The Reader Problem: A Working Guide to Academic Writing in Ten Tips
June 3, 2026
Most PhD students learn to write academically by absorbing what they read. They internalise the field’s conventions through years of immersion in monographs, journal articles, and the running commentary of supervisors. By the time they sit down to draft a thesis, they know what academic writing sounds like. What they often don’t know is how it works.
Academic writing has a specific job, and that job is harder than the writing itself implies. A doctoral thesis is read, in the first instance, by an examiner who is doing this evaluation alongside their own research, their teaching, their administrative load, and another half-dozen review tasks they have promised to complete that month. They are not reading for pleasure. They are reading to assess whether the work in front of them is a genuine contribution to knowledge, and they are doing this under a time constraint they will never acknowledge. The writing has to make their job easier. Every sentence is in competition with their inbox.
This is the unstated condition of academic writing, and most of the craft that distinguishes good academic writing from competent academic writing comes back to it. The introductions that work are the ones that earn the reader’s attention before the reader has to ask for it. The paragraphs that read smoothly are the ones the writer has structured so the reader does not have to. The literature reviews that feel substantive are the ones written for someone who already knows the literature and is asking what the writer makes of it. The conclusions that land are the ones that tell the reader something they did not have before reaching the page.
There is no single move that makes a piece of academic writing read well. There are about ten, and they tend to be taught implicitly rather than explicitly, picked up over years of revisions and reviewer comments. This guide is an attempt to make them explicit. The ten tips that follow are not a style guide; they are a practical map of where academic writing tends to fail and what to do about it. They organise into three pillars: structural choices that make an argument findable, voice choices that make the writer visible on the page, and process choices that make the work sustainable across the months it takes to finish.
Ten tips, three pillars
Every craft move in academic writing belongs to one of three pillars. Architecture makes an argument findable; authorial presence makes the writer visible on the page; process makes the work possible at all.
Architecture
Structure that makes an argument findable.
- 01The introduction funnel
- 02Topic sentences for every paragraph
- 03Signposting at three levels
- 04Cohesion between paragraphs
- 05Conclusions that add, not repeat
Authorial presence
Voice and stance on the page.
- 06Synthesise, do not summarise
- 07Hedge to the strength of the evidence
- 08Description, analysis, and “so what”
- 09Voice: precision, stance, register
Process
How the work actually gets made.
- 10Separate drafting from editing
Architecture: the structure that makes an argument findable
Five of the ten tips concern structure. This is not coincidence. Structural choices are the ones a reader feels first — and is the least likely to consciously notice. They determine whether the reader can locate the argument, whether the paragraphs feel like they are going somewhere, whether the chapter has a discernible shape. When the architecture is right, none of this is visible; the reader simply finds the argument easily. When it is wrong, the reader works harder than they should, gradually loses patience, and eventually stops reading carefully.
The five architectural moves below operate at different scales. The introduction funnel works at the chapter or paper level. Topic sentences work at the paragraph level. Signposting and cohesion operate across paragraphs and sections. Conclusions work at the level of the whole work. Together they form the skeleton of a piece of academic writing. Done well, they make the work readable in the basic sense: a reader can follow the argument without effort, and the writer’s thinking is given the chance to be evaluated on its merits rather than on whether the reader can find it.
Architecture across the scales of a chapter
Each structural move operates at a different level. The reader receives the cumulative effect, not the individual moves, which is why no single one of these tips is sufficient on its own.
The introduction funnel
The single most rewritten paragraph in any thesis is the first one. Students draft it, abandon it, rewrite it from scratch, abandon it again, and finally commit to a version they are not sure about because they have run out of time. The trouble is almost never that they cannot write well; it is that they are trying to do too many things in too few sentences, in too small a portion of the page.
A working introduction does three things, in order, and almost nothing else. It establishes a context broad enough that an intelligent stranger can locate the topic in their own mental map of the world. It identifies a gap — something missing, contested, or insufficiently understood within that context. And it announces a contribution: the work the present study does to address that gap. That is the funnel structure, and it has been a convention in academic writing for so long that breaking from it tends to confuse readers more than impress them.
The reason it works is that it matches how readers actually approach a new piece of writing. A reader who does not know the topic needs context before they can evaluate a gap; a reader who does not know the gap cannot evaluate a contribution. Trying to compress all three into one paragraph, or to lead with the contribution, requires the reader to do exhausting backward work. Most readers stop trying.
What this means in practice is that the most common bad introductions are not stylistically bad; they are structurally premature. They open with a dictionary definition, or with “Since the dawn of time...”, or with a generic statement about the importance of the broad field, because the writer is trying to delay the moment when they have to commit to a specific framing of the work. The fix is to stop delaying. State, in the very first sentence, something specific enough to be wrong. A good introduction makes the reader feel that the study was inevitable, and overdue.
Topic sentences for every paragraph
Once the introduction has done its job, the chapter that follows depends on a sequence of paragraphs each of which must do its own. The mistake almost every doctoral writer makes — and continues to make for years — is to treat the paragraph as a container for evidence rather than as a vehicle for an argument. A paragraph without a topic sentence is a paragraph in which the reader has to guess what they are reading and why.
A topic sentence is the first sentence of a paragraph, and its job is to state, in advance of any evidence, the single claim the paragraph is going to make. Everything else in the paragraph exists to support, qualify, or develop that claim. The reader who has read the topic sentence knows what the paragraph is about; the reader who has only read the second or third sentence is, structurally, lost.
The reverse-outline test is the easiest way to diagnose a draft. Read only the first sentence of each paragraph in sequence, ignoring everything else. If those first sentences form a coherent argument on their own, the paragraphs are doing their work and the chapter is structured. If they do not — if reading them in order produces a sequence of disconnected observations or a list of facts — then the topic sentences are not yet doing their job, and rewriting them is the most efficient way to fix the chapter.
This is a structural fix masquerading as a writing fix. The reason many drafts feel scattered or rough is not that the prose is awkward but that the topic sentences are missing or weak. A paragraph that ends with its point rather than opening with it forces the reader to read the whole paragraph before they understand why they read it. Cumulatively, that fatigue is what makes a draft hard to read.
The rule worth carrying everywhere: if you cannot write a topic sentence for a paragraph, you do not yet know what that paragraph is about. The paragraph is not the problem. The thinking is.
Signposting at three levels
Topic sentences tell a reader what each paragraph is about. Signposting tells a reader where they are in the whole argument. Both are forms of structural generosity, and both are routinely under-used because writers assume the structure they hold in their own head is also visible on the page. It is not.
Signposting operates at three scales. At the macro level it appears at the openings of chapters and sections: “This chapter argues that...”; “The following section examines...”. At the meso level it sits between sections, doing the work of transition: “Having established X, the chapter now turns to Y...”. At the micro level it threads through sentences: “However,” “By contrast,” “This suggests,” “A related issue is...”.
These phrases can feel redundant to the writer because the writer already knows what is coming next. They are not redundant to the reader. They are the verbal equivalent of road signs on a long drive: most are unnecessary if you already know the route, but the moment you are unsure, their absence becomes the only thing you notice. Academic readers are, structurally, unsure most of the time, because the writer has lived inside the work for years and the reader is encountering it cold.
The under-use of signposting is also a problem of confidence. Writers worry that announcing an argument in advance gives the punchline away, or that telling the reader what they are about to read sounds patronising. Neither concern survives contact with how readers actually behave. A reader who knows what is coming pays closer attention, not less; a reader who can predict the shape of the next two pages is reading actively rather than scrambling to catch up.
The instruction worth internalising is the oldest one in rhetoric: tell them what you are going to argue, argue it, then remind them what you argued. This is not repetition. It is the structural discipline by which a writer’s argument is allowed to be evaluated on its own merits, rather than on whether the reader managed to locate it.
Cohesion between paragraphs
Even with topic sentences and signposting, paragraphs can still feel disjointed. Each one is fine on its own, but moving from one to the next produces a slight jolt — a sense that the writing is jumping rather than flowing. This is a cohesion problem, and it is one of the most diagnostically useful signs that a draft has been written paragraph-by-paragraph rather than as a continuous argument.
Cohesion is built between paragraphs, not within them. Three techniques do most of the work. The first is echo: ending a paragraph with a key term and opening the next with a variation of it, so the reader feels a thread carrying forward. The second is the bridge phrase: an explicit verbal hinge between paragraphs (“This raises the question of...”, “A related issue is...”, “Building on this...”) that names the connection rather than leaving the reader to infer it. The third is backward reference: a brief gesture at what has just been established (“These inconsistencies, noted above, point to...”) before introducing the next move.
The before-and-after on this is striking. A draft can contain exactly the same information in two arrangements, one of which feels choppy and the other cohesive, with no change other than the addition of three or four bridging clauses. The reader of the cohesive version will say it “flows better.” What they will mean is that the verbal connections between paragraphs are present rather than implied.
Cohesion is a sentence-level discipline, not a global revision. It tends to be the last thing added to a draft, in the final pass, and it tends to produce the largest perceived improvement in quality per unit of editing effort. Writers who become consistently good at it tend to do so by reading their own drafts the way a stranger would: out loud, one paragraph at a time, listening for the moments where the connection between thoughts is in the writer’s head rather than on the page. The fix is almost always a short bridge, written between two sentences the writer had not realised were strangers.
Conclusions that add, not repeat
The conclusion is the last thing the examiner reads, which is to say it is the part of the work whose tone is held in memory longest. Most PhD conclusions are written, however, as if they were the first thing the examiner had ever seen of the thesis: they restate the research question, recap the findings, and stop. This is a summary, not a conclusion, and it leaves the reader with the impression that the work has ended without resolving.
A real conclusion does four things, in some order. It synthesises: it tells the reader what the findings, taken together, actually mean — not what each chapter found, but what the whole study now allows the field to say. It claims a contribution: it states, in the writer’s own words, what the work adds, whether theoretically, methodologically, or practically. It identifies limitations honestly, without apologising for them or hiding behind them. And it points forward: it identifies the questions that remain open, the directions the field might now move in, the work the present study has made possible.
A useful test for any conclusion draft is whether three specific sentences can be completed clearly. “Before this study, the field understood ___.” “This study has shown that ___.” “This matters because ___.” If all three can be filled in with specific content, the conclusion is doing real work. If any of them produces a hedge or an evasion, the conclusion is still a summary, and the writer’s contribution is still hidden inside the chapters rather than displayed at the end.
Two things conclusions should never do: introduce new arguments or evidence, and close with a vague “more research is needed”. The first is structurally wrong — arguments need development, and the conclusion is not the place for it. The second is a confession of having nothing to say. The reader who reaches the last page deserves something they did not have when they started reading it. The conclusion is where that something gets delivered.
Authorial presence: voice and stance on the page
Structure determines whether an argument can be found. Voice determines whether it sounds like the work of someone who knows what they are doing. The four tips that follow are about the writer’s presence in the text — the difference between a draft that reads like a thorough summary of what others have said and a draft that reads like a researcher who has done the thinking themselves and is now sharing it.
This is the area in which the standard advice to “write more academically” tends to be most damaging. The advice is usually correct in spirit but unhelpful in practice; it produces drafts that sound formal without being precise, or hedged without being honest, or impersonal without being authoritative. The four moves below decompose what is actually meant by academic voice into specific, addressable habits: synthesising sources rather than summarising them, calibrating hedges to the strength of evidence, distinguishing description from analysis, and combining precision with stance and register into something that reads recognisably as the writer’s own.
Synthesise, do not summarise
The literature review is the part of a thesis where the writer’s presence first becomes visible — or fails to. The default mode for early drafts is summary: “Smith (2020) found that... Jones (2021) argued that... Lee (2022) showed that...”. This is a thorough record of what has been read. It is not a literature review. A literature review is an argument about the literature, not a tour of it.
The structural shift from summary to synthesis is grammatical. In a summary, the author’s name leads the sentence and the citation does the work of authority. In a synthesis, the idea leads the sentence and the sources appear as evidence: “While earlier studies emphasised X (Smith, 2020; Lee, 2019), more recent work has shifted toward Y (Jones, 2021; Patel, 2023), leaving Z largely unexplored.” The same five sources can appear in both treatments. What changes is who is doing the thinking on the page.
The practical move is to organise the literature by theme or finding rather than by author or chronology. Within each theme, the writer’s job is to identify agreements, contradictions, and gaps, and to write topic sentences that make claims about the state of the literature rather than describing individual contributions to it. The sources then become evidence for those claims. This requires the writer to read across sources rather than down a single citation at a time, which is harder than it sounds and is the part of a literature review that most distinguishes a good one from a merely competent one.
The reader-side consequence is direct. A summary tells the examiner what the writer has read. A synthesis tells the examiner that the writer has read and thought. Hiring committees, examiners, and journal editors all read literature reviews as evidence of intellectual maturity. A summary suggests the writer has not yet developed a view of their own. A synthesis demonstrates that they have. In a literature review, the writer is the author; the sources are the evidence. The job is to write like it.
Hedge to the strength of the evidence
Academic claims are almost never absolute. Some findings are robust; some are suggestive; some are tentative; almost none are universal. The writer’s job is to signal, in the language of the claim itself, how confident the evidence allows them to be. This is hedging, and it is one of the most consistently misunderstood features of academic writing.
Two failure modes are common, and they pull in opposite directions. The first is the overclaim: stating a finding from a single study as if it applied universally. “This study proves that...” is an overclaim. “The results show that all hotels should...” is an overclaim. A reviewer reading either of these stops trusting the writer immediately, not because the underlying work is wrong but because the writer is being careless about what their own evidence actually allows. The second failure mode is the over-hedge: qualifying everything to the point where no claim is being made at all. “It could potentially be the case that, in some contexts, certain participants may have indicated...” reads as evasive rather than careful. The reader cannot tell what the writer believes.
Between these extremes is a calibrated vocabulary that academic writing has built up over centuries. Verbs (suggest, indicate, appear to, tend to), adverbs (largely, generally, often, typically, potentially), and phrases (“the findings suggest,” “this may indicate,” “in most cases”) each carry a specific level of confidence. The writer’s job is to pick the hedge that matches the strength of the evidence, rather than defaulting to the strongest claim possible or hedging out of nervousness.
A well-hedged claim does not sound uncertain to a reviewer. It sounds honest. The reviewer reads “these findings may suggest that, within this context...” as the work of someone who knows the limits of their own evidence; they read “this proves that all hotels should...” as the work of someone who does not. Calibrated hedging is part of how peer reviewers determine whether to trust a writer’s judgment on the parts of the work they cannot independently verify. The hedge should reflect, as precisely as the writer can make it, what the evidence actually allows them to say.
Description, analysis, and the “so what” test
Almost every PhD writer is told, at some point in their training, to “be more analytical.” The instruction is true, and almost completely unhelpful, because nobody explains what it actually means. The result is that early-career writers either ignore the comment or layer on words like significant, suggesting, implies in the hope that they sound more analytical than they did before.
The actual distinction is structural. Description reports what exists: it states facts, summarises findings, recounts events. Analysis interprets what those facts mean: it explains why they occurred, what they imply, how they connect to other work, what follows from them. The same data can appear in both modes. “Hotel ICON received a 4.8 service rating” is description. “Hotel ICON’s 4.8 service rating may reflect the effectiveness of its student-staff training model, suggesting a link between educational immersion and guest experience outcomes” is analysis. The first reports the data; the second uses it.
Early PhD drafts are heavy on description because description feels safe. The writer can be confident in stating what their data showed; making an interpretive claim opens them up to challenge. The result is a draft that reads like a thorough report rather than an argument, and that elicits the standard supervisorial response: be more analytical. The fix is not stylistic. It is the disciplined habit of asking, after every finding stated, the simple question: so what?
Three moves convert description into analysis. Interpret: what does this finding mean in context, beyond what it literally reports? Connect: how does it relate to the broader argument, the theoretical framework, or the research question? Evaluate: does it confirm, complicate, or extend what existing work in the area has claimed? A finding that is followed by all three is participating in an argument. A finding that is followed by none of them is sitting on the page waiting to be analysed by someone else, and that someone else is usually the reader, who will not do the work.
Voice: precision, stance, register
Academic voice is the catch-all term for everything that distinguishes a draft that sounds like the writer from one that sounds like the institution they belong to. Most supervisors instruct early-career writers to develop one; very few explain what one is. The deconstruction worth carrying is that academic voice has three components, each of which can be worked on separately.
The first is precision: using exact terms where vague ones would do. “A significant increase” becomes “a 34% increase.” “Various researchers” becomes specific citations. “Some have argued” becomes a named claim by a named source. Precision is not the same as detail; it is the discipline of choosing the most specific term the evidence supports. The second is stance: the writer’s position is visible in the prose. A draft with strong stance commits to a view and signals it (“This study takes the position that...”); a draft without stance hides behind hedges, passive constructions, and vague attribution. The third is register: the texture of formality. Academic register is formal but not stuffy, confident but not arrogant, technical where the subject requires it and plain where it does not.
Four habits tend to undermine all three at once. Clichés (“in today’s fast-paced world...”) erode precision. Vagueness (“some researchers believe...”) erodes stance. Performative complexity — sentences whose only function is to sound sophisticated — erodes register. And hedging every single claim erodes all three by suggesting the writer is not willing to say anything definite. The combination is, unfortunately, the default voice of a first-year PhD draft, because it is the voice that feels safest.
Building voice is a daily practice, not a single revision. The exercise that works for most writers is to take one paragraph from a current draft, replace every vague term with a precise one, ensure every claim has a clear subject (either the writer or a cited source), and then read the result aloud. If it sounds like the writer performing academic writing, simplify. Academic voice is not about sounding smart. It is about being clear, precise, and intellectually honest — consistently, across the whole draft.
Process: how the writing actually gets made
The first nine tips concern what makes academic writing read well. The last concerns what makes it possible to produce academic writing at all. It is the most often-skipped category in writing instruction, and it is the one that most determines whether a PhD writer finishes on time, finishes thinly, or finishes at all. The craft of the prose can be perfected only if there is prose to begin with, and producing prose at the volumes a doctoral thesis demands requires a process that is sustainable across the months and years the work occupies.
Separate drafting from editing
The single most common cause of doctoral writer’s block is not lack of ideas, lack of evidence, or lack of discipline. It is the attempt to draft and edit at the same time. The writer types a sentence, immediately deletes it, retypes a slightly different version, deletes that, and after forty-five minutes has produced no usable text and a substantial amount of frustration. This is not a failure of writing; it is a failure of process. Drafting and editing are different cognitive modes, and trying to do both simultaneously produces neither.
Drafting is generative. Its goal is to get ideas onto the page in whatever form they arrive, however rough, in whatever order the writer’s thinking produces them. The only metric that matters is volume. A good drafting session ends with more words on the page than there were at the start of it, regardless of how good those words are. Editing is evaluative. Its goal is to make existing ideas precise, structured, and readable. The metric is quality, not volume. A good editing session may leave the writer with fewer words than they started with, but better ones.
The two modes are not just different; they are mutually destructive. Drafting requires permission to write badly, because the alternative is not writing at all. Editing requires the willingness to judge harshly, because the alternative is keeping mediocre prose in the final version. A writer who edits while drafting deletes the rough material that would have become good material with a second pass; a writer who drafts while editing keeps unrefined prose because they cannot bear to delete what they just produced. The two modes need to be physically separated — ideally into different sessions on different days — for either to work properly.
A workable structure is three sessions, with deliberate gaps between them. The first session is for drafting: write forward, do not look back, use placeholders (“EXPAND THIS”, “CITE HERE”) to keep moving past gaps. The second session, on a different day, is for structural editing: read the draft from the reader’s perspective, work at the paragraph level first, fix the topic sentences and signposting before touching individual sentences. The third session, after another gap, is for line-editing: cut before you add, simplify, check for precision. The gap between sessions is not laziness; it is the time required for the writer to read their own draft as a stranger would.
A first draft is supposed to be bad. That is not a failure of the writer. It is the process working correctly. Drafts get good through editing, not through being drafted perfectly. The writers who finish are the ones who have made peace with the bad first version — and learned to treat its imperfection as the precondition for the work, not a verdict on it.
The reader, the writer, the work
The ten tips above are individually small. They are paragraph-level habits, sentence-level disciplines, session-level routines. None of them is a methodological breakthrough; none of them, on its own, will transform a draft. Their power is cumulative. A draft in which all ten are operating is unmistakably different from a draft in which none of them is — not because the prose is more elegant but because the reader can do less work to get to the argument, and the writer’s thinking is allowed to be evaluated on its actual merits.
Behind all of them is a single principle, easier to state than to internalise: academic writing is a transmission problem. The writer has knowledge; the reader has limited time and motivation; everything in between is the writing’s job. Structure makes the argument findable. Voice makes the writer present. Process makes the work possible. The combination is not a style; it is a craft, and like most crafts, it gets better through deliberate practice on specific moves rather than through general aspiration to “write more academically.”
The reader the writer should be writing for, throughout, is not an idealised version of the writer themselves. It is an examiner, a peer reviewer, or a hiring committee member who is competent and overworked, who is reading the draft alongside three others, who wants to be persuaded but will not work unreasonably hard to find the argument. The ten tips above are the small concessions a writer makes, paragraph by paragraph, to that reader’s situation. A writer who consistently makes those concessions produces the kind of academic prose that gets read carefully, defended successfully, and cited later. A writer who does not is producing, at best, a draft the reader has to finish on their behalf.
None of this is new. Most of it has been said before in various forms across the history of academic style guides. What is rare is having it laid out as a practical map, with the underlying principle made explicit. The ten tips are the map. The reader problem is the principle. The work of the next few years is the writing itself.
Was this helpful?