The visibility of illicit drugs: implications for community-based drug control strategies.

OBJECTIVES: This study examined differences between the visibility of drugs and drug use in more than 2100 neighborhoods, challenging an assumption about drug use in poor, minority, and urban communities. METHODS: A telephone survey assessed substance use and attitudes across 41 communities in an evaluation of a national community-based demand reduction program. Three waves of data were collected from more than 42 000 respondents. RESULTS: Measures of neighborhood disadvantage, population density, and proportion of minority residents explained more than 57% of the variance between census tracts in visibility of drug sales but less than 10% of tract-to-tract variance in drug use. Visible drug sales were 6.3 times more likely to be reported in the most disadvantaged neighborhoods than in the least disadvantaged, while illicit drug use was only 1.3 times more likely. CONCLUSIONS: The most disadvantaged neighborhoods have the most visible drug problems, but drug use is nearly equally distributed across all communities. Thus, efforts to address drug-related problems in poorer areas need to take into account the broader drug market served by these neighborhoods.

Concurrence of drug users' self-report of current HIV status and serotest results.

This study examines the concurrence of drug users' self-reports of current HIV status with serotest results. The analyses are based on data obtained from face-to-face interviews with 7,256 out-of-treatment injection drug and/or crack users in 10 sites that participated in the Cooperative Agreement for AIDS Community-Based Outreach/Intervention Research Program (funded by the U.S. National Institute on Drug Abuse ). Although the degree of concurrence between HIV-negative individuals' self-reports of their current HIV status and their serotest results was high (specificity, 99%), this was not the situation for individuals who tested positive for HIV (sensitivity, 44%).

Modeling Relationships Between Two Categorical Variables When Data Are Missing: Examining Consequences of the Missing Data Mechanism in an HIV Data Set.

Analysts evaluating the strengths of relationships between variables in behavioral science research must often contend with the problem of missing data. Analyses are typically performed using data for cases that are either complete in all the variables, or assume that the data are missing at random. Often, these approaches yield biased results. Using empirical data, the current work explores the implications and consequences of using various statistical models to describe the association of two variables, one ordinal and one dichotomous, in which data are incomplete for the dichotomous variable. These models explicitly reflect the missing data mechanism; models that hypothesize nonignorable nonresponse are given particular attention. Both the statistical fit and substantive consequences of these models are examined. This new methodological approach to examining nonignorable nonresponse can be applied to many behavioral science data sets containing an ordinal variable.

Survey estimates of drug-use trends in urban communities: general principles and cautionary examples.

Surveys to depict substance abuse rates and monitor trends in specific areas have become increasingly important policy tools. Yet, as illustrated by two national multiwave surveys, using small sample survey data and making longitudinal comparisons is fraught with interpretative problems. In the case of the metropolitan area "oversample" of the National Household Survey on Drug Abuse, for example, interpreting apparent declines in drug use has to take account of the devastating effects of Hurricane Andrew in the Miami Metropolitan area. In the case of a 41-community survey sponsored by the Robert Wood Johnson Foundation to evaluate substance abuse prevention, the difficulty is how to interpret small differences in drug use, which seem to follow no reasonable pattern with respect to treatment or comparison sites. Inferences from such surveys are confounded with statistical anomalies and unforeseen events. They are limited by the sample size. In part, the solution to these problems is to use other survey and nonsurvey data to validate their conclusions and to note their limitations.

Zero effects in substance abuse programs. Avoiding false positives and false negatives in the evaluation of community-based programs.

The evaluation of community-based programs poses special design and analysis problems. The present article focuses on two major types of errors that can occur in such evaluations: false positive--incorrectly declaring a program to be effective--and false negatives--incorrectly declaring a program to be ineffective. The evaluation of a national demonstration of community-based programs to reduce substance abuse, Fighting Back, is used to illustrate several approaches to reduce the probability of errors. Both those errors that are affected by the design and those by analytic approaches are considered. Ways to assess multiple outcomes and to match the complexity of the program with design and analytic strategies are proposed. Community trials are complex interventions, and, although they can provide very useful information, their outcomes have to be understood in terms of the constructs they test and the contexts within which they are carried out.

Some Theory and Applications of Confirmatory Second-Order Factor Analysis.

This paper shows how confirmatory factor analysis can be used to test second- (and higher-) order factor models in the areas of the structure of abilities, allometry, and the separation of specific and error variance estimates. In the latter area, an idea of Joreskog's is extended to include a new conceptualization of the estimation of validity. Second-order models are placed in a hierarchy of factor analysis models, which shows how the fit of various models can be compared. The concept of discriminability is introduced to describe a situation in which two models may both be identified, and yet the goodness-of-fit for both will be the same. This problem can usually be avoided by careful design of a study. Several examples are discussed.

The value of latent class analysis in medical diagnosis.

Assessment of the value of diagnostic indicators such as symptoms and laboratory tests results from calculation of the sensitivity and specificity of the indicators. Knowledge of the rate of occurrence of the disease allows for additional calculations of the error rates in using an indicator. These calculations are accurate only when the data on which they are based are reliable. If the diagnosis, which is used as the criterion for computing the sensitivity and specificity, is not accurate, then the resulting calculations will be in error. We show how a statistical method, latent class analysis, allows for the estimation of the characteristics of indicators even when an accurate diagnosis is unavailable. In addition, the method deals with several indicators at once, and provides a way to combine the information from all the indicators to make a diagnosis.