Dichotomous Indicators Research Articles

An Analysis of Meehl's MAXCOV-HITMAX Procedure for the Case of Dichotomous Indicators

MAXCOV-HITMAX was invented by Paul Meehl for the detection of latent taxonic structures (i.e., structures in which the latent variable, _, is not continuously, but rather Bernoulli, distributed). It involves the examination of the shape of a certain conditional covariance function, and is based on Meehl's claims that: (R1) Given a latent taxonic structure, this conditional covariance function is single peaked; and that (R2), continuous latent structures produce a flat, rather than single-peaked, curve. While Meehl has recommended that continuous indicators be used as input into MAXCOV-HITMAX, the use of dichotomous indicators has become popular. The current work investigates whether (R1) and (R2) are true for the case of dichotomous indicators. The conclusions will be that, for dichotomous indicators: (a) (R1) is not true; (b) (R1) is made true given that there are a large number of indicators; and (c) (R2) is not true, certain unexceptional Rasch structures, for example, producing single-peaked curves. Implications are briefly discussed of these results for the case of MAXCOV-HITMAX with continuous indicators.

Early identification of children at risk for costly mental health service use.

Children and adolescents with serious and persistent conduct problems often require large public expenditures. Successfully diverting one high risk child from unfortunate outcomes may result in a net savings to society of nearly $2 million, not to mention improving the life of that child and his or her family. This figure highlights the potential of prevention, which often rests on the ability to identify these children at a young age. This study examined the ability of a short conduct-problems screening procedure to predict future need for mental health assistance, special education services, and the juvenile justice system during elementary school ages. The screen was based on teacher and parent report of child behavioral habits in kindergarten, and was used to identify children as either at risk or not at risk for behavioral problems. Service outcomes were derived from a service-use assessment administered to parents at the end of the sixth grade, while special education information was gathered through a survey of school records. Study participants (463 kindergarten children; 54% male, 44% African American) were from economically disadvantaged neighborhoods in four diverse communities across the United States. Results indicated that, while controlling for demographic background variables, the risk indicator strongly predicted which children would require services related to conduct disorder or behavioral/emotional problems. Additional analyses revealed that the dichotomous high risk indicator was nearly as strong as the continuous screening variable in predicting the service-use outcomes, and that the screening of both parents and teachers may not be necessary for determining risk status.

THE IMPACT OF PRICES AND CONTROL POLICIES ON CIGARETTE SMOKING AMONG COLLEGE STUDENTS

Smoking among youths and young adults rose throughout the 1990s. Numerous policies were enacted to try to reverse this trend. However, little is known about the impact these policies have on the smoking behavior of young adults. This article uses a dichotomous indicator of daily smoking participation in the past 30 days, an ordered measure representing the frequency of cigarette consumption, and a quasi‐continuous measure of the number of cigarettes smoked per day on average to examine the impact of cigarette prices, clean indoor air laws, and campus‐level smoking policies on the smoking behaviors of a 1997 cross section of college students. The results of the analysis indicate that higher cigarette prices are associated with lower smoking participation and lower levels of use among college student smokers. Local‐ and state‐level clean indoor air restrictions have a cumulative impact on the level of smoking by current smokers. Complete smoking bans on college campuses are associated with lower levels of smoking among current smokers but have no significant impact on smoking participation. Bans on cigarette advertising on campus as well as bans on the sale of cigarettes on campus have no significant effect on the smoking behavior of college students.

The impact of prices and control policies on cigarette smoking among college students

Smoking among youths and young adults has been rising throughout the 1990’s. Numerous policies have been enacted to try to reverse this trend. However, little is known about the impact that these policies have on the smoking behavior of young adults. This study uses a dichotomous indicator of daily smoking participation in the past thirty days, an ordered measure representing the frequency of cigarette consumption and a quasi-continuous measure of the number of cigarettes smoked per day on average to examine the impact of cigarette prices, clean indoor air laws and campus level smoking policies on the smoking behaviors of a 1997 cross section of college students. The results of the analysis indicate that higher cigarette prices are associated with lower smoking participation and lower levels of use among college student smokers. Local and state level clean indoor air restrictions have a cumulative impact on the level of smoking by current smokers. Complete smoking bans on college campuses are associated with lower levels of smoking among current smokers but have no significant impact on smoking participation. Bans on cigarette advertising on campus as well as bans on the sale of cigarettes on campus have no significant effect on the smoking behavior of college student smoking behavior.

Taxometric analysis with dichotomous indicators: the modified MAXCOV procedure and a case-removal consistency test.

Previous Monte Carlo simulations and empirical trials have demonstrated the power with which taxometric procedures distinguish between taxonic (categorical) and dimensional (continuous) latent structures. However, questions have been raised about results obtained using procedures modified to accommodate dichotomous indicators. The present investigation shows that the most widely used taxometric procedure (MAXCOV) can distinguish between latent structures when adapted for use with dichotomous indicators, although particularly large samples and substantially valid indicators are required. This study also illustrates the value of including a case removal consistency test as one component of a taxometric investigation.

Empirical evidence on the demand for carve-outs in employment group mental health coverage.

BACKGROUND AND AIMS OF THE STUDY: The use of specialized behavioral health companies to manage mental/health benefits has become widespread in recent years. Recent studies have reported on the cost and utilization impacts of behavioral health carve-outs. Yet little previous research has examined the factors which lead employer-based health plans to adopt a carve-out strategy for mental health benefits. The examination of these factors is the main focus of our study. Our empirical analysis is also intended to explore several hypotheses (moral hazard, adverse selection, economies of scale and alternate utilization management strategies) that have recently been advanced to explain the popularity of carve-outs. METHODS: The data for this study are from a survey of employers who have long-term disability contracts with one large insurer. The analysis uses data from 248 employers who offer mental health benefits combined with local market information (e.g. health care price proxies, state tax rates etc), state regulations (mental health and substance abuse mandate and parity laws) and employee characteristics. Two different measures of carve-out use were used as dependent variables in the analysis: (1) the fraction of health plans offered by the employer that contained carve-out provisions and (2) a dichotomous indicator for those employers who included a carve-out arrangement in all the health plans they offered. RESULTS: Our results tended to support the general cost-control hypothesis that factors associated with higher use and/or costs of mental health services increase the demand for carve-outs. Our results gave less consistent support to the argument that carve-outs are demanded to control adverse selection, though only a few variables provided a direct test of this hypothesis. The role of economies of scale (i.e., group size) and the effectiveness of alternative strategies for managing moral hazard costs (i.e., HMOs) were confirmed by our results. DISCUSSION: We considered a number of different hypotheses concerning employers' demands for mental health carve-outs and found varying degrees of support for these hypotheses in our data. Our results tended to support the general cost-control hypothesis that factors associated with higher use and/or costs of mental health services increase the demand for carve-outs. LIMITATIONS: Our database includes a small number of relatively large employers and is not representative of employers nationally. Our selection criteria, concerning size and the requirement that some employees are covered by LTD insurance, probably resulted in a study sample that offers richer benefits than do employers nationally. Our employers also report a higher percentage of salaried employees relative to the national data. Another deficiency in the current study is the lack of detailed information on the socio-demographic and behavioral characteristics of covered employees. Finally, the cross-sectional nature of our analysis raises concerns about susceptibility of our findings to omitted variables bias. IMPLICATIONS FOR FURTHER RESEARCH: Research with more information on covered employee characteristics will allow for a stronger test of the general hypothesis that factors associated with a higher demand for services are also associated with a higher demand for carve-outs. Also, future analyses that capture the experience of states that have recently passed mandate and parity laws, and that use pooled data to control for omitted variables bias, will provide more definitive evidence on the relationship between these laws and carve-out demand.

Thurstonian modeling of ranking data via mean and covariance structure analysis

Although Thurstonian models provide an attractive representation of choice behavior, they have not been extensively used in ranking applications since only recently efficient estimation methods for these models have been developed. These, however, require the use of special-purpose estimation programs, which limits their applicability. Here we introduce a formulation of Thurstonian ranking models that turns an idiosyncratic estimation problem into an estimation problem involving mean and covariance structures with dichotomous indicators. Well-known standard solutions for the latter can be readily applied to this specific problem, and as a result any Thurstonian model for ranking data can be fitted using existing general purpose software for mean and covariance structure analysis. Although the most popular programs for covariance structure analysis (e.g., LISREL and EQS) cannot be presently used to estimate Thurstonian ranking models, other programs such as MECOSA already exist that can be straightforwardly used to estimate these models.

Statistical Issues Encountered in the Comparison of Health-Related Quality of Life in Diseased Patients to Published General Population Norms: Problems and Solutions

The objectives of this study were (1) to illustrate the statistical problems encountered when comparing health-related quality of life (HRQL) measured by the Medical Outcome Study Short Form-36 (SF-36) in a diseased group to general population norms, and (2) to define age- and gender-standardized dichotomous indicator variables for each health concept and show that these indicator variables facilitate comparisons between the diseased sample and the general population. Our “diseased” group consisted of 136 sequentially consenting patients referred to the syncope clinic for assessment and treatment. Participants completed the SF-36 questionnaire before undergoing diagnostic testing. General population norms for the SF-36 are available from the responses of 2474 participants in the National Survey of Functional Health Status, conducted in 1990 in the United States. Comparison of the SF-36 in a diseased sample with general population norms is difficult, owing to skewed and unusual distributions in both groups. In addition, making comparisons within age and gender strata is difficult if the within strata sample size is small. We propose a dichotomous indicator variable for each health concept that classifies an individual as having impaired health if he or she scored lower than the 25th percentile for the appropriate age and gender general population strata. By definition, the prevalence of impaired health in the general population is 25% for all eight health concepts. Comparison between the eight health-concept variables is easy because the population norm is the same for each of them. These indicator variables are age and gender adjusted, so that even if the sample did not have the age and gender distribution as the general population, comparisons can still be made with the value of 25.

The Price of Alcohol, Wife Abuse, and Husband Abuse

Alcohol consumption has been frequently linked to family violence. The purpose of this paper is to examine the direct relationship between the price of alcohol, which determines consumption, and violence towards spouses. The data come from the 1985 cross section and the 1985-1987 panel of the National Family Violence Survey. The 1985 data are a nationally representative sample while the panel oversamples violent individuals. Dichotomous indicators of severe violence towards wives and husbands are used. A reduced for violence equation is estimated, and individual-level fixed effects are used to control for unobserved characteristics in the panel. A consistent result that emerges from this paper is that an increase in the pure price of alcohol, as measured by a weighted average of the price of alcohol from beer, wine, and liquor, will serve to reduce severe violence aimed at wives. By contrast, the evidence on the propensity of an increase in the price of alcohol to lower violence towards husbands is mixed. When individual level characteristics are not controlled for, the price is not a predictor of violence towards men. However, once the individual traits are controlled for, a negative relationship between the price and violence emerges. Published: Southern Economic Journal, 67, No.2 (October 2000), 279-303.

Issues for covariance analysis of dichotomous and ordered categorical data from randomized clinical trials and non-parametric strategies for addressing them.

Analysis of covariance is an effective method for addressing two considerations for randomized clinical trials. One is reduction of variance for estimates of treatment effects and thereby the production of narrower confidence intervals and more powerful statistical tests. The other is the clarification of the magnitude of treatment effects through adjustment of corresponding estimates for any random imbalances between the treatment groups with respect to the covariables. The statistical basis of covariance analysis can be either non-parametric, with reliance only on the randomization in the study design, or parametric through a statistical model for a postulated sampling process. For non-parametric methods, there are no formal assumptions for how a response variable is related to the covariables, but strong correlation between response and covariables is necessary for variance reduction. Computations for these methods are straightforward through the application of weighted least squares to fit linear models to the differences between treatment groups for the means of the response variable and the covariables jointly with a specification that has null values for the differences that correspond to the covariables. Moreover, such analysis is similarly applicable to dichotomous indicators, ranks or integers for ordered categories, and continuous measurements. Since non-parametric covariance analysis can have many forms, the ones which are planned for a clinical trial need careful specification in its protocol. A limitation of non-parametric analysis is that it does not directly address the magnitude of treatment effects within subgroups based on the covariables or the homogeneity of such effects. For this purpose, a statistical model is needed. When the response criterion is dichotomous or has ordered categories, such a model may have a non-linear nature which determines how covariance adjustment modifies results for treatment effects. Insight concerning such modifications can be gained through their evaluation relative to non-parametric counterparts. Such evaluation usually indicates that alternative ways to compare treatments for a response criterion with adjustment for a set of covariables mutually support the same conclusion about the strength of treatment effects. This robustness is noteworthy since the alternative methods for covariance analysis have substantially different rationales and assumptions. Since findings can differ in important ways across alternative choices for covariables (as opposed to methods for covariance adjustment), the critical consideration for studies with covariance analyses planned as the primary method for comparing treatments is the specification of the covariables in the protocol (or in an amendment or formal plan prior to any unmasking of the study.

Issues for covariance analysis of dichotomous and ordered categorical data from randomized clinical trials and non‐parametric strategies for addressing them

Analysis of covariance is an effective method for addressing two considerations for randomized clinical trials. One is reduction of variance for estimates of treatment effects and thereby the production of narrower confidence intervals and more powerful statistical tests. The other is the clarification of the magnitude of treatment effects through adjustment of corresponding estimates for any random imbalances between the treatment groups with respect to the covariables. The statistical basis of covariance analysis can be either non-parametric, with reliance only on the randomization in the study design, or parametric through a statistical model for a postulated sampling process. For non-parametric methods, there are no formal assumptions for how a response variable is related to the covariables, but strong correlation between response and covariables is necessary for variance reduction. Computations for these methods are straightforward through the application of weighted least squares to fit linear models to the differences between treatment groups for the means of the response variable and the covariables jointly with a specification that has null values for the differences that correspond to the covariables. Moreover, such analysis is similarly applicable to dichotomous indicators, ranks or integers for ordered categories, and continuous measurements. Since non-parametric covariance analysis can have many forms, the ones which are planned for a clinical trial need careful specification in its protocol. A limitation of non-parametric analysis is that it does not directly address the magnitude of treatment effects within subgroups based on the covariables or the homogeneity of such effects. For this purpose, a statistical model is needed. When the response criterion is dichotomous or has ordered categories, such a model may have a non-linear nature which determines how covariance adjustment modifies results for treatment effects. Insight concerning such modifications can be gained through their evaluation relative to non-parametric counterparts. Such evaluation usually indicates that alternative ways to compare treatments for a response criterion with adjustment for a set of covariables mutually support the same conclusion about the strength of treatment effects. This robustness is noteworthy since the alternative methods for covariance analysis have substantially different rationales and assumptions. Since findings can differ in important ways across alternative choices for covariables (as opposed to methods for covariance adjustment), the critical consideration for studies with covariance analyses planned as the primary method for comparing treatments is the specification of the covariables in the protocol (or in an amendment or formal plan prior to any unmasking of the study). © 1998 John Wiley & Sons, Ltd.

Associations between Lung Function and Estimated Average Exposure to NO2 in Eight Areas of Switzerland

A common strategy for modeling dose-response in epidemiology is to transform ordered exposures and covariates into sets of dichotomous indicator variables (that is, to factor the variables). Factoring tends to increase estimation variance, but it also tends to decrease bias and thus may increase or decrease total accuracy. We conducted a simulation study to examine the impact of factoring on the accuracy of rate estimation. Factored and unfactored Poisson regression models were fit to follow-up study datasets that were randomly generated from 37,500 population model forms that ranged from subadditive to supramultiplicative. In the situations we examined, factoring sometimes substantially improved accuracy relative to fitting the corresponding unfactored model, sometimes substantially decreased accuracy, and sometimes made little difference. The difference in accuracy between factored and unfactored models depended in a complicated fashion on the difference between the true and fitted model forms, the strength of exposure and covariate effects in the population, and the study size. It may be difficult in practice to predict when factoring is increasing or decreasing accuracy. We recommend, therefore, that the strategy of factoring variables be supplemented with other strategies for modeling dose-response.

Health expectancy by level of education in Finland

Using the method first presented by Sullivan, the article presents results on health expectancy by level of education and gender in the late 1980s in Finland. The life tables by level of education cover the years 1986–1990. Indicators of disability and poor health were based on three variables from the nationwide 1986 Survey on Living Conditions ( N = 12,057): limiting long-standing illness, functional disability and poor self-perceived health. Two cutting points indicating different levels of severity of disability or poor health were used for each measure, giving six dichotomous indicators. Disability-free life expectancy and life expectancy with disability were found to depend strongly on the indicator of disability, but the patterns of differences both between genders and between educational categories were largely independent of the indicators used. Life expectancy as well as disability-free life expectancy showed a systematic relationship with level of education: the higher the level of education, the higher the life expectancy and disability-free life expectancy. The differences between educational categories in disability-free life expectancy were markedly larger than in total life expectancy. Life expectancy with disability was shortest among the more educated and longest among the less educated. Due to the higher life expectancy and the higher prevalence of disability among women, life expectancy with disability was longer among women than men according to all indicators.

Comparison of periodontal disease in HIV seropositive subjects and controls (I).

The aims of this study were to investigate the prevalence and severity of periodontal diseases in HIV seropositive and assumed HIV seronegative individuals in the Edinburgh area. 29 HIV seropositive subjects were examined at baseline and at 3-monthly intervals. Attachment loss and dichotomous indicators of periodontal disease were recorded. 27 control subjects were seen at baseline only. Although there was no significant difference between the 2 groups in terms of mean attachment loss at baseline, HIV seropositive subjects had experienced more severe attachment loss localised to the lower incisor region. HIV seropositive subjects had a significantly higher mean % of sites exhibiting suppuration and redness. A distinct subgroup of 9 HIV seropositive subjects with widespread attachment loss was detected, which included those subjects with severely affected lower incisors. Site-specific attachment loss of > or = 3 mm was observed in only 20 sites out of 2814 sites in HIV+subjects observed longitudinally. The incidence of aggressive periodontal disease in this study was lower than anticipated, despite the inclusion of patients with advanced HIV infection or AIDS.

Measures of socioeconomic status for child health research: Comparative results from Bangladesh and Pakistan

This paper examines the reliability and validity of several hypothesized indicators of socioeconomic status for use in epidemiologic research, particularly in studies of child health in the less developed world. Population-based surveys of child health and disability were completed in Bangladesh and Pakistan using standard questionnaires designed to measure four domains of household socioeconomic status: wealth, housing, parental education and occupation. Test-retest data indicate moderate to excellent reliability of most of the socioeconomic indicators in both countries. Loadings from factor analyses of the survey data provide further evidence of the reliability of the data, and confirm that the questionnaire measures housing and wealth as distinct domains in both countries. Parental education and occupation are correlated with housing and/or wealth in these data sets. Bivariate logistic regression analyses show that, although 11 of 12 dichotomous indicators of low socioeconomic status constructed from the data are predictive of child death in at least one of the four sub-populations studied (rural and urban Bangladesh, and rural and urban areas of Karachi, Pakistan), no single indicator is predictive of child death in all four sub-populations. These along with multivariate results demonstrate the importance of including multiple measures of distinct domains if the research aims include investigation and/or control of the effects of socioeconomic status on health in diverse populations.

A structural latent variable approach to modelling consumer perception: A case study of orange juice

Recognizing the unobservability of consumer perception, a structural latent variable model with multiple dichotomous indicators and multiple cause variables was employed to characterize consumer perception of orange juice. Results indicated that education, sex, race, urbanization, and household size were important determinants of consumer perception. Other results suggested that several of the included indicator variables (such as orange juice is good for breakfast, is good for health, is energizing, and is refreshing) represented important dimensions of the latent consumer perception variable. © 1993 John Wiley & Sons, Inc.

An application of a multiple cause variable model for consumer perception of orange juice

A multiple indicator and multiple cause model with dichotomous indicators was used to study consumers' perceptions of orange juice. The results indicate that recalls of orange juice advertising messages by the respondent had a positive impact on their perception of orange juice. The results also suggest that selected socio-economic variables were important determinants of consumer perception.

Health effects of military service lessons learned from the Vietnam experience

This article examines the methodologic difficulties encountered in studies of the long-term effects of the Vietnam War on the psychological and physical health of veterans. Alternate study designs and exposure and outcome measurements are examined in relation to psychological (post-traumatic stress disorder and alcohol use or abuse) and physical (mortality and birth defects in children) health outcomes. All major epidemiologic studies of post-traumatic stress disorder and alcohol use or abuse utilize cross-sectional research designs. Mortality and birth defect studies use cohort, case-control, or proportionate mortality strategies. Exposure is measured using either definitive dichotomous indicators of service in Southeast Asia or more complex indicators of the Vietnam experience (i.e., combat, herbicide exposure, atrocities, abusive violence) that are of suspect validity. In studies of psychological health, outcomes are based exclusively on self-reported symptoms, while investigations of mortality and birth defects use death certificates and hospital records. Epidemiologic research on the effects of the Vietnam conflict has been hampered by problems in research design and the inherent difficulties of measuring wartime exposures and long-term health outcomes.

Democracy, Stability, and Dichotomies

An early debate in the empirical study of political democracy concerned the measurement of democracy. Initial work employed dichotomous indicators and incorporated stability into political democracy measures. Evidence accumulated showing that this approach could adversely affect analyses, particularly in the study of income inequality. Despite this, some recent studies have renewed the flawed practices. We draw upon recent work to highlight the confusion that can result when democracy and stability are confounded and the problems with dichotomous indicators. We propose that stability is analytically distinct from political democracy and should be treated as such empirically, and we suggest ways to estimate incremental effects of political democracy. We further argue that political democracy is continuous and that measures of it should reflect this. It is important that the measurement history of this construct not repeat itself.

Prosthetic status: The formation of a schedule

I n England, and indeed worldwide, increasing attention is being focused on the dental needs of elderly persons. Journals are now including sections on oral gerontology, and surveys concerning the dental health and needs of elderly persons are becoming ever more frequent. Oral examination of elderly people is complex: not only may it involve examination for caries, periodontal diseases, oral mucosal lesions, and oral pathosis but also assessment of the patient’s dentures. In developed countries the proportion of elderly persons who are edentulous is high (Table I).le9 Yet, there is no internationally agreed method for examination of a denture during oral health surveys as has been established for caries and periodontal disease. Prosthetic examination of elderly populations has been undertaken by Neil1 and Phillips,‘O Grabowski and Bertram, Ritchie,” and by Davidson,‘* who used data for their examinations based on work by Kapur13 or Bergman et a1.14 Kapur,13 in a clinical evaluation of two denture adhesives, distinguished between various ridge forms and subdivided retention into four grades and stability into three grades. Although Kapur13 states, “The maxillary and mandibular dentures were clinically checked for their denture base fit and occlusal relationship,” and “The two dentures were further examined for their occlusal vertical and occlusal horizontal relationships,” he does not use these criteria in his scoring system. Bergman et al. ” described their method of assessing occlusion and articulation, but made no mention of retention or vertical height. Sprigg” described a method of denture analysis but did not give details of how he made the evaluation. RiseI used six indicators to assess denture conditi.on: defects of dentures, denture material, stability, retent.ion, occlusion, and denture related mucosal lesions. He found that measurement of differing degrees of complexity for each indicator, to produce a weighted score, only made the examination more complex to perform. He concluded that “examiners (using dichotomous indicators) will then only have to judge the presence or absence of unsatisfactory dentures and not to