Yes, this goes way back to the 1970’s. I was at a conference at Carnegie Mellon in 1978 when my professor presented a talk on some Inductive Inference work we had done. Another talk warned about the current toolkits of Statistical Analysis packages (to run on mainframes — it was the 1970’s) and the tendency the statisticians were finding that researchers were using the analysis methods in the packages instead of a more appropriate analysis simply because they were in the package and therefore easy and convenient to run.