Near-synonyms are not the same word, and in academic writing the gap can move a claim. A 2021 paper in the Journal of Bone and Joint Surgery by Neville Varady and colleagues found causal language misused in 60% of a random sample of 400 observational orthopaedic studies. The audits count wording rather than reader outcomes, and that limit shapes everything below.
Word choice is also a craft habit that formal training often skips. Where supervision fits into that learning is its own topic, covered in this look at how the advisor relationship shapes a doctorate. The three named studies below show what slips when precision fails.
Two of the audits track causal wording and one tracks hype in abstracts. More pieces on writing and academic work sit in the academia section. Together they give a style question real numbers, which is rare for writing advice.
Which Word Swaps Change a Claim?
Cause is the risky zone. Proves and causes assert a mechanism; suggests and is associated with record a pattern only. The JBJS commentary by Varady and colleagues notes that the word association can almost always replace a causal phrase when the design is observational, because causation generally needs randomization. Boosters deserve the same care, since they inflate certainty without adding data.
Softer pairs carry the same trap. Effect presumes a produced change, and impact often brings causal intent into observational text, which is the drift the audits measured. Method and technique drift less, yet they still blur when a methods section must name one exact procedure.
What Do the Numbers Say?
A 2022 survey in the Journal of Clinical Epidemiology, led by Mina A. Han with Gordon Guyatt’s team at McMaster University, randomly selected 199 systematic reviews of observational studies from 2019. The draw was stratified by use of the GRADE evidence-rating approach, and the search spanned EMBASE, MEDLINE, and Epistemonikos. Among reviews with causal intent, 78.8% used causal language somewhere, while only 4.4% kept it consistent. More than half of the reviews without causal intent, 51.2%, still used causal words.
Hype follows the same drift. A 2015 analysis in The BMJ by Christiaan Vinkers, Joeri Tijdink, and Willem Otte scanned PubMed abstracts from 1974 to 2014. Positive words rose from 2.0% of abstract text to 17.5%, a relative jump of 880%. Words such as novel and unprecedented led the climb, and the pattern held at high-impact journals. Books over the same period showed no matching rise, which points at journal culture rather than language drift.
Where Do These Audits Fall Short?
Each study covers a slice, not the whole. Varady’s sample sat in one surgical field, and the paper is a commentary with its own sample, so its 60% figure is a field snapshot. Han’s survey covered reviews from a single year, and Vinkers tracked abstracts rather than full papers.
None of the designs can tie wording to harm. They record how often language overreaches; they do not test whether a hardened verb changed a clinical decision. A research synonym list cannot make that call either, because the swap is an argument about evidence, not spelling. The audits end in the same place: the design, not the dictionary, sets the word.
Conclusion: Match the Verb to the Design
The rule from the evidence is plain: word strength should track design strength. Observational data earns association, randomized data can carry cause, and abstracts deserve fewer superlatives than most now hold. The three audits give that rule numbers instead of taste. Readers following this should also see The Two-Tier Faculty: What Tenure-Track and Adjunct Economics Look Like.
For a writer, the practical step is to name the design first and pick the verb second. For a reader, the step is to treat any causal verb in an abstract as a claim to check against the methods section, not as settled fact. Precision here is not polish. It is the claim itself.
