On the Estimation Accuracy of Causal Effects using Supplementary Variables

This paper focuses on a situation in which a set of treatments is associated with a response through a set of supplementary variables in linear models as well as discrete models. Under the situation, we demonstrate that the causal effect can be estimated more accurately from the set of supplementary variables. In addition, we show that the set of supplementary variables can include selection variables and proxy variables as well. Furthermore, we propose selection criteria for supplementary variables based on the estimation accuracy of causal effects. From graph structures based on our results, we can judge certain situations under which the causal effect can be estimated more accurately by supplementary variables and reliably evaluate the causal effects from observed data.     

Assessment of vocational opportunities and continuing job placement for persons with mental disabilities: factors indicating levels of necessary support

We determined primary factors in allocating support for vocational opportunities and continuing job placement for clients with mental disabilities. The study analyzed 29 items regarding work attitudes and vocational skills for 70 clients as assessed by the Tokyo Metropolitan Rehabilitation Center for the Physically and Mentally Disabled. The evaluations were performed using the situational assessment approach with a 3-point support scale system assigned through observation of behavior. The results of these evaluations placed 35 clients in employment groups (A and B) and 35 clients in a sheltered workshop group (C). The data was analyzed by multivariate analysis. Predictions of success and functional differences among the groups were indicated by the data. Differences were found among the three groups in work attitudes and skills such as "self-direction" and "directions for social participation". Out of the 35 clients currently employed, 11 were recognized as needing ongoing assistance from supported employment offices. For the supported employment group (B), the major work attitudes and skills for which they required continuous services for maintaining stable jobs were 'emotional stability' and 'learning of working skills and accomplishment of work.' In a previous study, the major differences between the competitive employment group (A) and the sheltered workshop group (C) were in areas of 'responsibilities,' 'goals for life,' 'motivation for the work and understanding of contents for work,' and 'safety during work.' Our study supports these findings. In addition, transition from the sheltered workshop to supported employment required training to improve skills and support for 'responsibilities,' 'goals for life,' 'participation in leisure activities and community-based activities,' and 'preparation and cleaning for work.'     

Family System Adjustment and Adaptive Reconstruction of Social Reality among the 1995 Earthquake Survivors

Abstract  The 1995 Great Hanshin-Awaji earthquake resulted in immense imbalances among and within natural ecosystems, the built environment, and human systems. The current study examined the relationship among familial adjustment, adaptive construction of social reality, and recovery of built environment. A random sample mail survey was conducted on 3,300 earthquake victims and 993 questionnaires were returned. The survey questionnaire included the following four scales that measured the family system adjustment on family cohesion and adaptability, the adaptive construction of new reality as evidenced by citizenship orientations, the current level of physical and psychological stress, and a subjective evaluation of life recovery. The results were as follows: (1) Those families that exhibited high cohesion and a clear leadership structure in the emergency period were more functional than others. (2) Those families that reported a balanced level of cohesion and adaptability during the recovery period were the most functional in promoting present individual recovery and in alleviating current stress. (3) The rise of civic-mindedness was observed among those who survived the disaster. (4) Those with high civic-mindedness tend to be better recovered with less current physical and psychological stress. This paper presented a human ecological model that described the relationships among five components: (1) the earthquake hazard, (2) built environment conditions such as disruption of the lifeline and its recovery, (3) opportunity costs for engaging in exchanges with either basic-trust-based ties or social-trust-based ties, (4) the optimal family system adjustment to corresponding exchange relations, and (5) adaptive construction of new social reality.     

Effects of 10-week walking and walking with home-based resistance training on muscle quality,muscle size,and physical functional tests in healthy older individuals

Background

Older individuals have been shown to present muscle atrophy in conjunction with increased fat fraction in some muscles. The proportion of fat and connective tissue within the skeletal muscle can be estimated from axial B-mode ultrasound images using echo intensity (EI). EI was used to calculate the index of muscle quality. Walking, home-based weight-bearing resistance training, and its combinations are considered simple, easy, and practical exercise interventions for older adults. The purpose of this study was to quantify the effects of walking and walking with home-based resistance training on muscle quality of older individuals.

Methods

Thirty-one participants performed walking training only (W-group; 72?±?5?years) and 33 participants performed walking and home-based resistance training (WR-group; 73?±?6?years). This study was a non-randomized controlled trial with no control group. All participants were instructed to walk 2 or 3 sets per week for 10?weeks (one set: 30-min continuous walking). In addition, the WR-group performed home-based weight-bearing resistance training. EI was measured as a muscle quality index using axial B-mode ultrasound images of the rectus femoris and vastus lateralis of the mid-thigh. We further averaged these parameters to obtain the EI of the quadriceps femoris (QF). Participants further performed five functional tests: sit-ups, supine up, sit-to-stand, 5-m maximal walk, and 6-min walk.

Results

QF EI was significantly decreased in both groups after training (W-group 69.9?±?7.4?a.u. to 61.7?±?7.0?a.u., WR-group 64.0?±?9.5?a.u. to 51.1?±?10.0?a.u.; P?<?0.05), suggesting improved muscle quality. QF EI was further decreased in the WR-group compared with the W-group. The sit-up test in both groups and the sit-to-stand and 5-m maximal walk tests in the W-group were significantly improved after training.

Conclusion

These results suggest that training-induced stimulation is associated with a decrease in EI in some thigh regions. Furthermore, the addition of home-based resistance training to walking would be effective for a greater reduction of EI.

     

Finite-sample analysis of impacts of unlabeled data and their labeling mechanisms in linear discriminant analysis

It is widely believed that unlabeled data are promising for improving prediction accuracy in classification problems. Although theoretical studies about when/how unlabeled data are beneficial exist, an actual prediction improvement has not been sufficiently investigated for a finite sample in a systematic manner. We investigate the impact of unlabeled data in linear discriminant analysis and compare the error rates of the classifiers estimated with/without unlabeled data. Our focus is a labeling mechanism that characterizes the probabilistic structure of occurrence of labeled cases. Results imply that an extremely small proportion of unlabeled data has a large effect on the analysis results.     

Enhancing vowel discrimination using constructed spelling

In a computerized task, an adult with intellectual disabilities learned to construct consonant-vowel-consonant words in the presence of corresponding spoken words. During the initial assessment, the participant demonstrated high accuracy on one word group (containing the vowel-consonant units it and un) but low accuracy on the other group (containing the units ag and ed). Errors occurred almost exclusively in the vowel position. Training the group with low accuracy increased accuracy on the trained word group but decreased the accuracy on the untrained, but initially accurate, group. High accuracy on both groups together occurred only after added training in which all five vowels (a, e, i, u, and o) were taught together in words that differed only in the vowel. These findings indicate the importance of carefully arranging examples and nonexamples to sharpen stimulus control. The findings also illustrate a promising step in the development of effective instructional programming for remediation of medial-vowel errors, which can be a source of difficulty in early reading instruction.     

Bayesian analysis for mixtures of discrete distributions with a non-parametric component

Bayesian finite mixture modelling is a flexible parametric modelling approach for classification and density fitting. Many areas of application require distinguishing a signal from a noise component. In practice, it is often difficult to justify a specific distribution for the signal component; therefore, the signal distribution is usually further modelled via a mixture of distributions. However, modelling the signal as a mixture of distributions is computationally non-trivial due to the difficulties in justifying the exact number of components to be used and due to the label switching problem. This paper proposes the use of a non-parametric distribution to model the signal component. We consider the case of discrete data and show how this new methodology leads to more accurate parameter estimation and smaller false non-discovery rate. Moreover, it does not incur the label switching problem. We show an application of the method to data generated by ChIP-sequencing experiments.     

Asymptotic distributions of some test statistics for a linear functional relationship

We consider the testing problems of the structural parameters for the multivariate linear functional relationship model. We treat the likelihood ratio test statistics and the test statistics based on the asymptotic distributions of the maximum likelihood estimators. We derive their asymptotic distributions under each null hypothesis respectively. A simulation study is made to evaluate how we can trust our asymptotic results when the sample size is rather small.     

Outpatient cocaine abuse treatment: predictors of success

In 1990 more than 238,000 patients in the United States entered treatment, with cocaine as their primary diagnosis. This represented 36 percent of the treatment admissions for substance abuse. This research project uses data from the U.S. Army's Alcohol and Drug Abuse Prevention and Control Program to analyze which treatment and demographic variables have an effect on cocaine treatment outcomes. The sample included 3602 soldiers entered into treatment between 1983 and 1992 for cocaine abuse. The logistical regression results suggested the following treatment variables had an effect on outcomes: the type of treatment provided, the length of time in treatment, the number of different types of drugs being abused prior to treatment, and the length of time since the patient had abused cocaine prior to entering treatment. Race of the patient had an impact on treatment outcomes, but gender, educational level, age, rank of the soldier, marital status, and length of time in the military did not have an impact on treatment outcomes.