Proposal Description

Ben Parker examines the role of optimization in the design of experiments on networks — an increasingly important setting as A/B testing, clinical trials, and field experiments are conducted on social, logistical, and biological networks where treatment spillovers create interference between units. The talk explores how constrained optimization, integer programming, and combinatorial algorithms can be used to solve treatment assignment problems in a way that maximizes statistical power and minimizes estimator variance, while accounting for network structure. Applications in agricultural field trials and industrial marketing experiments ground the methodological contributions in practical contexts where network interference is a genuine challenge.

Byran Smucker focuses on the coordinate-exchange algorithm, one of the most widely used heuristics for constructing optimal exact experimental designs, and extends its application to supersaturated experiments in which the number of factors exceeds the number of runs. Such designs are common in screening experiments where resources are limited and the goal is to identify active factors efficiently. The talk provides an accessible overview of coordinate exchange and demonstrates its use in designing supersaturated experiments in a high-throughput screening drug discovery context.

Aldo Corbellini examines optimization in robust statistical modelling through the monitoring of regression fits and transformation choice. Their work focuses on tuning decisions that balance protection against contamination with efficiency under the assumed model. By replacing a single robust fit with monitoring trajectories indexed by subset size or breakdown point, the proposed framework provides insight into the stability of parameter estimates, outlier detection, and model selection. The study further explores the role of Box–Cox and Yeo–Johnson transformations, offering practical tools for reproducible and transparent statistical analysis.

Matilde Bini brings the session’s optimization theme into the domain of comparative labour market policy. Their study investigates the relationship between minimum wage intensity — measured via the Kaitz index — and youth labour market exclusion, captured by NEET rates, across 18 European countries from 2005 to 2024. The empirical strategy deploys dynamic fixed-effects models with Driscoll–Kraay standard errors, country-specific trend specifications, and Dynamic Common Correlated Effects estimators to address the econometric challenges of persistence, cross-sectional dependence, and unobserved common factors. The study contributes to the literature on minimum wages and youth inclusion by examining how the relationship varies across wage-setting regimes and institutional contexts.