A scientific paper follows the classic literary plot structure. Each section follows in sequence from the previous sections, so that no individual section (with the exception of the Introduction) can be fully understood without having read the previous ones. If you pick up a novel and read the last page first, you might find out whether Jack dies, but you won’t have any idea who killed him and why. Those details are important.
In terms of understanding the Results and Discussion sections of a paper, the Methods section is critical. Results should never be read as a standalone text. The only way you, the reader, can judge if my results are valid and meaningful is if you know how I collected and analysed the data.
The sexy summary sentence in an empirical paper’s abstract doesn’t necessarily apply to everywhere and everything – there’s a context. Which is why journals that hide the Methods at the end of the paper, or in supplementary material, are doing Science a huge disfavour.
The other important function of the Methods is to explain all the variable names, terms and statistical notation that will be encountered in the results, and probably during the discussion. This is especially important for ecological studies, where commonly used metrics like ‘richness’ and ‘diversity’ have multiple meanings and can be measured and calculated in many different ways. The Methods is like a specialised dictionary to set the reader up with the language needed to interpret the results and discussion. Because it’s a standalone section, with its own unique function, it ensures that the Results and Discussion are a lot easier to read.
For example, say I count the number of flowering plant species within a 1m x 1m quadrat as a measure of flowering plant richness at each site. It’s a lot easier to explain this once in the Methods and then refer to the metric as ‘floral richness’ for the rest of the paper, rather than writing (and making you read) ‘total number of flowering plant species in a 1m x 1m quadrat’ every time.
This is a scientific standard that most students are taught in high school. The first mention of a metric or statistical test contains the long detail with its associated simple notation, and then the author can safely use the simple notation for the rest of the report knowing their reader understands what’s going on.
So it surprises me when a reviewer expects me to repeat the long description of variables and details for my entire paper. Recently a reviewer highlighted a sentence in my Results where I had used the term ‘bee richness’ (after explaining it’s meaning in the Methods). They asked “What kind of richness? You need to clarify which metric was used!” complete with the exclamation mark. They then picked up the same term a few more times throughout the rest of the manuscript, each time asking me to specify which metric I was using.
This tells me they didn’t read my Methods, where I clearly explained all details of all metrics used. Alternatively, if they did read my Methods, are they suggesting that I need to spell everything out for the whole paper, for the benefit of the people who don’t want to read the Methods? I hope not!
What do you think? Do you read or skim Methods sections? Should specific terms and variable names be explained once, or with each mention?
© Manu Saunders 2017
One explanation in the Methods ought to be sufficient (as a grad student who, admittedly, skims the methods until I come across a term I need to go back and look for!)–but for papers with a lot of variables/terms, especially computational work, I think a reminder or two in the discussion is very helpful. (Honestly, for mathematical modeling where no two people are able to set aside ego long enough to use the same variable for the same thing, is a table of variables too much to ask? Sorry, rant over.)
Or at least in figure captions, if they’re busy figures.
LikeLiked by 1 person