Allison Tracy has over 15 years’ experiences providing methodological and statistical consultation for researchers in a wide variety of disciplines, research topics, and institutions – academic, applied, and commercial. Her approach to consulting is to translate researchers’ articulated research questions and hypotheses into statistical models and to translate results of these models back into plain English that can be understood by individuals both within and outside academia.
She has technical expertise in a wide range of statistical techniques used in the social sciences, including structural equation modeling, confirmatory factor analysis and MIMIC approaches to measurement, path modeling, regression analysis (e.g., linear, logistic, Poisson), latent class analysis, hierarchical linear models (including growth curve modeling), latent transition analysis, mixture modeling, item response theory, as well as more commonly used techniques drawing from classical test theory (e.g., reliability analysis through Cronbach’s alpha, exploratory factor analysis, uni- and multivariate regression, correlation, ANOVA, etc). She also has expertise in missing data analysis and power analysis. She has a strong background in program evaluation and measurement development. She is currently expanding her expertise to include Rasch modeling and Generalizability Theory approaches to measurement.
Her role at the Wellesley Centers for Women was to provide statistical and methodological advice and support to WCW researchers, assist in the development of grant proposals, conduct statistical analyses, and pursue her own substantive and methodological interests. Recently, Dr. Tracy headed a research program exploring ways to empirically identify distinct social classes among teens, emerging adults, and adults and typical patterns of change in social class over time as teens establish their own households. Her approach was to model social class using multiple indicators of class, including but not limited to income and wealth, home ownership, debt, educational level, occupational prestige, social networks, leadership and volunteer activities. The theoretical framework informing her choice of analysis models reflected an ecological approach in which social class status is multiply-determined by individuals, their immediate social contexts such as families and households, as well as other relevant contexts such as schools, neighborhoods, and workplaces. She hypothesized that social class and change in social class over time have implications for health behaviors, lifestyle choices, and health problems in emerging adulthood and beyond.