Most advice on problem statements tells you what a problem statement should do. Almost none of it shows you one. This page does the opposite: six fully written problem statements, one per discipline, each followed by a short reading of the work each sentence performs. Then the four moves that produce them, a template you can fill in, and the five objections Canadian committees raise most often.
Read the specimens as shapes, not as claims. Every example below is a structural model. None contains a real study, a real statistic or a finding attributed to a real author, and nowhere is an invented number dressed up as evidence. Where your own evidence belongs, the text carries a marker: [cite] for a source you must supply and [figure] for a number you must measure or find. Fill those in from your own reading. A problem statement whose numbers you cannot source is not a problem statement; it is a guess with a footnote.
What a problem statement has to do
A problem statement is not a summary of your topic and it is not a research question. It is the paragraph that makes a reader agree, before you have shown any data, that something is wrong and that finding out about it would be worth someone’s time. In a Canadian thesis or major research paper it usually sits at the end of the introduction, running between 150 and 400 words, and it earns its place by doing four things in order.
- The context. One or two sentences establishing the setting and why a reader outside your subfield should care.
- The known. What the literature has settled — compressed, and cited.
- The gap or tension. The precise thing that is not known, or the two findings that do not fit together. This is the sentence the whole paragraph exists to deliver.
- The consequence. What follows from leaving the gap unfilled — for practice, policy, theory or a population.
Miss the third move and you have a literature summary. Miss the fourth and you have a curiosity. Committees notice both.
1. Nursing (Master of Nursing thesis)
Transitions from hospital to home are a recognised point of risk for older adults, and discharge planning has become a core nursing responsibility across Canadian acute-care settings. The literature consistently reports that structured discharge interventions reduce readmission for patients with single chronic conditions
[cite], and that patient-reported preparedness at discharge predicts subsequent help-seeking[cite]. Far less is known about patients who are discharged with three or more concurrent conditions and no primary-care attachment — a group that makes up roughly[figure]per cent of medical-unit discharges at the study site. Existing preparedness instruments were validated in populations with a family physician to return to[cite], so it is not established whether they measure the same construct for unattached patients. Without that evidence, discharge nurses are applying a screening tool to a population for which its interpretation is unknown, and the patients least connected to follow-up care are the ones least well served by it.
Why it works. The context is one sentence and it is about practice, not about the literature. The known is two citations wide and no wider. The gap is doubly specific — a named population and a measurement problem, not “more research is needed”. And the consequence lands on the people the reader already cares about. Note that the only number is marked [figure]: it is the writer’s job to produce it from the study site’s own data, and a committee will ask where it came from.
2. Education (Master of Education thesis)
Provincial ministries have invested substantially in classroom assessment reform over the past decade, and the rationale — that assessment for learning improves achievement more than assessment of learning — is well established in the research base
[cite]. Studies of implementation have concentrated on elementary classrooms and on teachers who volunteered for professional learning[cite]. What has not been examined is how secondary teachers who did not opt in interpret the same policy language, particularly in subject areas where a provincial examination still determines a proportion of the final grade. Preliminary board-level data suggest that reported use of formative practices varies by as much as[figure]across departments within a single school[cite]. If that variation reflects competing accountability pressures rather than differences in teacher skill, then professional development aimed at skill will not close it, and the assessment reform will continue to reach students unevenly.
Why it works. The tension is structural, not personal: the writer has spotted that two policy instruments pull in opposite directions. That is a genuine research problem rather than a complaint. The consequence sentence also quietly rules out the obvious intervention, which is what makes the study worth doing.
3. Psychology (thesis or MRP)
Self-report measures of emotion regulation are among the most widely used instruments in clinical and developmental psychology, and their factor structures have been replicated across several adult samples
[cite]. Two findings sit uneasily together. Scores on these measures predict clinical outcomes reliably in treatment-seeking samples[cite], yet the same scores show weak or inconsistent associations with behavioural and physiological indices of regulation in laboratory tasks[cite]. The usual explanation — that self-report and laboratory measures index different constructs — has been asserted more often than tested, and the studies that could adjudicate between explanations have mostly used undergraduate samples in a single language of administration. It therefore remains unclear whether the divergence is a property of the construct or an artefact of who is being asked, which matters because the same instruments are used to make treatment decisions for populations that were never part of their validation.
Why it works. This is the “two findings that do not fit” version of the gap, which is usually stronger than the “nobody has studied X” version, because nobody-has-studied-X invites the reply that nobody needed to. Notice the explicit naming of a competing explanation and the admission that it has not been tested. That sentence is what turns a puzzle into a study.
4. Civil engineering (MASc thesis)
Municipalities across Canada are replacing buried water infrastructure faster than it was installed, and replacement decisions are typically made from break-history models that assume failure rates depend on pipe material, diameter and age
[cite]. These models perform acceptably on aggregated networks[cite]but are known to degrade at the scale at which decisions are actually taken — the individual segment. Freeze-thaw cycling and soil-moisture variation are plausible contributors to that degradation and are increasingly well instrumented, yet published break models rarely incorporate either, in part because break records and environmental series are held by different bodies and are not aligned in time. The result is that segment-level renewal priorities are set with models whose largest source of unexplained variance has never been quantified, and a municipality with a fixed annual budget has no way to know how much of it is being spent in the wrong street.
Why it works. The gap is given a cause — the data live in two organisations — which tells the reader why the obvious study has not been done and implies the contribution. The last clause converts a modelling weakness into money, which is how an engineering committee reads significance.
5. Public policy or public administration (MPP / MPA major research paper)
Program evaluation in the Canadian public sector is governed by a formal requirement that spending be assessed for relevance and performance, and evaluation reports are published on a predictable cycle
[cite]. Research on evaluation use has largely asked whether findings are adopted[cite]. It has asked much less often what happens to evaluations whose findings are inconvenient to a program’s stated rationale, and almost nothing is known about the interval between a critical finding and any visible change in program design. Because evaluation reports and subsequent program documents are both public, that interval is in principle measurable, yet no study has attempted it for a defined set of programs. Until it is measured, arguments about whether evaluation functions as accountability or as ritual rest on assertion, and departments have no benchmark against which to judge their own responsiveness.
Why it works. The gap includes its own feasibility argument: the documents are public, so the study can be done. Committees kill proposals on feasibility more often than on interest, and a problem statement that pre-empts the objection saves a meeting.
6. Computer science (MSc thesis)
Retrieval-augmented systems are now a standard architecture for question answering over private document collections, and benchmark results on open-domain datasets improve steadily
[cite]. Those benchmarks share a property that deployed systems do not: the answer is present in the corpus. In an organisational deployment, a substantial share of queries —[figure]per cent in the only published audit we could locate[cite]— concern information the corpus does not contain, and the correct behaviour is abstention rather than retrieval. Evaluation practice offers no widely adopted way of scoring abstention, so systems are tuned on a metric that rewards answering and deployed in a setting that punishes it. The consequence is that reported gains on benchmarks provide little information about how a system will behave on the queries most likely to cause harm.
Why it works. The move here is a mismatch between how the field measures and what the field deploys, which is a reliably fundable problem shape in computing. It also names the deliverable implicitly: a way to score abstention.
The fill-in template
Write these five sentences, in this order, then cut.
- In [setting or field], [practice, policy or phenomenon] is [established fact about its importance].
- Research has established that [what is known]
[cite]and that [second thing known][cite]. - However, [the specific thing that is not known, or the two findings that conflict], particularly for [named population, setting or condition].
- This gap persists because [reason: data access, measurement, cost, disciplinary separation].
- Until it is addressed, [who does what worse, or which decision is made blind].
Sentence four is optional in some fields and decisive in others. If your gap is one a reader might think should already have been filled, include it.
Five objections Canadian committees raise, and the fix
| What you will hear | What it means | The fix |
|---|---|---|
| “This is a topic, not a problem.” | You have moves one and two and no move three. | Write the sentence that begins “However” and make it about one specific unknown. |
| “Who says?” | An unmarked assertion is doing load-bearing work. | Every claim in moves one and two carries a citation, or it comes out. |
| “Is this feasible in two years?” | The gap is real but the study implied by it is not. | Narrow the population or the setting in move three, not the ambition in move four. |
| “Your gap is that nobody has done it here.” | Geography alone is rarely a research problem. | Say what is different about here that makes the existing finding uncertain, and link that to your theoretical framework. |
| “Where is your number from?” | A statistic appeared without a source. | Source it, or replace it with the qualitative claim you can actually support. Canadian microdata routes are set out in the Statistics Canada PUMF, DLI and RDC guide. |
How the problem statement constrains everything downstream
The paragraph you write here is not decoration; it is a set of promises. Move three fixes your research questions, because a question that does not address the stated gap will read as a non-sequitur. It fixes your population and therefore your sample size justification. It constrains your design, which then has to be defended in the methodology chapter. And if your gap concerns a specific population, it shapes what your research ethics board will want to see about recruitment and consent.
Write the problem statement before the proposal, then read it again after your literature review. If the review has not changed it, the review was not thorough enough. The rest of the introduction is set out in the literature review guide.
Draft it, then make it yours
The hardest part of a problem statement is the first version — the one that lets you see what you are actually arguing. Tesify drafts from your sources and your notes so you have something concrete to cut down, and every word stays yours to change. Start your thesis with Tesify — 100% written by you.
Frequently asked questions
How long should a thesis problem statement be?
Usually 150 to 400 words, or one to two paragraphs. If it runs past a page you have written a literature review; if it runs under a hundred words you have probably skipped the consequence.
Where does the problem statement go?
At the end of the introduction in most Canadian theses, immediately before the research questions or objectives. Some programmes ask for a separate “Statement of the Problem” heading — check your department’s template.
What is the difference between a problem statement and a research question?
The problem statement establishes that something is unknown and that it matters. The research question is what you will actually ask about it. One problem statement usually supports two to four questions.
Can my problem be that nobody has studied this in Canada?
Only if you can say why the Canadian case would differ — a different policy regime, a different population, a different funding structure. “It has not been done here” on its own is the objection committees raise most.
Do I need statistics in a problem statement?
No, and an unsourced one is worse than none. If a number strengthens your gap, source it properly; otherwise make the claim qualitatively.
How many citations should it carry?
Enough that every claim about what is known is attributable, which in practice is usually three to six. The gap sentence itself is normally uncited — that is the point of it.
Can I write it before the literature review?
Write a draft, yes. Expect to rewrite it once the review is done. A problem statement that survives a thorough review unchanged usually means the review stayed inside what you already believed.
Is the problem statement the same in a major research paper?
The moves are the same; the scale is smaller. An MRP problem statement typically narrows to a single population or dataset rather than a field-level tension.
What if my supervisor and my committee want different problems?
Write both versions in full and take them to a meeting. Disagreement at this stage is cheap; disagreement after data collection is not.
Should I use the word “gap”?
You can, but you do not have to, and overusing it makes the paragraph read as formulaic. What matters is that the unknown is specific enough that a reader could tell whether you had filled it.