The above remark, from Alan Dunne, applies to mature fields more than to new fields. For example, I guess the textbooks on deep learning are pretty recent, so anything a few years old really would be out of date. Even in subfields that have been around for awhile, it can take a while for textbook writing to settle down. The first textbook on Bayesian data analysis is 25 years old, but people such as Richard McElreath are still figuring out new ways of presenting and extending these ideas. For more mature and higher-volume topics, though, sure, I think Dunne has a point. Textbooks in intro American Politics keep coming out in their 9th or 13th editions or whatever, and don’t get me started on intro stat textbooks.
Dunne was reacting to this post by David Myers from last year, “Psychological Science: Full of Surprises?” Dunne writes:
Myers is I think a reasonable representative of the commonplace or “conventional wisdom” in psychology (he is a textbook writer) giving “findings of psychology” divided into categories of increasing surprisingness.
It could be interesting to look into the evidence for each asserted discovery, and particularly whether there were patterns of amounts of evidence by level of surprisingness. I do not have the spare time and energy to do that and do not suppose you do but if you post the link or Myers’ post itself perhaps someone in search of a project or a new angle on the state of psychological research will take it on.
I would also be interested to read your thoughts on textbooks generally (besides advertising your own, of course), how they can be done better, and how they can avoid presenting jumped-to conclusions as fact (and having readers remember them as facts) and give a fair overview of what is known, and indication that things are mostly not known, about a subject. For example your (or your commenters’) thoughts on the effects of and any opportunities arising from, textbook publishing being a field marked by constant turnover, so that textbooks more than a very few years old cannot even be given away, but in which new textbooks are mostly made by copying from former ones. The impact of textbooks on the general public’s thinking about specific things and about sciences about humans as a whole could also be interesting to explore.
I think a good textbook should have as at least one base for it a systematic sample of the systematic reviews found in databases of research in the field—or of metaanalyses or whatever the closest equivalent tagged in the subject’s databases are; I believe even the most backward disciplines usually have literature reviews or review articles.
In his brief post, Myers lists 34 successes of psychology research. 34 is a lot! His list includes 12 “unsurprising but important findings (significant facts of life for our students to understand),” 11 “surprising findings that may challenge our beliefs and assumptions,” and 11 “surprising findings [that] reveal things unimagined.”
I don’t agree with 100% of Myers’s examples, but I like how he focuses on important real-world topics rather than things like Stroop which might be important for our scientific understanding but is not a big deal in most of our lives. The main thing that seems to be missing from Myers’s lists is developmental psychology. I guess that’s just not a topic that interests him very much.
Regarding Dunne’s question about the textbooks: Yeah, I dunno. Once a misconception gets in, it can be hard to root it out.