental and control group design with random assignment, while the qualitative aspect utilized the methods of case study, observation and interview. A 12-session psychodrama program extending over 3 months was implemented with the experimental subjects, while the control subjects underwent a placebo study. The quantitative findings showed that participating in the group psychodrama program was effective on the anxious attachment style of young adults. However, no meaningful difference existed between the experimental and control groups regarding the anxiety dimension. The qualitative findings, on the other hand, suggested that psychodrama participants improved themselves in understanding the self; developing insight; having awareness of attachment styles; natural and forthcoming in their relationships; having self-confidence; developing listening, empathy and coping skills; and seeing life from a more hopeful perspective.

We explore the problem of distributing a group of indivisible objects, some of which incorporate a primitive assignment of use, or tenancy right. Within this framework we analyze the existence of rules always selecting an ex-ante efficient allocation, conditioned to a preservation of the tenancy rights. We realize that a probabilistic version of the Deferred Acceptance procedure (Gale and Shapley, 1962), is efficient and, from an ordinal point of view, superior to the randomized approaches of the Top Trading Cycles (Abdulkadiroglu and Sonmez, 1999) and the New House 4 mechanism employed in the MIT.

Levels of evidence differ according to the audience addressed. Implementation of universal newborn hearing screening requires responses to a complex myriad of diverse groups: the general public, families with children who are deaf or hard of hearing, the deaf and hard of ...

Social experiments have been used in research since the 1960s, yet the technique of controlled experimentation still arouses controversy among social scientists. The crucial element that distinguishes a controlled experiment from other forms of research is random assignment of treatment to the observational units of study. Because treatment differences in the sample occur as a result of random chance, the effects of the treatment on behavior can be measured with high reliability. This paper examines the advantages and disadvantages of experiments in comparison with other research techniques and describes the circumstances where randomized trials should be preferred over other methods.

ACKNOWLEDGMENTS I am grateful to my advisor and mentor, Geoffrey Smith, for guiding me though my most significant learning experience, showing me the wonders of theoretical computer science, and teaching me about life, integrity, and wisdom. I am also thankful to the members of my dissertation committee Peter Clarke, Dev Roy, and Jinpeng Wei for their many suggestions, advice, and diligence.

This paper uses meta-analysis to investigate whether random assignment (or experimental) evaluations of voluntary government-funded training programs for the disadvantaged have produced different conclusions than nonexperimental evaluations. Information includes several hundred estimates from 31 evaluations of 15 programs that operated between 1964 and 1998. The results suggest that experimental and nonexperimental evaluations yield similar conclusions about the effectiveness of training programs, but that estimates of average effects for youth and possibly men might have been larger in experimental studies. The results also suggest that variation among nonexprimental estimates of program effects is similar to variation among experimental estimates for men and youth, but not for women (for whom it seems to be larger), although small sample sizes make the estimated differences somewhat imprecise for all three groups. The policy implications of the findings are discussed.

The aim of aetiologic studies in epidemiology is to investigate whether factors are causally related to diseases and therefore become a potential target for therapeutic interventions. Mendelian randomization enables estimation of causal relationships in observational studies using genetic variants as instrumental variables. An instrumental variable is a variable that can be considered to mimic the coin toss in a randomized study. Given the random assignment of alleles in gamete formation, the use of genetic variants is an alternative method to control for confounding. This educational article describes the approach of Mendelian randomization, its underlying rationale and its necessary assumptions.

The report was prepared for the Institute of Education Sciences under Contract No. ED-01-CO-0039/0009. The project officer is Elizabeth Warner in the National Center for Education Evaluation and Regional Assistance. IES evaluation reports present objective information on the conditions of implementation and impacts of the programs being evaluated. IES evaluation reports do not include conclusions or recommendations or views with regard to actions policymakers or practitioners should take in light of the findings in the reports.

Quantitative criminology focuses on straightforward causal questions that are ideally addressed with randomized experiments. In practice, however, traditional randomized trials are difficult to implement in the untidy world of criminal justice. Even when randomized trials are implemented, not everyone is treated as intended and some control subjects may obtain experimental services. Treatments may also be more complicated than a simple yes/no coding can capture. This paper argues that the instrumental variables methods (IV) used by economists to solve omitted variables bias problems in observational studies also solve the major statistical problems that arise in imperfect criminological experiments. In general, IV methods estimate causal effects on subjects who comply with a randomly assigned treatment. The use of IV in criminology is illustrated through a re-analysis of the Minneapolis Domestic Violence Experiment.

We consider the assignment problem between two sets of N random points on a smooth, two-dimensional manifold Ω of unit area. It is known that the average cost scales as E Ω (N) ∼ 1 2π ln N with a correction that is at most of order √ ln N ln ln N. In this paper, we show that, within the linearization approximation of the field-theoretical formulation of the problem, the first Ω-dependent correction is on the constant term, and can be exactly computed from the spectrum of the Laplace-Beltrami operator on Ω. We perform the explicit calculation of this constant for various families of surfaces, and compare our predictions with extensive numerics.

Studies have found that naltrexone, a long-acting opiate antagonist, owing to poor patient compliance, is of limited value in preventing relapse. The current study investigates compliance with a 9-month course of naltrexone (25-50 mg daily) given with counseling after ultra-rapid opiate detoxification which uses clonidine and naltrexone under general anesthesia. Eighty-three of 113 randomly selected patients (out of 640), who were detoxified more than 1 year prior (average 1.5 years), responded to phone interviews. Phone questionnaire asked about patients' compliance with naltrexone, counseling and drug use since detoxification. Similar interviews were also conducted with patients' significant other. Non-relapse patients (n= 47, 57%) took naltrexone an average of 2 months longer than did relapse patients (n= 36, 43%). About half of the non-relapse patients completed at least 5 months of naltrexone, 30% completed at least 7 months and about 20% completed 9 months. Fifty-five percent of the relapse patients stopped using naltrexone by the end of the 3rd month, and by the end of 7th month 10% continued to take it. After the first 2 months the decline in naltrexone compliance was about the same for relapse and non-relapse patients. These results are more encouraging about the use of naltrexone for relapse prevention than previous studies. This method of using naltrexone should be further tested in prospective random assignment controlled studies.

In the study, we investigated effects of 2 different versions of a web-based tutoring system to provide 5th-grade students with strategy instruction about text structure, which was an intervention to improve reading comprehension. The design feature assessed varied in individualization of instruction (individualized or standard). The more individually tailored version was developed to provide remediation or enrichment lessons matched to the individual needs of each student. Stratified random assignment was used to compare the effects of 2 versions of the 6-month web-based intervention. Students in the individualized condition made greater improvements from pretest to posttest on a standardized reading comprehension test (d ϭ 0.55) than did students in the standard condition (d ϭ 0.30). Students receiving more individualized instruction demonstrated higher mastery achievement goals when working in the lessons than did students receiving the standard instruction (d ϭ 0.53). Students receiving more individualized instruction showed greater improvement in using signaling, better work in lessons, and more positive posttest attitudes toward computers than did students receiving standard instruction. Students in both conditions improved their recall of ideas from texts and their use of the text structure strategy and comparison signaling words.

While the role of parenting style as a determinant of adolescent outcomes has a long history in developmental psychology and sociology, determining causality remains a challenge. In this study we estimate the impact of parental discipline and parental monitoring on a variety of risky behaviors. We use information on random assignment to parental training programs to identify the impact of alternative child rearing practices. Our results indicate that increases in parental supervision can reduce risky behavior, whereas the establishment of rules or disciplines has no significant impacts. Further, the estimated impacts are found to increase in magnitude as a child ages. Finally, we find that child rearing practices should be treated as endogenous and that rich data on child rearing practices is needed since many parenting strategies occur simultaneously to identify impacts of independent behaviors. We are grateful to CLSRN and SSHRC for research support. We wish to thank participants at the 2007 Canadian Labour Market and Skills Researcher Network Annual conference for numerous suggestions and helpful comments. We also wish to than Yi Chen for providing exceptional research assistance. The usual caveat applies.

Background-Atrial fibrillation (AF) after cardiac surgery is associated with increased risk of complications, length of stay, and cost of care. Observational evidence suggests that patients who have undergone previous statin therapy have a lower incidence of postoperative AF. We tested this observation in a randomized, controlled trial. Methods and Results-Two hundred patients undergoing elective cardiac surgery with cardiopulmonary bypass, without previous statin treatment or history of AF, were enrolled. Patients were randomized to atorvastatin (40 mg/d, nϭ101) or placebo (nϭ99) starting 7 days before operation. The primary end point was incidence of postoperative AF; secondary end points were length of stay, 30-day major adverse cardiac and cerebrovascular events, and postoperative C-reactive protein (CRP) variations. Atorvastatin significantly reduced the incidence of AF versus placebo (35% versus 57%, Pϭ0.003). Accordingly, length of stay was longer in the placebo versus atorvastatin arm (6.9Ϯ1.4 versus 6.3Ϯ1.2 days, Pϭ0.001). Peak CRP levels were lower in patients without AF (Pϭ0.01), irrespective of randomization assignment. Multivariable analysis showed that atorvastatin treatment conferred a 61% reduction in risk of AF (odds ratio 0.39, 95% confidence interval 0.18 to 0.85, Pϭ0.017), whereas high postoperative CRP levels were associated with increased risk (odds ratio 2.0, 95% confidence interval 1.2 to 7.0, Pϭ0.01). The incidence of major adverse cardiac and cerebrovascular events at 30 days was similar in the 2 arms. Conclusions-Treatment with atorvastatin 40 mg/d, initiated 7 days before surgery, significantly reduces the incidence of postoperative AF after elective cardiac surgery with cardiopulmonary bypass and shortens hospital stay. These results may influence practice patterns with regard to adjuvant pharmacological therapy before cardiac surgery. (Circulation. 2006;114:1455-1461.

This article summarizes the results from two studies that used experimental data to evaluate nonexperimental methods of program evaluation. These studies compared the experimental estimates—of an employment and training program that used random assignment to allocate its participants into training positions—to estimates from nonexperimental procedures. Both studies indicate that the nonexperimental methods may not accurately replicate the experimental estimates, and that recently developed methods for constructing comparison groups are no more likely to yield accurate estimates of the impact of training than the more conventional econometric procedures.

The genetic and morphological variation of Pomatomus saltatrix were studied based on morphometric and meristic analyses of samples collected throughout the Black Seas, Marmara, Aegean and eastern Mediterranean Seas. In discriminant function analysis, plotting first and second discriminant functions explained 61 and 77% of the between-group variation for morphometric and meristic analyses, respectively, and indicated existence of three morphologically differentiated groups of P. saltatrix. Consistent differences in the summer and winter samples of the eastern Black Sea, the adjacent sea populations of western Black Sea, Marmara Sea and Aegean Sea, which were overlapping and northeastern Mediterranean samples were observed. The detected pattern of morphological differentiation also reflects their geographic isolation. The overall random assignment of individuals into their original population was moderate for morphometric (54%) and meristic (64%) characters. Both for morphometric and meristic analysis, the highest reclassification rate was observed for the eastern Black Sea and northeastern Mediterranean Sea samples, the most clearly isolated groups. Examination of the contribution of each morphometric and meristic character to the first and second discriminant functions revealed that the observed differences were mainly from head measurements, body depth, lateral scale and vertebrate numbers, indicating these characters to be important in the description of population characteristics. Mantel's test showed that the Mahalanobis distance between the populations both for morphometric (r = −0.282; P = 0.226) and meristic (r = 0.632; P = 0.119) were not significantly associated with their geographical distances.

Research on hindsight bias indicates that awareness of event outcome influences how individuals interpret information and form judgments. We extend this earlier work to suggest that the effect of this bias on lay perceptions of third-party consent to warrantless searchers of residences may be contingent upon the presence versus absence of the search target (the suspect). A study using random assignment to experimental conditions in a between-subjects design explored this possibility. The experiment indicated that hindsight bias in perceived rights of the third-party consentor is influenced, not only by search outcome, but also by a web of overlapping and potentially competing social obligations and personal prerogatives, the salience of which is influenced by situational dynamics.

Because random assignment is not possible in observational studies, estimates of treatment effects might be biased due to selection on observable and unobservable variables. To strengthen causal inference in longitudinal observational studies of multiple treatments, we present 4 latent growth models for propensity score matched groups, and evaluate their performance with a Monte Carlo simulation study. We found that the

Evolutionary spatial 2 × 2 games between heterogeneous agents are analyzed using different variants of cellular automata (CA). Agents play repeatedly against their nearest neighbors 2 × 2 games specified by a rescaled payoff matrix with two parameteres. Each agent is governed by a binary Markovian strategy (BMS) specified by 4 conditional probabilities [p R , p S , p T , p P ] that take values 0 or 1. The initial configuration consists in a random assignment of "strategists" among the 2 4 = 16 possible BMS. The system then evolves within strategy space according to the simple standard rule: each agent copies the strategy of the neighbor who got the highest payoff. Besides on the payoff matrix, the dominant strategy-and the degree of cooperationdepend on i) the type of the neighborhood (von Neumann or Moore); ii) the way the cooperation state is actualized (deterministically or stochastichally); and iii) the amount of noise measured by a parameter ǫ. However a robust winner strategy is [1,0,1,1].

As part of a school-based alcohol misuse prevention study, questionnaires were administered to 2,589 fifth and sixth grade students to determine levels of use of alcohol, marijuana, and cigarettes, intentions to use these substances, and problems resulting from alcohol misuse. The questionnaire also included 45 items concerning susceptibility to peer pressure, self-esteem, and health locus of control. These 45 items were factor analyzed separately for two groups formed by random assignment. Six factors were identified which were both internally consistent and replicable, and indices were constructed. The indices measuring susceptibility to peer pressure, self-esteem, and internal health locus of control were significantly and negatively correlated with most of the substance use, misuse, and intention items, and an external health locus of control index was not significantly related to most of the substance use, misuse, and intention items. The "Susceptibility to Peer Pressure" index correlated more highly with the adolescent substance use, misuse, and intention items than the self-esteem or the health locus of control indices, and it had the highest alpha coefficient. Implications for the design of school-based substance abuse prevention programs are discussed.

We thank the editor and two anonymous referees for helpful comments which greatly improved this draft of the paper. We have benefited from comments received on early drafts of this paper at two brown bag lunches at the Statistics Department, University of Chicago, hosted by Stephen Stigler. We thank all of the workshop participants. In addition, we thank

Social experiments are powerful sources of information about the effectiveness of interventions. In practice, initial randomization plans are almost always compromised. Multiple hypotheses are frequently tested. "Significant" effects are often reported with p-values that do not account for preliminary screening from a large candidate pool of possible effects. This paper develops tools for analyzing data from experiments as they are actually implemented. We apply these tools to analyze the influential HighScope Perry Preschool Program. The Perry program was a social experiment that provided preschool education and home visits to disadvantaged children during their preschool years. It was evaluated by the method of random assignment. Both treatments and controls have been followed from age 3 through age 40. Previous analyses of the Perry data assume that the planned randomization protocol was implemented. In fact, as in many social experiments, the intended randomization protocol was compromised. Accounting for compromised randomization, multiple-hypothesis testing, and small sample sizes, we find statistically significant and economically important program effects for both males and females. We also examine the representativeness of the Perry study.

Eighteen students (K-11th grade) with emotional/behavioral disorders who were at-risk for change of placement to more restrictive settings participated. Construct validity of functional behavioral assessment (FBA) was assessed with a multifunction-multimethod matrix that showed excellent convergent and divergent agreement with combined FBA methods and functional analysis results. Treatment validity was assessed with random assignment to either a FBA with consultation condition (FBC) or to a behavioral consultation without FBA condition (BC). Growth curve analysis showed a significant decrease in inappropriate behavior from baseline to treatment for both treatment groups compared to a control group. The efficiency of teachers' implementation effected treatment validity. Effect sizes showed that regardless of treatment condition that interventions with good fidelity realized the largest effect (d ϭ Ϫ1.14). Social and habilitative validity was assessed with telephone interviews conducted a year after treatment. Sixty-five percent attributed the students' problem behaviors to their thoughts or feelings, whereas only 25% attributed it to classroom variables. Many (57%) indicated the most helpful intervention would be individual counseling, whereas only 14% indicated a more involved classroom management plan would be useful.

In the international system for granting patents for inventions, first patent filings can be followed by subsequent filings at other patent offices within one year. Each such group of related filings constitutes a patent family. Tests are developed as to whether the observed number of first filings that leads to subsequent filings (r) is in agreement with a random process of assignment of the hits from the subsequent filings. An exact expression for the random distribution can be used for small sized data sets. Its behaviour and also the behaviour of an asymptotic Poisson approximation as well as a censored binomial distribution for r are assessed. The approach is stimulated by the Fisher-Wright model in population genetics and possible parallel applications to other biological processes are sought, such as transformations of stem cells and cancer.

We present new evidence on the randomization methods used in existing experiments, and new simulations comparing these methods. We find that many papers do not describe the randomization in detail, implying that better reporting is needed. Our simulations suggest that in samples of 300 or more, the different methods perform similarly. However, for very persistent outcome variables, and in smaller samples, pair-wise matching and stratification perform best and appear to dominate the rerandomization methods commonly used in practice. The simulations also point to specific recommendations for which variables to balance on, and for which controls to include in the ex post analysis. (JEL C83, C93, O12)

This Methodological Brief introduces the reader to the regression discontinuity design (RDD), which is a method that when used correctly can yield estimates of research treatment effects that are equivalent to those obtained through randomized control trials and can therefore be used to infer causality. However, RDD does not require the random assignment of individuals to treatment and control groups, making it very attractive for applied researchers in educational settings. This Brief introduces the method, discusses applications and limitations, and illustrates an idealized example as well as some potential pitfalls and their relevance to the context of gifted education research.

Objective: To generate data on optimal shoulder position comparing two ultrasound-guided extracorporeal shock wave therapy techniques for the treatment of calcifying tendinitis of the shoulder. Design: Random assignment to two groups of treatment with three months follow-up. Setting: The data were collected in outpatients. Subjects: Thirty-five subjects affected by calcifying tendinitis of the shoulder were examined. Interventions: Each subject received three sessions of ultrasound-guided extracorporeal shock wave therapy (performed weekly). Neutral position technique was used in group A (n = 17, mean age 53 ± 9.2 years) and the hyperextended internal rotation technique was used in group B (n = 18, mean age 52.2 ± 10.8 years). Main outcome measures: The Constant and Murley method and radiographs were used to evaluate each subject before the treatment and at three months follow-up. Results: There were no significant differences between changes in Constant total score and pain, activi...

Objective: To describe and compare two extracorporeal shock wave therapy techniques for the treatment of painful subcalcaneal spur. Design: Random assignment to two groups of treatment with two and eight months follow-up. Setting: The data were collected in outpatients. Subjects: Forty-five subjects with a history of at least six months of heel pain were studied. Interventions: Each subject received a three-session ultrasound-guided extracorporeal shock wave therapy (performed weekly). Perpendicular technique was used in group A (n=22, mean age 59.3 ± 12 years) and tangential technique was used in group B (n= 23, mean age 58.8 ± 12.3 years). Main outcome measures: Mayo Clinical Scoring System was used to evaluate each subject before the treatment and at two and eight months follow-up. Results: Mayo Clinical Scoring System pretreatment scores were homogeneous between the groups (group A 55.2 ±18.7; group B 53.5 ± 20; P>0.05). In both groups there was a significant (P<0.05) incr...

Improving response rates to minimize nonresponse bias is often a key goal in survey research. How response rates are increased can determine the remaining nonresponse bias in estimates. Studies often target sample members that are most likely to be interviewed to maximize response rates. Yet the now widely accepted stochastic model for survey participation would suggest that this may not reduce nonresponse bias, further supported in the empirical literature. Counter to current practice, we suggest targeting of likely nonrespondents from the onset of the study with a different protocol, in order to minimize nonresponse bias. To achieve such targeting, various sources of information can be incorporated: paradata collected by the interviewers, demographic and substantive survey data from prior waves, and administrative data. Using these data, sample cases are identified on which a more effective, often more costly, survey protocol can be employed to gain respondent cooperation. This pa...

This article summarizes the results from two studies that used experimental data to evaluate nonexperimental methods of program evaluation. These studies compared the experimental estimates—of an employment and training program that used random assignment to allocate its participants into training positions—to estimates from nonexperimental procedures. Both studies indicate that the nonexperimental methods may not accurately replicate the experimental estimates, and that recently developed methods for constructing comparison groups are no more likely to yield accurate estimates of the impact of training than the more conventional econometric procedures. Hor the last two decades, economists and policymakers havei studied the impact of employment and training programs on the labor market earnings of their participants. These studies have estimated the difference between the participants ’ postprogram earnings and the earnings they would have received in the absence of the ...

This article summarizes the results from two studies that used experimental data to evaluate nonexperimental methods of program evaluation. These studies compared the experimental estimates—of an employment and training program that used random assignment to allocate its participants into training positions—to estimates from nonexperimental procedures. Both studies indicate that the nonexperimental methods may not accurately replicate the experimental estimates, and that recently developed methods for constructing comparison groups are no more likely to yield accurate estimates of the impact of training than the more conventional econometric procedures.

Despite the advances in psychotherapy outcome research, findings are limited because they do not fully generalize to the way therapy is conducted in the real world. Research's clinical validity has been compromised by the medicalization of outcome research, use of random assignment of clients without regard to appropriateness of treatment, fixed number of therapy sessions, nature of the therapy manuals, and use of theoretically pure therapies. The field needs to foster a more productive collaboration between clinician and researcher; study theoretically integrated interventions; use process research findings to improve therapy manuals; make greater use of replicated clinical case studies; focus on less heterogeneous, dimensionalized clinical problems; and find a better way of disseminating research findings to the practicing clinician.

Statistical analysis requires a probability model: commonly, a model for the dependence of outcomes Y on confounders X and a potentially causal variable Z. When the goal of the analysis is to infer Z's eects on Y , this requirement introduces an element of circularity: in order to decide how Z aects Y , the analyst first determines, speculatively, the manner of Y 's dependence on Z and other variables. This paper takes a statistical perspective that avoids such circles, permitting analysis of Z's eects on Y even as the statistician remains entirely agnostic about the conditional distribution of Y given X and Z, or perhaps even denies that such a distribution exists. Our assumptions instead pertain to the conditional distribution Z|X, and the role of speculation in settling them is reduced by the existence of random assignment of Z in a field experiment as well as by poststratification, testing for overt bias before accepting a poststratification, and optimal full matc...

It was hypothesized that node-link mapping. a tool for visually representing client issues during drug abuse counseling, would reduce communication barriers between counselors and their African American and Mexican American clients. Three hundred twenty daily opioid users participated in methadone maintenance treatment programs in three cities for 6 months or longer following random assignments to mapping-enhanced or standard counseling. Clients in the mapping condition had fewer drug-positive urines (with respect to opiates and cocaine). missed fewer scheduled counseling sessions. and were rated more positively by their counselors on rapport, motivation. and self-confidence during the first 6 months of treatment. Significant interactions involving ethnicity and counseling condition suggested that mapping is more effective for African Americans and Mexican Americans than for Whites. The use of mapping appears to help reduce cultural. racial. and clas\ communication barriers by providing a visual supplement and a common language that enhances counselor-client interchanges. There is substantial evidence that mental health services for identifiable ethnic/racial populations need to be improved (Sue & Sue, 1990; Sue & Zane, 1987). Three factors make this particularly critical in the area of drug abuse counseling: There is a disproportionately greater prevalence of drug abuse (especially injection drug use) in certain ethnic/racial populations (Brown. 1993; Brown & Primm. 1988). HIV seroprevalence rates in intravenous drug users appear to be higher in African Americans and Hispanics than in Whites (Allen, Onorato. & Green, 1992). There is evidence that African Americans and Hispanics drop out of some types of drug abuse treatment programs at higher rates than Whites (DeLeon. Melnick.

The purpose of this study was to examine the feasibility and effectiveness of a nutrition education intervention for elementary school children, grades K-6, called the Cookshop Program. It was designed to increase preferences for and consumption of minimally processed whole grains and vegetables through cooking these foods in the classroom along with multiple exposures to the same foods in the cafeteria and parent involvement. It emphasized how eating a plant-based diet was not only important for personal health but was also resource conserving or more ecologically sustainable.This outcome evaluation study compared cooking as an educational strategy with the more usual active participatory educational methods. The study used a quasi-experimental, pre/post intervention-comparison group design with 39 classes (590 students) assigned approximately equally, through matching, to four conditions: cookshops plus food and environment lessons (CS + FEL), cookshops only (CS), food and environment lessons only (FEL), and comparison condition (Com), receiving neither cookshops nor lessons. Data were analyzed using a 2 x 2 factorial design using analysis of covariance. Class was the unit of analysis. Main positive effects were obtained for CS on preferences, knowledge, and plate waste in both younger and older children and on behavioral intention in younger children and cooking self-efficacy in older children. Main positive effects were obtained for FEL in knowledge for both age groups.The results suggest that actual cooking experiences and eating food with peers, accompanied by cognitive learning, may provide a promising approach to nutrition education, especially for younger children. Lack of random assignment of classes to conditions was a limitation of the study. Further research is needed to evaluate this approach using randomized designs.

In this paper, we consider what may be done when researchers anticipate that in the implementation of field experiments, random assignment to experimental and control groups is likely to be flawed. We then reanalyze data from the Minneapolis Spouse Abuse Experiment in a manner that explicitly models violations of random assignment. As anticipated, we find far larger treatment effects than previously reported. The techniques developed should be useful in a wide variety of settings when random assignment is implemented imperfectly.

* We are very grateful to the Research Institute of Economy, Trade and Industry (RIETI) for allowing us to perform the pilot study of the survey on the health and living status of the 50s and beyond. Among the RIETI staffs, we thank to Yuji Hosoya, Yoshifumi Miyamoto, Kozo Oikawa and Yasuo Tanabe at RIETI, especially for Masaru Yoshitomi and Naoto Nagase. We also thank to the other project collaborators on the new economics of aging for their constructive comments; Naohito Abe, Hideki Hashimoto, Katsunori Kondo, Haruko Noguchi and Yasuyuki Sawada. The views expressed in this study are personal and do not represent those of any institutions including RIERI.

Growing concerns about low awareness and take-up rates for government support programs like college financial aid have spurred calls to simplify the application process and enhance visibility. This project examines the effects of two experimental treatments designed to test of the importance of simplification and information using a random assignment research design. H&R Block tax professionals helped low-to moderate-income families complete the FAFSA, the federal application for financial aid. Families were then given an estimate of their eligibility for government aid as well as information about local postsecondary options. A second randomly-chosen group of individuals received only personalized aid eligibility information but did not receive help completing the FAFSA. Comparing the outcomes of participants in the treatment groups to a control group using multiple sources of administrative data, the analysis suggests that individuals who received assistance with the FAFSA and information about aid were substantially more likely to submit the aid application, enroll in college the following fall, and receive more financial aid. These results suggest that simplification and providing information could be effective ways to improve college access. However, only providing aid eligibility information without also giving assistance with the form had no significant effect on FAFSA submission rates.

Social experiments are powerful sources of information about the effectiveness of interventions. In practice, initial randomization plans are almost always compromised. Multiple hypotheses are frequently tested. "Significant" effects are often reported with p-values that do not account for preliminary screening from a large candidate pool of possible effects. This paper develops tools for analyzing data from experiments as they are actually implemented. We apply these tools to analyze the influential HighScope Perry Preschool Program. The Perry program was a social experiment that provided preschool education and home visits to disadvantaged children during their preschool years. It was evaluated by the method of random assignment. Both treatments and controls have been followed from age 3 through age 40. Previous analyses of the Perry data assume that the planned randomization protocol was implemented. In fact, as in many social experiments, the intended randomization protocol was compromised. Accounting for compromised randomization, multiple-hypothesis testing, and small sample sizes, we find statistically significant and economically important program effects for both males and females. We also examine the representativeness of the Perry study.

In this work we prove the bivariate uniqueness property of the Logistic xed-p oint equation, which arise in the study of the random assignment problem, as discussed by Aldous (4). Using this and the general framework of Aldous and Bandyopadhyay (2), we then conclude that the associated recursive tree process is endogenous, and hence the Logistic variables dened in Aldous' work (4) are measurable with respect to the -eld generated by the edge weights. The method involves construction of an explicit recursion to show the uniqueness of the associated integral equation.

This paper evaluates the incidence and extent of support as well as the remuneration of formal and informal caregivers for recipients of a professionally assisted consumer-directed program, a personal budget, compared to recipients of standard types of home care benefits of the German long-term care insurance, i.e. agency services and cash payments. The evaluation makes use of a long-run social experiment at seven different sites in Germany with a random assignment into a treatment group of personal budget recipients and a control group of standard benefit recipients. Despite the experimental design, we have to deal with selection problems stemming from nonrandom assignment and non-random panel attrition. Hence, we apply a difference-indifferences estimator taking account of the non-random nature of panel attrition. The results show that personal budgets could be cost efficient compared to agency services, but are likely to be cost inefficient compared to cash payments.