Academic writing is not complicated writing. It is precise writing — every claim is supported, every hedge matches the strength of the evidence, every term is used consistently. This article identifies the features that mark a passage as academic and how to apply them.
Sign in to mark this article as read and track your progress.
Academic writing is not difficult vocabulary and long sentences. It is a set of conventions that serve a specific purpose: to make arguments that are clear enough to be evaluated, specific enough to be verified, and honest enough to acknowledge what they do not settle.
Four features mark a passage as academic:
Formal register — academic writing avoids contractions (don't → do not), colloquialisms (a lot of → many; got → obtained), vague quantifiers (some, lots, many without numbers), and first-person opinion markers used as evidence ("I think," "I believe"). This is not about sounding impressive; it is about keeping the reader's attention on the evidence rather than the writer's personality.
Evidence-based claims — every claim in an academic paper is either a result the paper itself demonstrates, or a claim from a cited source. Unsupported assertions do not belong in academic writing. "Phishing is a major threat" is an assertion. "Phishing accounted for 36% of all data breaches in 2023 (Verizon DBIR, 2023)" is a supported claim.
Precise attribution — academic writing names its sources specifically. "Studies show that MFA is effective" names no study and gives the reader no way to verify the claim. "Li et al. (2022) found that SMS-based MFA reduced successful phishing credential harvesting by 78% across 14,000 enterprise accounts" is verifiable.
Hedging that matches the evidence — a claim should be stated with the same confidence the evidence actually supports. This is not weakness; it is accuracy.
Hedging language calibrates how strongly you state a claim. The degree of hedging should match the quality of the evidence behind the claim.
| Hedging level | Language | When to use |
|---|---|---|
| Strong | "demonstrates," "shows," "proves" | Only when evidence is direct and replicated |
| Moderate | "suggests," "indicates," "appears to" | For single studies with clear findings |
| Weak | "may," "could," "it is possible that" | For speculative or indirect inference |
The most common error is understating claims from your own results ("the results seem to suggest…") while overstating claims from cited literature ("Smith (2020) proves that…"). Apply hedging to match the evidence, not to appear modest about your own work.
The convention in older scientific writing was to use passive voice throughout: "The model was trained on..." "Results were obtained by..." This avoided first-person constructions and kept the focus on the procedure rather than the researcher.
Modern guidance is more nuanced. Passive voice is appropriate when:
Active voice is appropriate when:
The real rule is: write the sentence that is clearest. Do not use passive voice as a default, and do not use active voice to avoid all formality.
A well-formed academic paragraph has three parts: a topic sentence that states one claim, evidence from cited sources or your own results that supports it, and analysis that explains how the evidence supports the claim and links to the next paragraph.
Weak paragraph: "Many classifiers have been proposed for malware detection. These include random forests, support vector machines, and deep learning. Each has advantages and disadvantages."
This paragraph has three topic sentences and no evidence. It reads like a list of things the writer knows exist.
Stronger paragraph: "Gradient-boosted classifiers consistently outperform simpler baselines on static malware features, but their advantage narrows significantly when tested on data from later time periods (Raff et al., 2018; Pendlebury et al., 2019). This temporal degradation suggests that the features learned by these classifiers capture artefacts of the training-period malware ecosystem rather than invariant properties of malicious behaviour — a distinction that becomes critical when evaluating classifiers intended for deployment rather than retrospective analysis."
This paragraph makes one claim, supports it with two citations, and provides analysis connecting the evidence to a broader point.
| Instead of | Use |
|---|---|
| a lot of / lots of | many / a large number of |
| really important | significant / critical |
| showed that it works | demonstrated effectiveness |
| I think / I believe | the results suggest / the evidence indicates |
| obviously / clearly | (delete — if it were obvious, you would not need to state it) |
| stuff / things | (name the specific things) |
| very accurate | highly accurate / accurate to within ± X |
The following paragraph is from a student draft. Identify every academic writing problem present, and rewrite it to correct them: "Ransomware is a really bad type of malware that lots of companies have to deal with. Some studies have shown that having good backups helps a lot. I think organisations should definitely make sure they test their backups regularly because it obviously makes a difference."
