Wikipedia describes Freedman’s paradox briefly as
a problem in model selection whereby predictor variables with no explanatory power can appear artificially important
with a reference to a 1983 paper by David Freedman.
The term "Freedman's paradox" appears in the title of Model selection bias and Freedman’s paradox by Lukacs et al. from 2010. They write
Freedman (1983) demonstrated that variable selection methods based on testing for significant $F$ statistics commonly include explanatory variables with no relation to the response and spuriously inflate $R^2$ when such irrelevant variables are present. This problem is often referred to as “Freedman’s paradox.” Freedman’s paradox results in spurious effects appearing important.
I haven't been able to find other sources that refers to the problem with variabel selection as Freedman’s paradox, and I am curious about the origin of the terminology.
My question is therefore if anybody knows earlier references than Lukas et al. (2010) that use the term?
Edit: Following @Tim's advice (searching on Google Scholar), I found the technical report Model selection and accounting for model uncertainty in linear regression models by Raftery et al. from 1993, where they write
In the extreme case where there are many candidate predictors but there is no relationship between any of them and the response, standard variable selection procedures often choose some subset of variables that yields a high $R^2$ and a highly significant overall $F$ value. We refer to this unfortunate phenomenon as "Freedman's Paradox" (Freedman, 1983).
This may be the origin of the term, but I welcome any additional early references and other comments about its usage.