Jeff Smith College Quality and the Garden of Forking Paths Abstract: Empirical economics papers report standard errors that quantify the uncertainty associated with sampling variation but rarely consider non-sampling variation in a systematic way. Nonsampling variation arises from researcher study design choices related to measurement of key variables, functional form, tuning parameters, model selection procedures, and so on. This paper documents the current state of play regarding non-sampling variation and describes approaches from inside and outside of economics to quantify such variation more systematically. We provide a worked example in the form of an analysis of the labor market and educational effects of college quality on degree completion and earnings using data from the National Longitudinal Survey of Youth 1997 cohort (NLSY-97). In our analysis, we consider multiple ways to create our college quality index, multiple ways to deal with item non-response in our conditioning variables, multiple ways to code our earnings outcome measures, and multiple ways to choose a specification for our conditioning variables. We find that in this context, in which sampling variation matters a lot due to the relatively small sample size of the NLSY-97 and relatively high residual variance of our outcomes, these dimensions of non-sampling variability imply uncertainty in our parameter estimates on a par with the sampling variation. Co-authors: Heather Little & Lois Miller
The Dindigul Agreement stands as a pioneering framework designed to eliminate gender-based violence and harassment in apparel factories. Established among global brands (H&M, Gap Inc., and PVH Corp), an international NGO (Global Labor Justice), a union alliance (Asia Floor Wage Alliance), a local union (Tamilnadu Textile and Common Labor Union), and an Indian apparel manufacturer (Eastman Exports), this agreement is widely recognized as a potential model for the apparel industry, where a majority of employees are young women and gender-based violence remains a persistent challenge. As the agreement reaches its conclusion, this webinar from the Cornell ILR Global Labor Institute will present the first official evaluation and explore a crucial question: Is the Dindigul Agreement achieving its intended impact? Our panel will examine the effectiveness of this collaborative effort between factory management and unions in addressing gender-based violence and harassment in the workplace and discuss its implications for the broader apparel industry. Register Now! What You'll Learn How the Dindigul Agreement has performed and what metrics define its successKey factors driving the agreement's effectiveness and best practices for implementationStrategies for adapting and scaling this model across the apparel industryWays this framework can strengthen corporate human rights due diligence effortsEvidence-based insights into how the agreement empowers women to combat gender-based violence both in the workplace and their communities
