The New APA Style: Try to Contain Your Excitement—and Watch Out for Dud Copies

This is a post for the nerds, fine people that we are. Actually, everyone needs to think about reporting style, especially for statistical stuff.

The APA Publication Manual is the bible of APA Style, also perhaps the bane of some students’ lives. In ITNS we used the 6^th edition (APA, 2010). Now there’s a 7^th edition (APA, 2020). I guess we need to switch for ITNS, second edition, tho’ I’m finding it hard to get enthusiastic.

Figures: captions out, headings in

For me, the biggest change is the demise of detailed figure captions placed below the figure. Now there’s a brief heading above a figure, and Note below to give the description. That matches style for tables, but I’m grieving already.

The OLD style for figures, as in ITNS

Figure 1.1. Support for Proposition A, in percent, as reported by the poll. The dot marks the point estimate, and the two lines display the margin of error (2%) either side of the dot. The full interval, from 51% to 55%, is the 95% confidence interval.

The NEW style for figures, as in ITNS2?

Figure 1.1

Support for Proposition A, as Reported by the Poll

Note. Support is expressed in percent. The dot marks the point estimate, and the two lines display the margin of error (2%) either side of the dot. The full interval, from 51% to 55%, is the 95% confidence interval.

What else is new in the 7^th edition

A guide to the changes from 6^th to 7^th editions is here. A video that illustrates the main changes is here. It shows how the new style is a bit simpler—hooray—despite the new edition being more detailed and having more than 50% more pages than the old.

There are small changes to reference formats and how references are cited in text. Reference formats, and examples, are provided for more than 100 types of items, including social media items, blog posts and comments on blog posts, TED talks, songs, and just about every other type of item you can think of.

For the first time, format for student papers is discussed, with great scope given for instructors to specify how student work should be presented. Good.

JARS, the Journal Article Reporting Standards, are highly detailed, and now comprise a whole chapter (Chapter 3). As well as quantitative, they cover qualitative and mixed methods research. There’s particular mention of meta-analyses, and replication studies. Yay.

Only the super-nerds need notice that now only one space is required at the end of a sentence, not two. A massive saving of virtual trees?

Open Science?

There’s brief mention of Open Science badges, but the index includes no other item for ‘Open Science’. There’s no index entry for ‘preregistration’, but the ‘registration’ entry points to a couple of mentions of study registration, in particular of clinical trials. There’s advice on how to include access information about open data and open materials provided online. Overall, however, there is no strong advocacy of Open Science practices. Sadly.

The new Publication Manual: Beware counterfeits

Last year I bought a copy of the 2020 Manual on the Australian site of Amazon. I noticed small errors, for example incorrect italics. I enquired of APA, sent them photos of a few pages, and was told I had a counterfeit copy. It came from a third-party seller via Amazon. I returned the copy to Amazon, as requested, and gave the seller a blistering review—they even had the cheek to ask me to withdraw this as spoiling their business. Happily, APA sent me a replacement copy, although I suspect that’s not usual practice.

There’s now a blog post warning of counterfeits, as well as advice here and here on spotting duds.

Are journals using the new style?

Not surprisingly, APA journals are using the new style. The change is in progress. Skim through recent issues and see a mixture of articles using the old style for figures, and those using the new. For APS journals, the online guidelines for authors still refer to the old 2010 style. I enquired of Elaine Walker, Chair of the APS Publications Committee, who kindly explained that APS intends to swap to the 7^th edition, and is currently planning the move. So, I guess Bob and I need to switch. Sigh. I’m already nostalgic for fully explanatory captions under figures.

Why we need reporting standards

Years ago some students and I scanned articles in economics, business, and chemistry journals to see how these disciplines reported statistical significance testing. This was pilot testing, never written up. We found that reporting was often sketchy and obscure. We saw text such as: “Figure 1 illustrates that only groups Groups A and C showed an effect.” It took investigation and guesswork to figure out that the authors had probably used a criterion of 2xSE to identify effects as existing or not. Dichotomous decision making, without even coming clean, stating the criterion being used, and admitting that any difference just less than 2xSE was regarded as not existing.

In Psychology, long-standing APA style requirements have meant that NHST is almost always more full reported. We should be told at least the summary statistics, what’s being tested, the df, the p value, and then a conclusion.

Now that researchers are using computer scanning of very large numbers of journal articles to study, for example, the use of NHST and the distribution of reported p values, it’s more vital than ever that authors report fully and comply with style standards in reporting statistical analyses.

At least to some extent, we all need to embrace our inner nerd. The 2020 version of course.

Geoff