Time-invariant Explanatory Variables Research Articles

The roles of depression and social relationships in the onset and course of loneliness amongst Nigerian elders.

Few studies in Africa have investigated the risk profile and course of loneliness in old age. This study examined the risk factors for onset and chronicity, as well as the predictors of recovery from loneliness in a large representative sample of community dwelling older Africans. A household multistage probability sample of Nigerians who were 65 years or older was drawn from a geographical area with approximately 25 million population. Loneliness was measured using the 3-item University of California (UCLA) scale in 2007 and annually in 2008 and 2009. Social engagement, social network, and depression were evaluated using the WHO Composite International Diagnostic Interview. Respondents were also administered the 30-item Geriatric Depression Scale. Multivariate logistic regression models were used to explore for risk factors. Estimates of adjusted hazard ratios (HR) for recovery were derived with the discrete time version of the Cox regression model for time invariant explanatory variables. Of 1704 respondents, 1525 were free of loneliness, using the UCLA scale, in 2007. A total of 209 (18.8%) persons developed new onset of loneliness in 2008 and 2009. Depression (O.R = 2.9, 95% C.I = 1.3-6.7), unmarried status (OR = 2.1, 95% C.I = 1.2-3.9) and social isolation (OR = 1.8, 95% CI = 1.0-3.2) independently predicted loneliness onset. Baseline demographic, health, social, and lifestyle factors were not associated with a chronic course of loneliness. The overall recovery rate estimated over two years was 89.5% (95% CI = 75.3-106.4). Being male (HR = 1.3, 95% C.I = 1.0-1.6), ≥80 years (HR = 1.4, 95% CI = 1.2-1.8) and having good social engagement at baseline (HR = 1.5, 95% C. = 1.1-2.0) independently predicted recovery from loneliness. Over a 2-year period, nearly one in five community-dwelling Africans developed new onset loneliness in old age, with a similar proportion having a chronic course of the emotional experience. While depression and indices of social isolation at baseline were associated with onset, good social engagement predicted recovery from loneliness.

The Association between Sport Activities and Educational Achievements: Evidence from Russian Longitudinal Data

This paper empirically studies the association between sport activities and educational achievements of school students from 1st to 11th grade. The used sample included observations over the period 2010-2015 taken from the Russia Longitudinal Monitoring Survey (waves 19-24), which is a unique nationally representative survey. The method consisted of logit regressions with panel data, which allows control for time-invariant explanatory variables. The empirical analysis was divided into sport activities at school (in class) and out-of-school (before or after classes). Furthermore, the regression analysis examined the effect of three large groups of sport activities: 1) Combat sports, such as karate, judo, self-defense, wrestling, and boxing, 2) Ball sports, such as tennis, soccer, basketball, and volleyball, 3) Athletic sports, such as track and field, skiing, and skating. General speaking, the findings indicated that sport activities at school do not have significant associations with educational achievements. On the other hand, sport activities out-of-school showed some positive relationships. Specifically, participation in athletic and combat sports increases the probabilities of boys and girls, respectively, of being classified as high-performing students. Moreover, male students practicing ball sports out-of-school are less likely to be classified as low-performing students. The time that students spent on these sports does not influence these probabilities. However, male students spending more than 10 hours per week on sports (high-performance sportsmen) are more likely than other students of being linked to the group of low-performing students.

Estimation of Some Nonlinear Panel Data Models With Both Time-Varying and Time-Invariant Explanatory Variables

The so-called “fixed effects” approach to the estimation of panel data models suffers from the limitation that it is not possible to estimate the coefficients on explanatory variables that are time-invariant. This is in contrast to a “random effects” approach, which achieves this by making much stronger assumptions on the relationship between the explanatory variables and the individual-specific effect. In a linear model, it is possible to obtain the best of both worlds by making random effects-type assumptions on the time-invariant explanatory variables while maintaining the flexibility of a fixed effects approach when it comes to the time-varying covariates. This article attempts to do the same for some popular nonlinear models.

Explanatory, Multilevel Person-Fit Analysis of Response Consistency on the Spielberger State-Trait Anxiety Inventory

Self-report measures are vulnerable to concentration and motivation problems, leading to responses that may be inconsistent with the respondent's latent trait value. We investigated response consistency in a sample (N = 860) of cardiac patients with an implantable cardioverter defibrillator and their partners who completed the Spielberger State-Trait Anxiety Inventory on five measurement occasions. For each occasion and for both the state and trait subscales, we used the l p z person-fit statistic to assess response consistency. We used multilevel analysis to model the between-person and within-person differences in the repeated observations of response consistency using time-dependent (e.g., mood states) and time-invariant explanatory variables (e.g., demographic characteristics). Respondents with lower education, undergoing psychological treatment, and with more post-traumatic stress disorder symptoms tended to respond less consistently. The percentages of explained variance in response consistency were small. Hence, we conclude that the results give insight into the causes of response inconsistency but that the identified explanatory variables are of limited practical value for identifying respondents at risk of producing invalid test results. We discuss explanations for the small percentage of explained variance and suggest alternative methods for studying causes of response inconsistency.

Which System GMM to Use? A Simulation Exercise to Identify the Ideal Estimator for Data of Different Sizes

Economists analyzing models using panel data often have to estimate the coefficients of time-invariant explanatory variables in their model. Although system Generalized Method of Moments (GMM) is the method of choice for many economists to estimate coefficients in their models, very little is known about which system GMM is best suited to provide valid estimates of coefficients of time-invariant variables under different sample sizes. I address this research question by running a series of Monte Carlo simulations on some of the more popular system GMMs (Arellano and Bover, 1995; Blundell and Bond, 1998). I find that Arellano and Bover (1995) system GMM provides the best estimates of the coefficient on time-invariant variable when the number of observation is large relative to the time periods available in the data, regardless of whether the time-varying explanatory variables are predetermined or exogenous. When the sample size is small relative to time period available in the dataset and the explanatory variables are predetermined, I find that the long-difference moment condition and estimation technique of Hahn et al (2007), combined with the time-mean moment condition of Arellano and Bover (1995) provides the best estimates of the coefficients on the time-invariant variable. I also find that this estimator provides valid estimates of the coefficients of time-varying variables under such circumstances.

Instrumental variable estimation of heteroskedasticity adaptive error component models

The linear panel data estimator proposed by Hausman and Taylor relaxes the hypothesis of exogenous regressors that is assumed by generalized least squares methods but, unlike the Fixed Effects estimator, it can handle endogenous time invariant explanatory variables in the regression equation. One of the assumptions underlying the estimator is the homoskedasticity of the error components. This can be restrictive in applications, and therefore in this paper the assumption is relaxed and more efficient adaptive versions of the estimator are presented.

Research expenditures, technology transfer activity, and university licensing revenue

In this paper we relate university licensing revenues to both university research expenditures and characteristics of the university and the university technology transfer office. We apply the Hausman–Taylor estimator for panel data with time-invariant explanatory variables and the Arellano–Bover dynamic panel model to unbalanced panels for the years 1991–2003 and balanced panels for the years 1995–2003. We find conflicting evidence regarding the short-term impacts of research expenditures on licensing revenues. On the other hand, both early initiation of technology transfer programs and staff size increase expected licensing revenues. Staff size and early entry appear to be substitutes, however. One-year lagged licensing revenue has strong predictive power for current licensing revenue. Further research is necessary to analyze changes in technology transfer office efficiency over time and the contribution of technology transfer to larger university missions.

Fixed-effect estimation of technical efficiency with time-invariant dummies

“Within” estimation of the fixed-effect stochastic frontier model does not identify parameters on time-invariant explanatory variables. If time-invariant variables are important production inputs, then standard efficiency estimates are biased. This note details bias correction, when time-invariant inputs are dummy variables.

Estimation of Growth Curve Models with Structured Error Covariances by Generalized Estimating Equations

The growth curve model is useful for the analysis of longitudinal data. It helps investigate an overall pattern of change in repeated measurements over time and the effects of time-invariant explanatory variables on the temporal pattern. The traditional growth curve model assumes that the matrix of covariances between repeated measurements is unconstrained. This unconstrained covariance matrix often appears unattractive. In this paper, the generalized estimating equation method is adopted to estimate parameters of the growth curve model. As a result, the proposed method allows a more variety of constrained covariance structures than the traditional growth curve model. An empirical application is provided so as to illustrate the proposed method.