Search

Background: Patterns of survey response and the characteristics associated with response over time in longitudinal studies are important to discern for the development of tailored retention efforts aimed at minimizing response bias. The Millennium Cohort Study, the largest and longest running cohort study of military personnel and veterans, is designed to examine the long-term health effects of military service and experiences and thus relies on continued participant survey responses over time. Here, we describe the response rates for follow-up survey data collected over 15 years and identify characteristics associated with follow-up survey response and mode of response (paper vs. web). Method: Patterns of follow-up survey response and response mode (web, paper, none) were examined among eligible participants (n=198,833), who were initially recruited in four panels from 2001 to 2013 in the Millennium Cohort Study, for a follow-up period of 3–15 years (2004–2016). Military and sociodemographic factors (i.e., enrollment panel, sex, birth year, race and ethnicity, educational attainment, marital status, service component, service branch, pay grade, military occupation, length of service, and time deployed), life experiences and health-related factors (i.e., military deployment/combat experience, life stressors, mental health, physical health, and unhealthy behaviors) were used to examine follow-up response and survey mode over time in multivariable generalized estimating equation models. Results: Overall, an average response rate of 60% was observed across all follow-up waves. Factors associated with follow-up survey response over time included increased educational attainment, married status, female sex, older age, military deployment (regardless of combat experience), and higher number of life stressors, mental health issues, and physical health diagnoses. Conclusion: Despite the challenges associated with collecting multiple waves of follow-up survey data from members of the U.S. military during and after service, the Millennium Cohort Study has maintained a relatively robust response rate over time. The incorporation of tailored messages and outreach to those groups least likely to respond over time may improve retention and thereby increase the representativeness and generalizability of collected survey data. [ABSTRACT FROM AUTHOR]

Zero-inflated Poisson (ZIP) and negative binomial (ZINB) models are widely used to model zero-inflated count responses. These models extend the Poisson and negative binomial (NB) to address excessive zeros in the count response. By adding a degenerate distribution centered at 0 and interpreting it as describing a non-risk group in the population, the ZIP (ZINB) models a two-component population mixture. As in applications of Poisson and NB, the key difference between ZIP and ZINB is the allowance for overdispersion by the ZINB in its NB component in modeling the count response for the at-risk group. Overdispersion arising in practice too often does not follow the NB, and applications of ZINB to such data yield invalid inference. If sources of overdispersion are known, other parametric models may be used to directly model the overdispersion. Such models too are subject to assumed distributions. Further, this approach may not be applicable if information about the sources of overdispersion is unavailable. In this paper, we propose a distribution-free alternative and compare its performance with these popular parametric models as well as a moment-based approach proposed by Yu et al. [Statistics in Medicine 2013; 32: 2390-2405]. Like the generalized estimating equations, the proposed approach requires no elaborate distribution assumptions. Compared with the approach of Yu et al., it is more robust to overdispersed zero-inflated responses. We illustrate our approach with both simulated and real study data. [ABSTRACT FROM AUTHOR]

The log-transformation is widely used in biomedical and psychosocial research to deal with skewed data. This paper highlights serious problems in this classic approach for dealing with skewed data. Despite the common belief that the log transformation can decrease the variability of data and make data conform more closely to the normal distribution, this is usually not the case. Moreover, the results of standard statistical tests performed on log-transformed data are often not relevant for the original, non-transformed data. We demonstrate these problems by presenting examples that use simulated data. We conclude that if used at all, data transformations must be applied very cautiously. We recommend that in most circumstances researchers abandon these traditional methods of dealing with skewed data and, instead, use newer analytic methods that are not dependent on the distribution the data, such as generalized estimating equations (GEE). [ABSTRACT FROM AUTHOR]

Exact forms of Taylor expansion for vector-valued functions have been incorrectly used in many statistical publications. We offer two methods to correct this error. [ABSTRACT FROM PUBLISHER]