Categorical Dependent Variables Research Articles

Farm size, seining practices, and salt use: Risk factors for Aeromonas hydrophila outbreaks in farm-raised catfish, Alabama, USA

In freshwater aquaculture systems, Aeromonas hydrophila is usually considered to be an opportunistic pathogen most often associated with secondary bacterial infections. Since 2009, the U.S. catfish industry, especially in West Alabama, has been affected by mortality from a strain of A. hydrophila that has been acting as a primary pathogen. Tens of millions of pounds of catfish production have been lost as a consequence of this disease. This study used data from two whole-population farmer surveys to examine farm-level risk factors for two A. hydrophila outbreaks in foodsize Alabama catfish, one in 2009 (surveyed in 2010), and one in 2011 (surveyed in 2012). The response to the 2010 survey was 85% and the response to the 2012 survey was 82%. Univariate analyses were used to examine biologically plausible variables (farm size, pond stocking density, seine exposure, use of salt (NaCl) in ponds), and used categorical disease outcome and dependent variables. Farm size was included in bivariate analyses with the other variables, because it was a potential confounding variable. For both study years, the odds of an A. hydrophila outbreak were significantly greater for farms larger than the mean size (2009: mean=132 acres (53.4 hectares), odds ratio (OR)=8.2; 95% confidence interval (CI)=3.3–20.6, p<0.001; 2011: mean=116 acres (46.9 hectares), OR=5.3, CI=1.7–17.0, p=0.009). Compared with 42% of control farms, every case farm was seined by a commercial or processing plant seining crew in 2009. The bivariate analysis of the 2011 variable “average number of times each pond was seined per year” indicated that regardless of farm size, farms with ponds that were seined more than twice per year had a significantly greater odds of an A. hydrophila outbreak (OR=4.1, CI=1.2–14.4, p=0.02). For 2009, the results of the bivariate analyses of chloride concentrations indicated that farms that had chloride concentrations >135ppm had a significantly lower odds of experiencing A. hydrophila outbreaks (OR=0.2, CI=0.05–0.6, p-value=0.004). To achieve economies of scale, catfish farmers raise fish on large farms at higher stocking densities, but this practice may result in increased susceptibility to disease outbreaks. Producers should prioritize implementing biosecurity measures such as improved seining practices and other management practices to protect fish grown at high population densities. Further work will determine what the detailed seining protocols should include, and whether the use of salt, and at what concentrations, reduces the risk of A. hydrophila outbreaks.

Addressing unobserved endogeneity bias in accounting studies: control and sensitivity methods by variable type

Together with their associated statistical routines, this paper describes the control and sensitivity methods that can be employed by accounting researchers to address the important issue of unobserved (omitted) variable bias in regression and matching models according to the types of variables employed. As with other social science disciplines, an important and pervasive issue in observational (non-experimental) accounting research is omitted variable bias (endogeneity). Causal inferences for endogenous explanatory variables are biased. This occurs in regression models where an unobserved (confounding) variable is correlated with both the dependent (outcome) variable in a regression model and the causal explanatory (often a selection) variable of interest. The Heckman treatment effect model has been widely employed to control for hidden bias for continuous outcomes and endogenous binary selection variables. However, in accounting studies, limited (categorical) dependent variables are a common feature and endogenous explanatory variables may be other than binary in nature. The purpose of this paper is to provide an overview of contemporary control methods, together with the statistical routines to implement them, which extend the Heckman approach to binary, multinomial, ordinal, count and percentile outcomes and to where endogenous variables take various forms. These contemporary methods aim to improve causal estimates by controlling for hidden bias, though at the price of increased complexity. A simpler approach is to conduct sensitivity analysis. This paper also presents a synopsis of a number of sensitivity techniques and their associated statistical routines which accounting researchers can employ routinely to appraise the vulnerability of causal effects to potential (simulated) unobserved bias when estimated with conventional regression and propensity score matching estimators.

Social and economic characteristics of street youth by gender and level of street involvement in Eldoret, Kenya.

BackgroundStreet-connected youth are a neglected and vulnerable population, particularly in resource-constrained settings. The development of interventions and supports for this population requires insight into how they live. This study describes the social and economic characteristics of a convenience sample of street youth (SY) in Eldoret, Kenya.MethodsParticipants were eligible if they were aged 12–21, living in Eldoret, spending days only (part-time), or nights and days on the street (full-time) and able and willing to consent or assent. Data were collected using a standardized interview conducted in English or Kiswahili. Binary dependent variables were having been arrested and/or jailed, and first priority for spending money (food vs. other). Nominal categorical dependent variables included major source of support, and major reason for being street-involved. Multivariable analysis used logistic regression models to examine the association of gender and level of street-involvement with social and economic factors of interest adjusting for age and length of time on the street. Data were analyzed using SAS 9.3.ResultsOf the 200 SY enrolled, 41% were female, mean age of 16.3 years; 71% were on the street full-time, and 29% part-time. Compared with part-time SY, full-time SY were more likely to have been arrested (Adjusted Odds Ratio [AOR]: 2.33, 95% Confidence Interval [95%CI]:1.01–5.35), name food as their first spending priority (AOR: 2.57, 95%CI:1.03–6.45), have left home due to violence (AOR: 5.54, 95%CI: 1.67–18.34), and more likely to report friends on the street as a major source of support (AOR: 3.59, 95% CI: 1.01–12.82). Compared with females, males were more likely to have ever been arrested (AOR: 2.66, 95%CI:1.14–6.18), and to have ever been jailed (AOR: 3.22, 95%CI:1.47–7.02).ConclusionsThese results suggest a high degree of heterogeneity and vulnerability among SY in this setting. There is an urgent need for interventions taking into consideration these characteristics.

Disease Diagnosis with a hybrid method SVR using NSGA-II

Early diagnosis of any disease at a lower cost is preferable. Automatic medical diagnosis classification tools reduce financial burden on health care systems. In medical diagnosis, patterns consist of observable symptoms and the results of diagnostic tests, which have various associated costs and risks. In this paper, we have experimented and suggested an automated pattern classification method for classifying four diseases into two classes.In the literature on machine learning or data mining, regression and classification problems are typically viewed as two distinct problems differentiated by continuous or categorical dependent variables. There are endeavors to use regression methods to solve classification problems and vice versa. To regard a classification problem as a regression one, we propose a method based on the Support Vector Regression (SVR) classification model as one of the powerful methods in intelligent field management. We apply the Non-dominated Sorting Genetic Algorithm-II (NSGA-II), a kind of multi-objective evolutionary algorithm, to find mapping points (MPs) for rounding a real-value to an integer one. Also, we employ the NSGA-II to find out and tune the SVR kernel parameters optimally so as to enhance the performance of our model and achieve better results. The results of the study are compared with the results of some previous studies focusing on the diagnoses of four diseases using the same UCI machine learning database. The experimental results show that the proposed method yields a superior and competitive performance in these four real-world datasets.

The association between genetic variants of RUNX2, ADIPOQ and vertebral fracture in Korean postmenopausal women

Contrary to the traditional belief that obesity acts as a protective factor for bone, recent epidemiologic studies have shown that body fat might be a risk factor for osteoporosis and bone fracture. Accordingly, we evaluated the association between the phenotypes of osteoporosis or vertebral fracture and variants of obesity-related genes, peroxisome proliferator-activated receptor-gamma (PPARG), runt-related transcription factor 2 (RUNX2), leptin receptor (LEPR), and adiponectin (ADIPOQ). In total, 907 postmenopausal healthy women, aged 60-79 years, were included in this study. BMD and biomarkers of bone health and adiposity were measured. We genotyped for four single nucleotide polymorphisms (SNPs) from four genes (PPARG, RUNX2, LEPR, ADIPOQ). A general linear model for continuous dependent variables and a logistic regression model for categorical dependent variables were used to analyze the statistical differences among genotype groups. Compared with the TT subjects at rs7771980 in RUNX2, C-carrier (TC + CC) subjects had a lower vertebral fracture risk after adjusting for age, smoking, alcohol, total calorie intake, total energy expenditure, total calcium intake, total fat intake, weight, body fat. Odds ratio (OR) and 95% interval (CI) for the vertebral fracture risk was 0.55 (95% CI 0.32-0.94). After adjusting for multiple variables, the prevalence of vertebral fracture was highest in GG subjects at rs1501299 in ADIPOQ (p = 0.0473). A high calcium intake (>1000 mg/day) contributed to a high bone mineral density (BMD) in GT + TT subjects at rs1501299 in ADIPOQ (p for interaction = 0.0295). Even if the mechanisms between obesity-related genes and bone health are not fully established, the results of our study revealed the association of certain SNPs from obesity-related genes with BMD or vertebral fracture risk in postmenopausal Korean women.

Early Detection of Potential Bank Bankruptcy Through Financial Ratio Analysis: Multinomial Logistic Regression Model

Prediction of potential bank bankruptcy based on financial ratios is a continuing research. This study is aimed to provide prediction model capable of explaining bank’s health, predicting or detecting early potential bankruptcy of bank, finding formula that can be applied to all banks, promoting sound banking and simultaneously creating economic prosperity of the country considering that bank is the country’s economic infrastructure. Statistical technique based on multinomial logistic regression model is used as method to test the model with categorical dependent variables given a set of independent variables. It is found that financial ratio related to bank’s capital adequacy is statistically significant (in two logit functions) in providing early detection of potential bank insolvency. The accuracy of predictions by the model is 75% for failed banks (BL), 62.50% for banks classified under special surveillance or banks in resolution (BDP), and 97.14% for healthy banks (BS). Financial ratios are believed to have contributed to the bankruptcy prediction model by 89.36%.

Assessment of demographic and socio-economic factors on ailments and disability among elderly

The present study makes an effort to study the influence of demographic and socio-economic factors on ailments and disability among elderly people living in urban and rural areas of India. Data is collected through a well-structured questionnaire, specifically designed for the purpose, to obtain information on demographic factors like gender, marital status, educational qualification, kinship, etc., and socio-economic factors such as annual income, immovable properties, income generating livestock, etc. A sample of 288 elderly men and women above the age of 60 years is considered from both urban and rural household from the State of Odisha, India. Logistic regression method is tailored to fit the categorical dependent variables. The results indicate that the variables which show significant effect on aliments and disability in rural setting are marital status, risk behaviour, and kinship. Similarly, the variables which show significant effect on urban setting are livestock, ownership of house and risk behaviour. Risk behaviour comes out to be a significant factor in both the settings. It is observed that the elderly, having bad habits especially consumption of alcohol and smoking, are prone to ailments leading to disability in the long run.

Travellers’ Satisfaction with Railway Transport: A Bayesian Network Approach

Survey data are often employed as a tool to support planners in defining and reviewing their intervention programs and policies. In the current literature, a broad range of studies have examined the determinants and components of consumers’ satisfaction. Most of these studies have exploited statistical models, such as probit or logit, to determine the categorical dependent variables. Less attention has been paid to the investigation of the structural properties of the data by testing alternative methods and procedures. The present work performs such an analysis using data from a survey on consumer satisfaction from railway transport in 14 EU countries. The data were collected in a large survey (more than 17,000 observations) conducted on behalf of the European Commission. By applying both ordered logistic regression and Bayesian Network (BN) analysis, the research question addressed by this paper is twofold. First of all, the performance of the two methodologies, defined in terms of their predictive capability, is assessed. Secondly, the main policy implications conveyed by the two models are compared. According to the results, BN analysis provides better predictions of the outcomes of customers’ satisfaction with railway transport. Moreover, the choice of the statistical methodology is a relevant issue for policy-making, since the policy messages conveyed by the two models differ significantly.

Variable selection for multivariate generalized linear models

Generalized linear models (GLMs) are widely studied to deal with complex response variables. For the analysis of categorical dependent variables with more than two response categories, multivariate GLMs are presented to build the relationship between this polytomous response and a set of regressors. Traditional variable selection approaches have been proposed for the multivariate GLM with a canonical link function when the number of parameters is fixed in the literature. However, in many model selection problems, the number of parameters may be large and grow with the sample size. In this paper, we present a new selection criterion to the model with a diverging number of parameters. Under suitable conditions, the criterion is shown to be model selection consistent. A simulation study and a real data analysis are conducted to support theoretical findings.

Identifying the Factors that Influence Change in SEBD Using Logistic Regression Analysis

Multiple linear regression and ANOVA models are widely used in applications since they provide effective statistical tools for assessing the relationship between a continuous dependent variable and several predictors. However these models rely heavily on linearity and normality assumptions and they do not accommodate categorical dependent variables. The seminal contribution of John Nelder and Robert Wedderburn (1972) introduced the concept of Generalized Linear Models. GLMs overcome the limitations of Normal regression models and accommodate any distribution which is a member of the exponential family. Moreover, these models relate the dependent variable to the linear predictor (non-random component) through any invertible link function. Logistic regression models are GLMs that accommodate categorical dependent variables. They assume a Binomial distribution and Logit canonical link function. The iteratively re-weighted least squares algorithm using the Fisher scoring technique is employed to maximize the log-likelihood function in GLMs and estimate the model parameters. In this paper, Logistic regression analysis was used to identify the dominant factors that influence change in social, emotional and behaviour difficulties (SEBD) of Maltese children. The study comprised 486 pupils whose SEBD was assessed by both teachers and parents using the Strengths and Difficulties Questionnaire (Goodman 1997) when the children were aged 6 and 9 years old.

Using the Margins Command to Estimate and Interpret Adjusted Predictions and Marginal Effects

Many researchers and journals place a strong emphasis on the sign and statistical significance of effects—but often there is very little emphasis on the substantive and practical significance of the findings. As Long and Freese (2006, Regression Models for Categorical Dependent Variables Using Stata [Stata Press]) show, results can often be made more tangible by computing predicted or expected values for hypothetical or prototypical cases. Stata 11 introduced new tools for making such calculations—factor variables and the margins command. These can do most of the things that were previously done by Stata's own adjust and mfx commands, and much more. Unfortunately, the complexity of the margins syntax, the daunting 50-page reference manual entry that describes it, and a lack of understanding about what margins offers over older commands that have been widely used for years may have dissuaded some researchers from examining how the margins command could benefit them. In this article, therefore, I explain what adjusted predictions and marginal effects are, and how they can contribute to the interpretation of results. I further explain why older commands, like adjust and mfx, can often produce incorrect results, and how factor variables and the margins command can avoid these errors. The relative merits of different methods for setting representative values for variables in the model (marginal effects at the means, average marginal effects, and marginal effects at representative values) are considered. I shows how the marginsplot command (introduced in Stata 12) provides a graphical and often much easier means for presenting and understanding the results from margins, and explain why margins does not present marginal effects for interaction terms.

Relationships between categorical dependent variables and other variables and between waiting times and other variables

Relationships between categorical dependent variables and other variables and between waiting times and other variables

CONDITIONAL AND MULTINOMIAL LOGITS AS BINARY LOGIT REGRESSIONS

For a categorical variable with several outcomes, its dependence on the predictors is usually considered in the conditional or multinomial logit models. This work considers elasticity features of the binary and categorical logits and introduces the coefficients individual by observations. The paper shows that by a special rearrangement of data the more complicated conditional and multinomial models can be reduced to binary logistic regression. It suggests the usage of any software widely available for logit modeling to facilitate constructing for complex conditional and multinomial regressions. In addition, for binary logit, it is possible to obtain meaningful coefficients of regression by transforming data to the linear link function, which opens a possibility to obtain meaningful parameters of the complicated models with categorical dependent variables.

Categorization of Beekeepers from the North-West Region of Romania

The aim of the present study is to develop a categorization of beekeepers from the North-West Region of Romania using multiple correspondence analysis. Multiple correspondence analysis is an extension of correspondence analysis which allows the analysis of the pattern of relationships of several categorical dependent variables (Abdi and Valentin, 2007). The data was collected through a survey administered to a sample composed of 420 beekeepers from the North-West Region of Romania. The present categorization implies that the beekeepers form the sample are grouped into categories according to their age and the dimension of their beekeeping exploitation, for the specific purpose of distinguishing several aspects that are specific to each category. The results of the analysis illustrate a relationship between beekeepers and other factors such as: ways of commercializing beekeeping products, propensity for European Union measures, opinions regarding the current legislation in the beekeeping sector. Each category of beekeepers has its own characteristics and should be addressed according to its specificity by public measures. The average age with average to high beekeeping exploitation is characterized by greater openness to European Union Funding Programmes and is specific for Family business or Sole Proprietorship. Openness to specific measures for young beekeepers is obviously specific to young beekeepers with small beekeeping exploitations and who prefer the distribution of their products to several stores. Modernization trends are specific to beekeepers of average age and average beekeeping exploitation size. Higher experience, advanced age and large-scale beekeeping exploitations characterises the veterans in apiculture who are all discontented with the state measures in this sector. Amateur beekeepers have a slightly above average age and want to sell their products to friends or acquaintances. A categorization of beekeepers from the North-West Region of Romania is extremely important for a better understanding of the factors and mechanisms that determine the development of the sector. Taking into consideration the resource constraints faced by beekeeping exploitations and their vulnerability to the environment, policy measures must be implemented to support small scale beekeepers as well as beekeepers that have large beekeeping exploitations.

Tabulating SPost Results Using Estout and Esttab

The SPost user package (Long and Freese, 2006, Regression Models for Categorical Dependent Variables Using Stata [Stata Press]) is a suite of postestimation commands to compute additional tests and effects representations for a variety of regression models. To facilitate and automate the task of tabulating results from SPost commands for inclusion in reports, publications, and presentations, we introduce tools to integrate SPost with the estout user package (Jann, 2005, Stata Journal 5: 288–308; 2007, Stata Journal 7: 227–244). The estadd command can retrieve results computed by the SPost commands brant, fitstat, listcoef, mlogtest, prchange, prvalue, and asprvalue. These results can then be tabulated by esttab or estout.

One adult who is crazy about you: Can natural mentoring relationships increase assets among young adults with and without foster care experience?

During emerging adulthood, most youth receive family support to help them weather the difficulties associated with transitioning to independence. When foster youth emancipate, they confront the challenges associated with emerging adulthood, and are at risk of having to transition without family support. Many are in danger of failing to meet minimal levels of self-sufficiency. A caring adult who offers social support is normative for adolescent development and protective for youth across many risk conditions. Natural mentoring can cultivate such relationships. This study examines the association between natural mentor relationship characteristics, and material hardship and asset-related outcomes during the emerging adulthood period in both a normative sample of young adults and young adults identified as former foster youth. This study also considers the potential mediating effect of future expectations. Data from Wave 3 of the National Longitudinal Study of Adolescent Health that pertain to 15,197 respondents are used. Path models with categorical dependent variables were estimated using a maximum likelihood method with standard errors that are robust to non-normality and non-independence of observations. “Like a parent”, “role model”, and “guidance/advice” were significantly associated with assets among both groups. This study contributes to the growing body of literature on natural mentoring and former foster youth, and highlights the value of increasing our understanding of natural mentor roles for intervention development. The focus on asset-related outcomes is a novel approach to investigate the benefits of natural mentoring to the healthy development of youth. This paper is the first to consider the association between natural mentoring and asset building among both former and nonformer foster youth.

The FAQs on Data Transformation

Although we often hear that data speak for themselves, their voices can be soft and sly. (Mosteller, Fienberg, & Rourke, 1983, p. 234) The job of the data analyst, strangely enough, is to find rand...

Mental Health Treatment Received by Youths in the Year Before and After a New Diagnosis of Bipolar Disorder

Despite a marked increase in treatment for bipolar disorder among youths, little is known about their pattern of service use. This article describes mental health service use in the year before and after a new clinical diagnosis of bipolar disorder. Claims were reviewed between April 1, 2004, and March 31, 2005, for 1,274,726 privately insured youths (17 years and younger) who were eligible for services at least one year before and after a service claim; 2,907 youths had new diagnosis of bipolar disorder during this period. Diagnoses of other mental disorders and prescriptions filled for psychotropic drugs were assessed in the year before and after the initial diagnosis of bipolar disorder. The one-year rate of a new diagnosis of bipolar disorder was .23%. During the year before the new diagnosis of bipolar disorder, youths were commonly diagnosed as having depressive disorder (46.5%) or disruptive behavior disorder (36.7%) and had often filled a prescription for an antidepressant (48.5%), stimulant (33.0%), mood stabilizer (31.8%), or antipsychotic (29.1%). Most youths with a new diagnosis of bipolar disorder had only one (28.8%) or two to four (28.7%) insurance claims for bipolar disorder in the year starting with the index diagnosis. The proportion starting mood stabilizers after the index diagnosis was highest for youths with five or more insurance claims for bipolar disorder (42.1%), intermediate for those with two to four claims (24.2%), and lowest for those with one claim (13.8%). Most youths with a new diagnosis of bipolar disorder had recently received treatment for depressive or disruptive behavior disorders, and many had no claims listing a diagnosis of bipolar disorder after the initial diagnosis. The service pattern suggests that a diagnosis of bipolar disorder is often given tentatively to youths treated for mental disorders with overlapping symptom profiles and is subsequently reconsidered.

Nonparametric Methods for Modeling Nonlinearity in Regression Analysis

The linear model and related generalized linear model (GLM) are important tools for sociologists. If the relationships between y (or in the case of the GLM, the linear predictor η) and the xs are linear, these methods provide elegant summaries of the data. However, these methods fail to adequately model underlying relationships if they are characterized by complex nonlinear patterns. In such cases, nonparametric regression, which allows the functional form between y and x to be determined by the data themselves, is more suitable. There are many types of nonparametric simple regression. I focus on locally weighted scatterplot smoothing (lowess or loess) and smoothing splines because they are the most widely used. I also describe additive and generalized additive models (GAM), which allow modeling of categorical dependent variables, and I explain how these methods can handle both parametric and nonparametric (i.e., lowess and smoothing splines) effects for many predictors. Finally, I briefly introduce the more recent development of the vector generalized additive model (VGAM), which further extends the GAM to handle multivariate dependent variables, and the generalized additive mixed model (GAMM), which allows specification of smooth functions within the mixed model framework.

Smoking cessation and periodontal health – a missed opportunity?

This was a cohort study. COHORT SELECTION: A longstanding prospective study was carried out of a birth cohort born in Dunedin (New Zealand) in 1972-1973. EXPOSURE MEASUREMENT: Periodontal examinations of the study participants were carried out when they reached the age of 26 and 32 years by calibrated examiners; smoking exposure was measured by questioning at ages 15, 18, 21, 26 and 32 years. Socioeconomic status (SES) was measured at the age of 26 years by categorising adult occupation using standard New Zealand occupation indices. Dental visiting behaviour was assessed by questioning at age 25 and 32 years. Chi-square tests were used to examine the statistical significance of differences observed with categorical dependent variables. Analysis of variance was used for continuous variables. Logistic regression modelling was used to examine smoking exposure and periodontitis prevalence and incidence while controlling for sex, SES, dental plaque accumulation and the use of dental services. Complete data were available for 810 individuals of whom 48.9% had smoked at some point (31.5% were current smokers). Table 1 shows the prevalence, for a selection of smoking exposure groups of attachment loss at age 32 and the odds ratio compared with never-smokers and Table 2 the incidence of 3+ mm attachment loss. Sites with >5 mm attachment loss and were more likely to be incident cases after age 26 (OR, 5.2 and 3.2, respectively). Two thirds of new cases in individuals older than 26 years were attributable to smoking. There were no significant differences in periodontal health between never-smokers and those who had quit smoking after age 26. Current and long-term smoking in young adults is detrimental to periodontal health, but smoking cessation may be associated with a relatively rapid improvement in the periodontium.