The paper employs data from 2884 matches in the English Football League Championship. It builds a model of determinants of attendance designed to yield results relevant to decision-taking at individual clubs. The model has two innovatory features. It controls for the market size of home and away teams precisely by including local population measures constructed from the application of GIS software and information on competition from other clubs. It incorporates these time-invariant covariates in a Hausman–Taylor random effects estimator to take explicit account of variables typically excluded in earlier studies based on fixed effects models. Unlike fixed effects results, Hausman–Taylor estimates permit assessment of the role of market size and quality of the playing squad in determining attendance. Results also quantify the reduction in attendance from televising a match and show that attendance diminishes when a match is played simultaneously with a televised game in a higher status competition.