[Show abstract][Hide abstract] ABSTRACT: IntroductionEvery statistical model has its own underlying “assumptions” that must be verified to validate the results. In some situations, violations of these assumptions will not change substantive research conclusions, while in others, violation of assumptions can be critical to meaningful research. For a meaningful and conclusive data analysis by Analysis of Variance (ANOVA), the following assumptions are needed:(a)Errors be normally distributed (b)Errors have same variances (homogeneity of variances) (c)Errors be independently distributed However, the question arising is What would be the effects of any departure from the assumptions of the model on the inferences made? The answer is simple: It may either influence the probability of making Type I error (i.e., incorrectly rejecting null hypothesis) or a Type II error (i.e., failing to reject a null hypothesis when it is false). For a thorough discussion of the topic, the reader is referred to Scheffé (1959), Miller (1986) ...
[Show abstract][Hide abstract] ABSTRACT: Students in statistics service courses are frequently exposed to dogmatic approaches for evaluating the role of randomization in statistical designs, and inferential data analysis in experimental, observational and survey studies. In order to provide an overview for understanding the inference process, in this work some key statistical concepts in probabilistic and nonprobabilistic sampling are discussed. The statistical model constituting the basis of statistical inference is postulated and a brief review of the finite population descriptive inference and a quota sampling inferential theory are provided. Some comments on distinct approaches for conducting inferences in probabilistic and nonprobabilistic samples are adduced.
No preview · Article · Nov 2010 · International Journal of Mathematical Education
[Show abstract][Hide abstract] ABSTRACT: We propose a mixed-effects linear model for analyzing growth curves data obtained using a two-way classification experiment. The model combines an unconstrained means model and a regression model on the time, in which the coefficients are con-sidered random. The model allows for experimental unit covariates so as to study the trend and the variability of the individual growth curves. Comments on data analysis strategies are provided. An application of the model is illustrated using a data-set comes from a chrysanthemum growth experiment. Resumen Proponemos un modelo lineal de efectos mixtos para analizar datos de curvas de crecimiento de un experimento con dos criterios de clasificación. El modelo combina un modelo no restringido de medias y un modelo de regresión sobre el tiempo, en el cual los coeficientes son considerados aleatorios. El modelo considera covariables a nivel de la unidad experimental para estudiar la tendendia y la variabilidad de las curvas de crecimiento. Se proporcionan comentarios sobre estrategias de analisis de datos. Se ilustra la aplicación del modelo usando un conjunto de datos de un experimento de crecimiento de crisantemos. Palabras clave: Modelos de regresión lineal multinivel, modelos de coeficientes aleatorios, modelos de medias, estrategias de analisis de datos.
[Show abstract][Hide abstract] ABSTRACT: In this paper we review the main proposed approximations to percentiles of the noncentral t-distribution. The approximations are examined for their acuracy over a wide range of values of the parameters of the distribution and for several percentil values. Tables summarizing the approximations are included.
[Show abstract][Hide abstract] ABSTRACT: In this work simultaneous confidence intervals for the variance components in the two-way balanced crossed random effects model with interaction have been derived under the usual assumptions of normality and independence of random effects. The intervals are conservative in the sense that the true confidence coefficient is as large as preassigned value. The formulas are illustrated using published data with SAS outputs.
Full-text · Article · Jan 2009 · Investigacion Operacional
[Show abstract][Hide abstract] ABSTRACT: Drug users have been found to be at high risk of mortality but the mortality experience of Hispanic drug users remains understudied. This study assessed mortality among Puerto Rican injection drug users (IDUs) in New York City (NY), and in Puerto Rico (PR). Study subjects were 637 IDUs from NY and 319 IDUs from PR. Mortality was ascertained using data from the National Death Index. Annual mortality rate of the NY cohort was 1.3 per 100 person years compared to the PR cohort with a rate of 4.8. Compared to the Hispanic population of New York City, the standardized mortality ratio (SMR) of the NY cohort was 4.4. Compared to the population of Puerto Rico, the SMR of the PR cohort was 16.2. The four principal causes of death were: NY-HIV/AIDS (50.0%), drug overdoses (13.3%), cardiovascular conditions (13.3%), and pulmonary conditions (10.0%); PR-HIV/AIDS (37.0%), drug overdoses (24.1%), sepsis (13.0%), and homicide (11.1%). Modeling time to death using Cox proportional hazards regression, the relative risk of mortality of the PR cohort as compared to the NY cohort was 9.2. The other covariates found to be significantly associated with time to death were age, gender, education, social isolation, intoxication with alcohol, and HIV seropositivity. The large disparity in mortality rates found in this study suggests that health disparities research should be expanded to identify intra-group disparities. Furthermore, these results point to an urgent need to reduce excess mortality among IDUs in Puerto Rico.
Preview · Article · Dec 2006 · Journal of Urban Health
[Show abstract][Hide abstract] ABSTRACT: This paper reports findings on 334 out-of-treatment drug users in Puerto Rico and 617 in New York City, at the 6-month follow-up interview of a Longitudinal Survey. Main outcomes were health care and drug treatment utilization since baseline, assessed by asking participants if they had received physical or mental health services (including HIV medications), and if they had been in methadone maintenance, inpatient or outpatient drug treatment, or drug treatment while incarcerated. Chi-square tests were used to evaluate associations between gender and the various correlates. Logistic regression was used to calculate the contribution of each variable in predicting use of drug treatment. The analysis suggests that women in both sites were likely to suffer from disparities in both health care and drug treatment utilization when compared with men, albeit women in New York utilized more drug treatment resources and were more embedded in the immediate family than their female peers in Puerto Rico. Further research to specify the impact of contextual factors at the organizational and community levels, among members of the same ethnic group residing in different sites, may prove valuable in identifying the health needs and the factors that impede or facilitate drug-using women in obtaining the most appropriate treatment. Findings from these studies can help in developing appropriate public health policy and science-based drug treatment programs to eliminate disparities such as the ones detected in this study.
[Show abstract][Hide abstract] ABSTRACT: This study examined factors associated with drug treatment dropout among injection drug users (IDUs) in Puerto Rico, a group that has contributed significantly to the self-sustaining AIDS epidemic in the island since the mid-1980s. A total of 557 IDUs were recruited from communities in a semirural region of Puerto Rico, as part of a longitudinal study testing the efficacy of a two-facet intervention model, based on motivational interviewing. Of 124 IDUs who had entered drug treatment at follow-up, 33 (26.6%) dropped out before completing all recommended sessions. Multiple logistic regression analysis showed that age, homelessness, and speedball use were significantly associated with drug treatment dropout. Conversely, participants who received the two-facet intervention were significantly less likely to drop out of drug treatment. Receiving psychiatric services also reduced the odds of treatment dropout. Improving adherence to drug treatment and reducing dropout rates are complex processes that need to be addressed at the individual behavioral and social support levels, as well as the program process and resource levels.
Full-text · Article · Mar 2005 · Addictive Behaviors
[Show abstract][Hide abstract] ABSTRACT: This paper reports results of an analysis of the association between alcohol intoxication and injection and sexual HIV risk behaviors among 557 Hispanic heroin and cocaine injectors, not in treatment, who were recruited in poor communities in Puerto Rico. Subjects were part of a longitudinal prevention-intervention study aimed at reducing drug use and HIV risk behaviors. Participants reported a high prevalence of co-occurring conditions, particularly symptoms of severe depression (52%) and severe anxiety (37%), measured by Beck's Depression Index and Beck's Anxiety Index, respectively. Alcohol intoxication during the last 30 days was reported by 18% of participants. Associations were found between alcohol intoxication and both injection and sexual risk behaviors. In the bivariate analysis, subjects reporting alcohol intoxication were more likely to inject three or more times per day, pool money to buy drugs, share needles, and share cotton. They were also significantly more likely to have a casual or paying sex partner and to have unprotected sex with these partners. After adjustment, sharing needles and cotton, having sex with a paying partner or casual partner, and exchanging sex for money or drugs were significantly related to alcohol intoxication. HIV prevention programs, to be effective, must address alcohol intoxication and its relation to injection and sexual risk behaviors as a central issue in HIV prevention among drug injectors.
No preview · Article · Jan 2005 · Drug and Alcohol Dependence
[Show abstract][Hide abstract] ABSTRACT: Analysis of variance (ANOVA) models have become widely used tools and play a fundamental role in much of the application of statistics today. In particular, ANOVA models involving random effects have found widespread application to experimental design in a variety of fields requiring measurements of variance, including agriculture, biology, animal breeding, applied genetics, econometrics, quality control, medicine, engineering, and social sciences. This two-volume work is a comprehensive presentation of different methods and techniques for point estimation, interval estimation, and tests of hypotheses for linear models involving random effects. Both Bayesian and repeated sampling procedures are considered. Volume I examines models with balanced data (orthogonal models); Volume II studies models with unbalanced data (nonorthogonal models). Features and Topics: * Systematic treatment of the commonly employed crossed and nested classification models used in analysis of variance designs * Detailed and thorough discussion of certain random effects models not commonly found in texts at the introductory or intermediate level * Numerical examples to analyze data from a wide variety of disciplines * Many worked examples containing computer outputs from standard software packages such as SAS, SPSS, and BMDP for each numerical example * Extensive exercise sets at the end of each chapter * Numerous appendices with background reference concepts, terms, and results * Balanced coverage of theory, methods, and practical applications * Complete citations of important and related works at the end of each chapter, as well as an extensive general bibliography Accessible to readers with only a modest mathematical and statistical background, the work will appeal to a broad audience of students, researchers, and practitioners in the mathematical, life, social, and engineering sciences. It may be used as a textbook in upper-level undergraduate and graduate courses, or as a reference for readers interested in the use of random effects models for data analysis.
[Show abstract][Hide abstract] ABSTRACT: This study examined the effectiveness of a combined counseling and case management behavioral intervention, using motivational interviewing strategies, in engaging Hispanic injection drug users in treatment and reducing drug use and injection-related HIV risk behaviors. Follow up data are presented on 440 (79.0%) of 557 randomized participants, 6 months after the initial interview. Subjects in the experimental arm were significantly less likely to continue drug injection independent of entering drug treatment, and were also more likely to enter drug treatment. Subjects in both arms who entered drug treatment were less likely to continue drug injection. Among subjects who continued drug injection, those in the experimental arm were significantly less likely to share needles. Confirming the outcomes of this study in other Hispanic sites and populations could be a critical step towards reducing factors that contribute to the self-sustaining HIV/AIDS epidemic in Puerto Rico and communities in the U.S. mainland.
No preview · Article · Oct 2004 · Journal of Substance Abuse Treatment
[Show abstract][Hide abstract] ABSTRACT: This paper assesses mortality rate for a cohort of drug users in Puerto Rico compared with that of the Island's general population, examining causes of death and estimating relative risk of death. Date and cause of death were obtained from death certificates during 1998. Vital status was confirmed through contact with subjects, family, and friends. HIV/AIDS was the major cause of death (47.7%), followed by homicide (14.6%), and accidental poisoning (6.3%). Females had higher relative risk of death than males in all age categories. Not living with a sex partner and not receiving drug treatment were related to higher mortality due to HIV/AIDS. Drug injection was the only variable explaining relative risk of death due to overdose. Puerto Rico needs to continue developing programs to prevent HIV/AIDS among drug users. Special attention should be given to young women, who appear to be in greatest need of programs to prevent early mortality.
Full-text · Article · Jan 2004 · Puerto Rico health sciences journal
[Show abstract][Hide abstract] ABSTRACT: This study was conducted to identify factors accounting for differences in health care and drug treatment utilization between
Puerto Rican drug users residing in 2 separate locations. Survey findings from 334 drug users in Puerto Rico and 617 in New
York City showed that those in Puerto Rico were 6 times less likely than their counterparts in New York to have used inpatient
medical services and 13 to 14 times less likely to have used outpatient medical services or methadone. They also were less
likely to have health insurance or past drug treatment. After site was controlled for, health insurance and previous use of
physical or mental health services remained significant predictors of health care and drug treatment utilization during the
study period. Although Puerto Rican drug users in Puerto Rico are not an ethnic minority, they reported significant disparities
in health services use compared with Puerto Rican drug users in New York.
Full-text · Article · Jan 2004 · Clinical Infectious Diseases
[Show abstract][Hide abstract] ABSTRACT: It is known that vitamin C status is compromised in smokers. The vitamin C status of nonsmokers who are exposed to environmental tobacco smoke (ETS) is now being elucidated.
We assessed vitamin C status in children who were either exposed or not exposed to ETS, and we sought to associate changes in vitamin C status with the amount of ETS exposure.
The study group included 512 children aged 2-12 y; 50% of them were exposed to ETS in the home because their parents smoked. Dietary intake of vitamin C, obtained with a 24-h recall questionnaire, and blood ascorbate concentrations were compared in the exposed and unexposed groups. Smoke exposure was assessed by measuring a biomarker, urinary cotinine. Age, sex, and body mass index were examined as potential correlates of vitamin C status in each exposure category.
Plasma ascorbate concentrations were lower, by 3.2 micro mol/L on average, in ETS-exposed children than in unexposed children who consumed equivalent amounts of vitamin C; this was a highly significant difference (P = 0.002). This reduction in plasma ascorbate occurred even with very low exposure to ETS.
ETS can reduce concentrations of ascorbate, an important blood antioxidant, even when the amount of smoke exposure is minimal. Children exposed to ETS should be encouraged to consume increased amounts of foods rich in vitamin C or should be given the equivalent amount of this vitamin as a supplement.
No preview · Article · Feb 2003 · American Journal of Clinical Nutrition
[Show abstract][Hide abstract] ABSTRACT: The use of a project-based approach for designing a one-year graduate level programme in applied statistics is addressed. The pedagogical approach, academic setting, and learning activities are described in a multidisciplinary context. Some comments on implementation of such a programme based on the results from five successive graduating classes at the University of Veracruz (1994–2000) are included.
Full-text · Article · Jan 2003 · International Journal of Mathematical Education