Unpublished manuscripts cited in Richly Parameterized Linear Models: Additive, Time Series, and Spatial Models Using Random Effects.
Cui Y, Hodges JS, "Smoothing analysis of variance for general designs". This was part of Cui's dissertation that never got published. The most interesting part, to my mind, is Section 3, which identifies the destination of variation that gets smoothed out of the various effects in a balanced ANOVA.
Hodges JS, Clayton MK, "Random Effects Old and New". The material from this paper is mostly in the book's Chapter 13, though various pieces of it are in other chapters, e.g., Chapter 3's discussion of standard errors for fitted splines. Murray Clayton very generously gave me permission to do this, thus killing this paper as a separate publication. At that point, three journals had told us they were only willing to consider it if we completely re-wrote it, and we liked it the way it was (and we still do).
Reich BJ, Hodges JS, Carlin BP, "Spatial analysis of periodontal data using conditionally autoregressive priors having two types of neighbor relations". This is the tech-report version and original submission of Reich, Hodges, and Carlin JASA 2007. It was, in retrospect, an uneasy mix of theory and applications and JASA's editors told us to make it an applications paper, which we did. This version contains a lot of material that didn't make it into the JASA version.
Nick Salkowski's project from my class. The main burden of this project was giving the SemiPar package a good workout, but as part of that Nick re-analyzed the Slovenian stomach-cancer data using a 2-dimensional penalized spline and representing each municipality as a point location (these point locations are given in the dataset available on this website's Datasets page). These results are mentioned in Section 10.1's discussion of spatial confounding.