Writing a paper, results and discussion – day 5: evaluating results with certainty or moderation
Free PreviewWords for certainty and moderation
In some cases it will be correct to write evaluations with real certainty when you are describing things that everyone is already agreed on or results that you have found that are unequivocal. However, in most cases of new research that is breaking boundaries and changing our understanding, the results will not be entirely unequivocal. Questions around the findings will likely remain and more research will need to be done to help to complete our understanding. Therefore, it is important to appropriately moderate the language that we use to evaluate and highlight the findings in scientific research – this makes the writing more accurate. For example, rather than stating:
- These findings show that dogs only see in black and white.
A researcher may prefer to write:
- These findings indicate that dogs only see in black and white.
Sentence 2. may be more accurate that sentence 1. and so it is better to add the word ‘indicate’ because the results of the experiment cannot be seen as 100% proof of this evaluation.
Words of clarity
There are certain adjective and adverbs that indicate certainty such as:
clearly, evidently, definite, conclusive, undoubtedly, show, obvious
These results show conclusively that electric cars are the best.
There are also verbs that on their own tell a story of complete certainty:
be (is/are), demonstrates, proves, means, equals
These results demonstrate that electric cars are the best.
Adding moderation
Instead of using words that give total certainty, you could use adverbs and phrases that are more moderate and allow the possibility of other interpretations:
Relatively, possibly, likely, to some extent, somewhat, mostly, in the main, slightly, apparently
These results show that electric cars are likely to be the best for the environment.
The findings demonstrate a likely correlation between [x] and [y].
The findings show an apparent link between [x] and [y].
There are also several words that can be used to moderate the use of verbs:
may, probably, likely, tend, seems, appears, indicate, suggest,
These results seem to indicate that…
The cause of this may be that…
It seems likely that there is a correlation…
Have a read through the extract below for some examples.
Lipid levels in midlife and risk of atrial fibrillation over 3 decades—Experience from the Swedish AMORIS cohort: A cohort study
Discussion
In this large population-based study with more than 65,000 cardiovascular healthy individuals at baseline, aged 45 to 60 years, and followed for up to 35 years, a lower risk of AF in association with high TC and LDL-C levels was observed. However, these associations were only present during the first 5 years after blood measurement. For HDL-C and ApoA-I, low levels were associated with an increased risk of AF, and this did not change with time of follow-up. The association between a high TG and TG/HDL-C ratio and higher risk of AF also persisted over the follow-up. ApoB/ApoA-I ratio was not associated with AF, except for elevated ratio (≥0.94) in untreated participants with incident HF/CHD. The association between lipids and AF does not seem to differ by co-occurring HF/CHD.
…..
To our knowledge, our study is the first to examine the longitudinal relationship between apolipoprotein measures and incident AF. The ApoB level represents the total number of atherogenic particles in blood including cholesterol- and TG-containing lipoprotein particles (very low-density lipoprotein, intermediate-density lipoprotein, and LDL). ApoB/ApoA-I ratio has been shown in many reports to be a more informative and robust measure in predicting CHD than other lipid fractions [8,9]. In our study, ApoB and ApoB/ApoA-I ratio was not associated with incident AF, except for low ApoB (<1.05 g/L) and ApoB/ApoA-I ratio (<0.42) that was associated with higher AF risk among those who developed incident HF/CHD. One explanation could be that they developed AF on the basis of HF/CHD or other concomitant diseases associated with low cholesterol, such as cancer [19,20].
Extract is taken from: Ding M, Wennberg A, Gigante B, Walldius G, Hammar N, et al. (2022) Lipid levels in midlife and risk of atrial fibrillation over 3 decades—Experience from the Swedish AMORIS cohort: A cohort study. PLOS Medicine 19(8): e1004044. https://doi.org/10.1371/journal.pmed.1004044
Useful phrases from the extract
these associations were only present: ‘were only’ highlights a limited link
low levels were associated with… …The association between: ‘associated’ as a word to link two factors is a word that shows less than 100% certainty
does not seem to differ: ‘seem to’ adds moderation here
To our knowledge: the authors are admitting that there could be other studies and so moderates their claim to be te first to show this
ratio has been shown in: this is a moderate way of introducing the idea that follows
except for: highlights a known exception to the claim
One explanation could be: shows there are different explanations
Further study for this week
Hopefully this week you have been able to write a discussion section describing the main findings of a recent study. Try the short quiz below to test your understanding.