Shu Xu
Shu Xu
Scroll
Clinical Associate Professor of Biostatistics
-
Professional overview
-
Dr. Shu Xu’s work represents a balance of both statistical and applied aspects of quantitative methodology. Her primary quantitative interests include evaluating and developing statistical methods for longitudinal data analysis. Specifically, Dr. Xu’s research focuses on various aspects of latent growth models, missing data methods, and causal inference models.
Dr. Xu has collaborated with substance use, family, and health researchers to advance and share her knowledge of quantitative methodology and pursue a better understanding of the social sciences and public health. She has conducted research with the Family Translational Research Group at NYU and the Methodology Center at the Pennsylvania State University.
-
Education
-
BS, Psychology, East China Normal University, Shanghai, ChinaMS, Quantitative Psychology, University of California, DavisPhD, Quantitative Psychology, University of California, Davis
-
Areas of research and study
-
BiostatisticsFamily researchLongitudinal Data AnalysisMissing Data MethodsMixture ModelsQuantitative Research
-
Publications
Publications
Relationships Between E-cigarette Use and Subsequent Cigarette Initiation Among Adolescents in the PATH Study : an Entropy Balancing Propensity Score Analysis
AbstractXu, S., Coffman, D. L., Liu, B., Xu, Y., He, J., & Niaura, R. S. (n.d.).Publication year
2022Journal title
Prevention ScienceVolume
23Issue
4Page(s)
608-617AbstractThis study aimed to examine the relationship between electronic cigarette use and subsequent combustible cigarette use, controlling for confounding by using a propensity score method approach. Data from the first three annual waves of the Population Assessment of Tobacco and Health study were analyzed (n = 6309). Participants were tobacco-naïve at Wave 1; used e-cigarettes exclusively (n = 414), used combustible cigarettes exclusively (n = 46), or not used any tobacco products (n = 5849) at Wave 2. We conducted entropy balancing propensity score analysis to examine the association between exclusive e-cigarette or cigarette initiation and subsequent cigarette use at Wave 3, adjusting for non-response bias, sampling bias, and confounding. Among tobacco-naïve youth, exclusive e-cigarette use was associated with greater risk for subsequent combustible cigarette smoking initiation (OR = 3.42, 95% CI = (1.99, 5.93)) and past 30-day combustible cigarette use (OR = 2.88, 95% CI = (1.22, 6.86)) in the following year. However, the latter risk was comparatively lower than the risk if youth started with a combustible cigarette (OR = 25.79, 95% CI = (9.68, 68.72)). Results of sensitivity analyses indicated that estimated effects were robust to unmeasured confounding. Use of e-cigarettes in tobacco-naïve youth is associated with increased risk of subsequent past 30-day combustible cigarette use but the risk is an order of magnitude higher if they start with a combustible cigarette.Role of social-cognitive factors in the relationship between e-cigarette use and subsequent cigarette smoking among U.S. youth : A causal mediation analysis
AbstractXu, S., Coffman, D. L., Luta, G., Mai, A., Jiang, N., & Niaura, R. S. (n.d.).Publication year
2025Journal title
Addictive BehaviorsVolume
161AbstractObjective: E-cigarette use is associated with subsequent cigarette smoking among youth. The current study examined the mediating role of social-cognitive factors in this association. Methods: Data from four waves of the Population Assessment of Tobacco and Health study (2013/4 – 2017/8) were analyzed. Among youth who had heard about e-cigarettes at Wave 1 but never used cigarettes before Wave 2, we conducted both causal and traditional mediation analyses to examine the mediated effect of social-cognitive factors (including relative harm perception of e-cigarettes versus cigarettes, harm perception of e-cigarette use, perceptions of addictiveness of e-cigarette use, and e-cigarette use among best friends) in the association between e-cigarette use and subsequent cigarette ever or current smoking, adjusting for covariates. We included sampling weights in all analyses; hence, results are generalizable to the U.S. youth (12 – 14 years) from the 2013–2014 cohort. Results: Results from causal mediation analyses indicated that the total effect of e-cigarette use, compared to no use, increased the risk of cigarette ever smoking (20.9 %) and current smoking (4.6 %). A portion of this effect (4.2 % − 15.1 % for ever smoking; less than 10.6 % for current smoking) can be attributed to changes in social-cognitive factors induced by e-cigarette use. However, these mediated effects were small in magnitude relative to their standard errors and not statistically significant. Results from the traditional mediation analyses largely aligned with these findings, except for a few small sized pathways. Conclusions: For the U.S. youth population, social-cognitive factors may only minimally or not at all mediate the association between e-cigarette use and subsequent cigarette smoking. Further investigation into the mediation role of social-cognitive factors is warranted. Tobacco control interventions that focus on cigarette smoking initiation among youth should target other mediating factors."Self-report measures of coercive process in couple and parent-child dyads" : Correction
AbstractMitnick, D. M., Lorber, M. F., Smith Slep, A. M., Heyman, R. E., Xu, S., Bulling, L. J., Nichols, S. R., & Eddy, J. M. (n.d.).Publication year
2022Journal title
Journal of family psychology : JFP : journal of the Division of Family Psychology of the American Psychological Association (Division 43)Volume
36Issue
4Page(s)
522AbstractReports an error in "Self-report measures of coercive process in couple and parent-child dyads" by Danielle M. Mitnick, Michael F. Lorber, Amy M. Smith Slep, Richard E. Heyman, Shu Xu, Lisanne J. Bulling, Sara R. Nichols and J. Mark Eddy (Journal of Family Psychology, 2021[Apr], Vol 35[3], 388-398). In the original article, the full acknowledgment of funding was missing in the author note and should have read "This work was supported by the National Institutes of Health (NIH) Science of Behavior Change Common Fund Program and the National Institute of Dental and Craniofacial Research through an award administered by the National Institute of Dental and Craniofacial Research [1UH2DE025980-01]." The online version of this article has been corrected. (The following abstract of the original article appeared in record 2020-49926-001). One of the most influential behavioral models of family conflict is G. R. Patterson's (1982) coercive family process theory. Self-reports for behaviors related to coercion (e.g., hostility toward a family member) abound; however, there are no self-report measures for coercive process itself, which is, by definition, a dyadic process. Operationalizations of coercive process are measured with behavioral observation, typically including sequential analyzed, microcoded behaviors. Despite its objectivity and rigor, coding of behavior observation is not always feasible in research and applied settings because of the high training, personnel, and time costs the observation requires. Because coercive process has been shown to predict a host of maladaptive outcomes (e.g., parent-child conflict, aggression, negative health outcomes) and given the complete absence of self-report measures of coercive process, we recently designed brief questionnaires to assess coercive process in couple (Couple Coercive Process Scale [CCPS]) and parent-child interactions (Parent-Child Coercive Process Scale [PCCPS]) and tested them via Qualtrics participant panels in samples recruited to mirror socioeconomic generalizability to U.S. Census data. The CCPS and PCCPS exhibited initial evidence of psychometric quality in measuring coercive process in couple and parent-child dyads: Both measures are unifactorial; have evidence of reliability, especially at higher levels of coercive process; and demonstrate concurrent validity with constructs in their nomological networks, with medium to large effect sizes. (PsycInfo Database Record (c) 2022 APA, all rights reserved).Self-report measures of coercive process in couple and parent–child dyads.
AbstractMitnick, D. M., Lorber, M. F., Smith Slep, A. M., Heyman, R. E., Xu, S., Bulling, L. J., Nichols, S. R., & Eddy, J. M. (n.d.).Publication year
2021Journal title
Journal of Family PsychologyVolume
35Issue
3Page(s)
388-398AbstractOne of the most influential behavioral models of family conflict is G. R. Patterson’s (1982) coercive family process theory. Self-reports for behaviors related to coercion (e.g., hostility toward a family member) abound; however, there are no self-report measures for coercive process itself, which is, by definition, a dyadic process. Operationalizations of coercive process are measured with behavioral observation, typically including sequential analyzed, microcoded behaviors. Despite its objectivity and rigor, coding of behavior observation is not always feasible in research and applied settings because of the high training, personnel, and time costs the observation requires. Because coercive process has been shown to predict a host of maladaptive outcomes (e.g., parent–child conflict, aggression, negative health outcomes) and given the complete absence of self-report measures of coercive process, we recently designed brief questionnaires to assess coercive process in couple (Couple Coercive Process Scale [CCPS]) and parent–child interactions (Parent–Child Coercive Process Scale [PCCPS]) and tested them via Qualtrics participant panels in samples recruited to mirror socioeconomic generalizability to U.S. Census data. The CCPS and PCCPS exhibited initial evidence of psychometric quality in measuring coercive process in couple and parent–child dyads: Both measures are unifactorial; have evidence of reliability, especially at higher levels of coercive process; and demonstrate concurrent validity with constructs in their nomological networks, with medium to large effect sizes.Sensitivity analysis of mixed models for incomplete longitudinal data
AbstractXu, S., & Blozis, S. A. (n.d.).Publication year
2011Journal title
Journal of Educational and Behavioral StatisticsVolume
36Issue
2Page(s)
237-256AbstractMixed models are used for the analysis of data measured over time to study population-level change and individual differences in change characteristics. Linear and nonlinear functions may be used to describe a longitudinal response, individuals need not be observed at the same time points, and missing data, assumed to be missing at random (MAR), may be handled. While the mechanism giving rise to the missing data cannot be determined by the observations, the sensitivity of parameter estimates to missing data assumptions can be studied, for example, by fitting multiple models that make different assumptions about the missing data process. Sensitivity analysis of a mixed model that may include nonlinear parameters when some data are missing is discussed. An example is provided.Sensitivity Analysis of Multiple Informant Models When Data Are Not Missing at Random
AbstractBlozis, S. A., Ge, X., Xu, S., Natsuaki, M. N., Shaw, D. S., Neiderhiser, J. M., Scaramella, L. V., Leve, L. D., & Reiss, D. (n.d.).Publication year
2013Journal title
Structural Equation ModelingVolume
20Issue
2Page(s)
283-298AbstractMissing data are common in studies that rely on multiple informant data to evaluate relationships among variables for distinguishable individuals clustered within groups. Estimation of structural equation models using raw data allows for incomplete data, and so all groups can be retained for analysis even if only 1 member of a group contributes data. Statistical inference is based on the assumption that data are missing completely at random or missing at random. Importantly, whether or not data are missing is assumed to be independent of the missing data. A saturated correlates model that incorporates correlates of the missingness or the missing data into an analysis and multiple imputation that might also use such correlates offer advantages over the standard implementation of SEM when data are not missing at random because these approaches could result in a data analysis problem for which the missingness is ignorable. This article considers these approaches in an analysis of family data to assess the sensitivity of parameter estimates and statistical inferences to assumptions about missing data, a strategy that could be easily implemented using SEM software.The belief and modeling of aging
AbstractXu, S., Cui, L. J., Xu, V. S., & Wang, X. J. (n.d.).Publication year
2000Journal title
Chinese Journal of GerontologyVolume
20Page(s)
3Abstract~The impact of the 2014 military conflict in the east of Ukraine and the Autonomous Republic of the Crimea among patients receiving opioid agonist therapies
AbstractMeteliuk, A., Sazonova, Y., Goldmann, E., Xu, S., Liutyi, V., Liakh, T., Spirina, T., Lekholetova, M., Islam, Z., & Ompad, D. C. (n.d.).Publication year
2024Journal title
Journal of Substance Use and Addiction TreatmentVolume
160AbstractBackground: Opioid agonist therapies (OAT) for people with opioid use disorders (OUD) have been available in Ukraine since 2004. This study assessed the effect of 2014 Russian invasion of Ukraine on OAT re-enrollment and retention in conflict areas. Methods: We analyzed the Ukraine national registry of OAT patients containing 1868 people with OUD receiving OAT as of January 2014 in conflict areas (Donetsk, Luhansk, and the Autonomous Republic [AR] of the Crimea). We developed logistic regression models to assess the correlates of re-enrollment of OAT patients in government-controlled areas (GCA) from conflict areas and retention on OAT at 12 months after re-enrollment. Results: Overall, 377 (20.2 %) patients were re-enrolled at an OAT site in a GCA from confict areas, of whom 182 (48.3 %) were retained on OAT through 2021. Correlates of re-enrollment were residing in Donetsk (adjusted odds ratios (aOR) = 7.06; 95 % CI: 4.97–10.20) or Luhansk (aOR = 6.20; 95 % CI: 4.38–8.93) vs. AR Crimea; age 18–34 (aOR = 2.03; 95 % CI: 1.07–3.96) or 35–44 (aOR = 2.09; 95 % CI: 1.24–3.71) vs. ≥55 years, and being on optimal (aOR = 1.78; 95 % CI: 1.33–2.39) or high OAT dosing (aOR = 2.76; 95 % CI: 1.93–3.96) vs. low dosing. Correlates of retention were drug use experience 15–19 years (aOR = 3.69; 95 % CI: 1.47–9.49) vs.The lump-versus-split dilemma in couple observational coding : A multisite analysis of rapid marital interaction coding system data.
AbstractHeyman, R. E., Otto, A. K., Reblin, M., Wojda, A. K., & Xu, S. (n.d.).Publication year
2020Journal title
Journal of Family PsychologyVolume
35Issue
4AbstractHistorically, observational couple communication researchers have oscillated between splitting behaviors into narrowly defined discrete codes and grouping behaviors into broader codes—sometimes within the same study. We label this the “lump-versus-split dilemma.” Coding across a decade and 11 investigators were used to recommend the most meaningful number of codes to use when observing couples’ conflict. We combined data from 14 studies that used the Rapid Marital Interaction Coding System (RMICS) to score communication behavior during different-sex couples’ conflict interactions. In each study, couples completed at least one 10-min, video-recorded conflict discussion. Communication during these interactions was coded by trained research staff using RMICS; all codes were compiled into a single data set for descriptive analysis and exploratory factor analyses (EFAs). The final sample comprised N = 2,011 couples. Several RMICS codes were extremely infrequent—specifically, distress-maintaining attributions, psychological abuse, withdrawal, dysphoric affect, and relationship-enhancing attributions. By far, the most frequent code was constructive problem discussion. EFAs yielded two factors for both women and men. Factor 1 (Negative) contained two items: distress-maintaining attributions and hostility. Factor 2 (Nonnegative) contained constructive problem discussion and humor (and, for women only, acceptance). Results side heavily with the “lump” camp in the lump-versus-split dilemma in couple observational coding. These RMICS factor analysis results converge with those from other systems and imply that the microanalytic “splitting” era in couples coding should draw to a close, with future studies instead focused on negative, neutral, and positive codes. (PsycInfo Database Record (c) 2020 APA, all rights reserved)The Mediating Effect of E-Cigarette Harm Perception in the Relationship between E-Cigarette Advertising Exposure and E-Cigarette Use
AbstractJiang, N., Xu, S., Li, L., El-Shahawy, O., Freudenberg, N., Shearston, J. A., & Sherman, S. E. (n.d.).Publication year
2022Journal title
International journal of environmental research and public healthVolume
19Issue
10AbstractExposure to e-cigarette advertising is associated with e-cigarette use among young people. This study examined the mediating effect of e-cigarette harm perception on the above relationship. Cross-sectional survey data were collected from 2112 college students in New York City in 2017–2018. The analytic sample comprised 2078 participants (58.6% females) who provided completed data. Structural equal modeling was performed to examine if e-cigarette harm perception mediated the relationship between e-cigarette advertising exposure (via TV, radio, large signs, print media, and online) and ever e-cigarette use and susceptibility to e-cigarette use. About 17.1% of participants reported ever e-cigarette use. Of never users, 17.5% were susceptible to e-cigarette use. E-cigarette advertising exposure was mainly through online sources (31.5%). Most participants (59.4%) perceived e-cigarettes as equally or more harmful than cigarettes. Advertising exposure showed different effects on e-cigarette harm perception depending on the source of the advertising exposure, but perceiving e-cigarettes as less harmful than cigarettes was consistently associated with e-cigarette use and susceptibility. Low harm perception mediated the association between advertising exposure (via online, TV, and radio) and ever e-cigarette use and between online advertising exposure and e-cigarette use susceptibility. Regulatory actions are needed to address e-cigarette marketing, particularly on the Internet.Tutorial on Causal Mediation Analysis With Binary Variables : An Application to Health Psychology Research
AbstractXu, S., Coffman, D. L., Luta, G., & Niaura, R. S. (n.d.).Publication year
2023Journal title
Health PsychologyAbstractMediation analysis has been widely applied to explain why and assess the extent to which an exposure or treatment has an impact on the outcome in health psychology studies. Identifying a mediator or assessing the impact of a mediator has been the focus of many scientific investigations. This tutorial aims to introduce causal mediation analysis with binary exposure, mediator, and outcome variables, with a focus on the resampling and weighting methods, under the potential outcomes framework for estimating natural direct and indirect effects.We emphasize the importance of the temporal order of the study variables and the elimination of confounding. We define the causal effects in a hypothesized causal mediation chain in the context of one exposure, one mediator, and one outcome variable, all of which are binary variables. Two commonly used and actively maintained R packages, mediation and medflex, were used to analyze a motivating example. R code examples for implementing these methods are provided.Use of electronic nicotine delivery system (ENDS) devices among U.S. Youth and adults : Findings from the Population Assessment of Tobacco and Health Study Waves 1–5
AbstractJiang, N., Xu, S., Li, L., Cleland, C. M., & Niaura, R. S. (n.d.).Publication year
2023Journal title
Addictive BehaviorsVolume
139AbstractIntroduction: Electronic nicotine delivery system (ENDS) devices evolve rapidly and impact nicotine dependence. This study described the type of ENDS devices used most frequently by U.S. youth and adults from 2013/14 to 2018/19. Methods: We analyzed Waves 1–5 data of the Population Assessment of Tobacco and Health Study. Among current ENDS users, descriptive statistics summarized the most frequently used ENDS devices (i.e., disposable cigalike, refillable cartridge, nonrefillable cartridge, tank, mod, prefilled pod, disposable pod) among youth (12–17 years), young adults (18–34 years), and older adults (≥35 years) for each wave. Results: The proportion of current ENDS users who reported they most frequently used disposable cigalikes and cartridge-based devices declined over time across all age groups. At Waves 1–4, tank was generally the most popular type for all ages and an increasing proportion of ENDS users reported they most frequently used tanks. The primary use of mods decreased among youth, and fluctuated among young and older adults. At Wave 5, prefilled pods became the dominant type (youth: 55.0%; young adults: 44.7%; older adults: 42.7%), and 4.2–10.0% of ENDS users reported using disposable pods most often. The popularity of tanks, mods, and prefilled pods was more evident in youth and young adults, and primary use of disposable pods was more common in older adults. Conclusions: The primary use of ENDS devices changed over the years and varied by age. More research is warranted to continuously monitor the characteristics of ENDS devices in youth and adults to inform product regulations and intervention efforts.Using Security Questions to Link Participants in Longitudinal Data Collection
AbstractXu, S., Chan, A., Lorber, M. F., & Chase, J. P. (n.d.).Publication year
2020Journal title
Prevention ScienceVolume
21Issue
2Page(s)
194-202AbstractAnonymous data collection systems are often necessary when assessing sensitive behaviors but can pose challenges to researchers seeking to link participants over time. To assist researchers in anonymously linking participants, we outlined and tested a novel security question linking (security question linking; SEEK) method. The SEEK method includes four steps: (1) data management and standardization, (2) many-to-many matching, (3) fuzzy matching, and (4) rematching and verification. The method is demonstrated in SAS with two samples from a longitudinal study of adolescent dating violence. After an initial assessment during a laboratory visit, participants were asked to complete an online assessment either (a) once, 3 months later (Sample 1, n = 60), or (b) three times at 1-month intervals (Sample 2, n = 140). Demographics, eye color, and responses to nine security questions were used as key variables to link responses from the laboratory and online follow-up assessments. The rates of matched cases were 100% in Sample 1 and from 94.3 to 98.3% in Sample 2. To quantify the confidence in the data quality of successfully matched pairs, we reported the means and standard deviations of the number of matched security questions. In addition, we reported the rank order and counts of the mismatched components in key variables. Results indicate that the SEEK method provides a feasible and reliable solution to link responses in longitudinal studies with sensitive questions.Youth E-Cigarette Use and Functionally Important Respiratory Symptoms : The Population Assessment of Tobacco and Health (PATH) Study Waves 3 and 4
AbstractStevens, E. R., Xu, S., Niaura, R. S., Cleland, C. M., Sherman, S. E., Mai, A., Karey, E., & Jiang, N. (n.d.).Publication year
2022Journal title
International journal of environmental research and public healthVolume
19Issue
22AbstractRespiratory effects of e-cigarette use among youth are not fully understood. This study investigated the longitudinal association between e-cigarette use and a validated index of functionally important respiratory symptoms among US youth. Data from Waves 3–4 of the Population Assessment of Tobacco and Health Study were analyzed. The sample included youth (aged 12–17) without asthma at baseline (Wave 3), who completed a follow-up survey (Wave 4), and were not missing data for analytic variables (n = 3899). Exposure was e-cigarette use status (never, former, or current) at baseline. The outcome was a respiratory symptom index based on responses for seven wheezing items at Wave 4. An index of ≥2 was defined as having functionally important respiratory symptoms. Lagged logistic regression models examined the association between baseline e-cigarette use and functionally important respiratory symptoms at follow-up by combustible tobacco use status (never or ever), and controlling for baseline covariates. At baseline, 13.7% of participants reported former e-cigarette use, and 4.3% reported current use. Baseline e-cigarette use did not increase the odds of having functionally important respiratory symptoms at follow-up regardless of combustible tobacco use status. Future research on larger populations of e-cigarette users with longer follow-up periods will improve our understanding of the respiratory risks associated with e-cigarette use among youth.老年人的养老观念与养老模式
AbstractXu, S., & Xu, V. S. (n.d.).Publication year
2000Journal title
中国老年学杂志Abstract目的 探讨老年人的养老观念与较理想的养老模式.方法 采用自编调查表对上海市居家养老老人,养老院或老年公寓养老老人以及老年人的子女进行了调查.结果 (1)老年人“养儿防老”的观念正在发生转变,且这一养老观念的转变与老年人的性别和文化程度无关.(2)养老观念的变化使老年人的幸福感指向的对象也发生了变化,不再认为一定要与子女住在一起安度晚年才能幸福.由此导致老年人对养老居住方式有了不同的选择.