PhD Research Proposal Example for Australia

PhD Research Proposal Example for Australia

Below is a complete two-page Australian PhD research proposal (about 839 words before references), written on UWA's five headings and annotated section by section. It sits well under Sydney's 2,000-word ceiling, and with its title, headings and timeline table it fills about two pages at 11-point single spacing (our estimate), which is UTS's two-page floor.

Our view: an Australian proposal is read first by one supervisor deciding whether they can supervise this project for three and a half years. The sections that decide that are methods, data access, ethics and the timeline, not the literature review. A sample that spends half its words on literature is written for the wrong reader. So in this example the methodology is the longest section: 314 words against 222 for the background. If you are drafting yours now, choose GradPilot's PhD research proposal rubric for a free review.

Every university page quoted below was checked on 31 October 2026. For how long your own university wants the proposal, see PhD research proposal length in Australia. For what to send at first contact, see the Australian PhD expression of interest example.

Contents

• What this proposal is for
• Constructed example: the full proposal, annotated
• The 3.5-year timeline, and why it is 3.5
• The ethics and permissions paragraph
• What changes for Sydney, Adelaide, UQ and UNSW
• What changes by field
• Questions about Australian PhD proposals

What this proposal is for

UWA asks applicants for "a summary of your proposed research outlining" five things: "background to the topic", "aims of the research", "significance of the research", "the methodology you'll use" and "a brief reference list" (UWA). It publishes no length. Those five headings are the skeleton below.

The length fits three other published rules. Sydney asks for a "research proposal of up to 2000 words", finalised "In conjunction with your supervisor" (Sydney). UTS says there is "no set format or length" but that a proposal "should be at least two pages in length" (UTS). Adelaide University uses a shorter form, shown later on this page.

The reader is a prospective supervisor first, then the faculty. In Australia the supervisor usually agrees before the formal application: Sydney says "you must secure a supervisor" (Law School and Business School excepted), and UNSW says "You must provide proof of the supervisor's agreement when submitting your application" (UNSW 2025 HDR Guide).

Constructed example: the full proposal, annotated

Constructed example, written by us, not a real applicant statement. Context (invented): an applicant with a four-year Bachelor of Pharmacy from an Indian university, three years as a community pharmacist, and a coursework Master of Public Health with a dissertation. The supervisor, the datasets, the pharmacy chain and every result are invented. Square brackets mark facts the applicant must fill in from checked sources; the example states no real-world findings.

Title (19 words)

Do guideline revisions change what community pharmacies dispense for adult sore throat? A two-country study of pharmacy dispensing records

Background to the topic (222 words)

Guidelines for adult sore throat in [Australia] and [India] were revised in [year] and [year] to [state what each revision changed] [Ref 1; Ref 2]. Whether dispensing changes after a revision is usually measured from prescriptions written by doctors [check against Ref 3]. That misses the setting I worked in for three years: the community pharmacy counter, where in [Indian state] antibiotics [describe how they reach patients, with a source] [Ref 4]. Studies that use pharmacy dispensing records directly are [describe their scope, as Ref 5 does].

Dr [Surname]'s group has used [the linked dispensing dataset it holds for an Australian state] to measure prescribing after [an earlier guideline change] [Ref 6]. My MPH dissertation coded 2,400 dispensing records from four pharmacies in one district of [Indian state] and found [state the dissertation's main result in one clause]. The dissertation covered a single year, so it could not show whether dispensing changed after a revision. A pharmacy record also shows what was handed over the counter, which a prescription record does not.

The gap this project addresses is a comparison: the same guideline question, asked of pharmacy records in two systems, with one method. Ref 6 measured an Australian revision from dispensing data; no study I have found [check this before submitting] has applied the same design to [Indian state] pharmacy records.

What it does: the background ends on a bounded gap built from named prior work (Ref 3, Ref 5, Ref 6), not on "little is known". The supervisor's own study appears in the second paragraph, so fit is visible before the aims. The applicant's dissertation is one sentence with one number.

Aims and research questions (124 words)

Aim: to measure whether dispensing for adult sore throat changes after a guideline revision, in two dispensing systems, using one method.

RQ1. In [Indian state] pharmacy records, did the share of sore-throat dispensing events that included [the antibiotic the revision moved away from] change after the [year] revision?

RQ2. In [the Australian dataset], did the same share change after the [year] revision?

RQ3. Which recorded factors (pharmacy type, day of week, patient age band) are associated with dispensing that does not follow the revised guideline in each system?

RQ1 and RQ2 use the same outcome definition, so the two results can be compared directly.

Scope. Adults only, and one condition, so that a single coding rule can be tested and applied within the doctorate.

What it does: each question can be answered "yes, by this much" or "no". The last line tells the reader why RQ1 and RQ2 are two questions and not one.

Significance (102 words)

If pharmacy records show a change that prescription data miss, guideline evaluations in [India] are measuring the wrong setting. If they show no change, the revision has not reached the counter, and [the state's antibiotic stewardship program] would know where to act. For Dr [Surname]'s group, the project adds a second country to its dispensing work and tests whether its outcome definitions transfer to records that were not built for research. The expected outputs are three papers, one per research question, a documented dataset for the [Indian state] records where the data custodian permits it, and a coding manual others can reuse.

What it does: both possible results are useful, so the project cannot "fail" by finding nothing. The second sentence says what the supervisor's group gains, not only what the applicant gains.

Methodology (314 words)

Outcome. The monthly share of adult sore-throat dispensing events that include [the antibiotic the revision moved away from]. A sore-throat event is defined by [the indication field, or the coding rule in my dissertation manual]. The same rule is applied to both datasets, and in year 1 I will test it on a sample that two people code independently.

Design. An interrupted time series in each system, with monthly counts of adult sore-throat dispensing events for [24] months before and [24] months after each revision. Segmented regression will estimate the change in level and trend at the revision date [Ref 7]. RQ3 uses multilevel logistic regression: each event is coded as following or not following the revised guideline, with pharmacy as the second level.

Data, [Indian state]. De-identified dispensing records from [number] branches of the [pharmacy chain] where I worked. The chain's operations manager has agreed in principle to share them; a signed data-sharing agreement is not yet in place. Fallback: if it is not signed by month 9, I will extend my dissertation dataset (four independent pharmacies, already held with the owners' written consent) to eight pharmacies in the same district. That changes the comparison from a chain to independent pharmacies, and the thesis would say so.

Data, Australia. [The group's linked dataset], accessed under [the group's existing data-custodian approval, to confirm with Dr Surname]. I would need to be added as an approved user.

Ethics and permissions. The project needs approval from [the host university's] human research ethics committee, from an ethics committee in [Indian state] for the pharmacy records, and from each data custodian. None has been sought yet. Records will be de-identified before transfer, and no record will leave [India] with names or phone numbers.

Skills. I used segmented regression in my dissertation. I have not used multilevel models and would take [the university's advanced statistics unit] in year 1.

What it does: every research question has a method, and every method has a named data source with a route to it. The one unconfirmed dependency (the data-sharing agreement) has a date, a fallback and a stated cost. The skills paragraph names a gap and how it closes, which our guide to proposal methods without prior experience treats in more depth.

Timeline over 3.5 years (77 words)

Months	Work	Milestone
1–12	Ethics applications; data-sharing agreement; Australian data access; RQ2 analysis begins	[Your university's first-year milestone]; fallback decision by month 9
13–24	RQ1 and RQ2 analyses; paper 1 submitted	[Mid-candidature review, if your university has one]
25–36	RQ3 analysis; paper 2 submitted	Paper 3 drafted
37–42	Paper 3 submitted; thesis written and submitted	Submission at 3.5 years

The main risk is the pharmacy agreement. The month-9 fallback keeps RQ1 answerable if it is not signed.

What it does: the Australian data, which needs no new agreement, is analysed first, so year 1 produces results even if the pharmacy agreement is slow. Milestone names are left in brackets because they differ by university.

Brief reference list (not counted)

Ref 1 to Ref 7: [full citations, in the style your supervisor uses]

What the example adds up to

By our count, the five counted sections total 839 words: background 222, aims 124, significance 102, methodology 314, timeline 77 (our arithmetic; the title and references are not counted). The methodology is the longest section, and the background is about 70% of the methodology's length (222 ÷ 314, our arithmetic).

The 3.5-year timeline, and why it is 3.5

The number comes from the scholarship, not the degree. UNSW's guide gives a "Length of Scholarship 3.5 years for a PhD, 2 years for a research Masters" for its main awards. The University of Auckland, in New Zealand, says you submit your thesis "after three years and within four years of your start date" (Auckland PhD).

Our recommendation: plan the thesis to finish inside the scholarship, not inside the maximum candidature. A plan that needs the fourth year without funding is a plan the supervisor has to worry about.

Milestone names differ. Auckland enrols you "provisionally" and expects a "full thesis proposal" by the end of the first year. Other universities use their own names for a first-year review. Write the milestone name from your university's own page, or leave it generic as the example does. Do not borrow another university's term.

The ethics and permissions paragraph

The Auckland Bioengineering Institute's proposal form asks the three questions an ethics and permissions paragraph has to answer: "Which data are required? What major methods or procedures will be used to collect and analyse the data? Which ethics approval(s)/permissions will be required?" (ABI form).

Two constructed examples, written by us, not real applicant statements:

Weak: Ethical approval will be obtained before data collection begins.

Revised: The project needs approval from [the host university's] human research ethics committee, from an ethics committee in [Indian state] for the pharmacy records, and from each data custodian. None has been sought yet.

The weak version promises a process without naming it, so the reader cannot judge the delay. The revised version names three approvals and says plainly that none exists yet. Never write that a project "has ethics approval" before it does. Our data access and feasibility guide covers the difference between "agreed in principle" and permission.

What changes for Sydney, Adelaide, UQ and UNSW

• Sydney. You may expand to 2,000 words, written with your supervisor. Spend the extra words on the methods (the model specification, the fallback sites) and on the literature behind the gap, in that order.
• UQ. Do not lead with the full proposal. UQ's article on proposals says sending it as your first point of contact means you have "jumped the gun a little" (UQ, 23 June 2026). The UQ EOI guide covers UQ's stages.
• UNSW. The central guide wants a research description of "approximately one page" with a title, the research problem, how you will address it, and previous research. Keep the title, the three questions in one sentence, the design paragraph and the dissertation sentence. Arts, Design & Architecture is different: it asks for 1,000 to 3,000 words. See the UNSW EOI guide.
• Adelaide University. Its proposal form sets limits: a title of "no more than 20 words", a summary "up to 100 words" that "must be intelligible to a reader who is not a specialist in this field", project details of "200-500 words" and research methods "up to 200 words". Our title is 19 words, inside the limit. The summary is below.

Constructed example: the same project in Adelaide's 100-word summary

Constructed example, written by us, not a real applicant statement. (91 words)

When sore-throat guidelines change, do pharmacies change what they hand over the counter? This project measures that directly, using dispensing records from pharmacies in [Indian state] and an Australian dataset held by [Dr Surname]'s group. It compares the two systems with the same method, month by month, before and after each guideline revision. If dispensing changes, the revision has reached the counter. If it does not, health services will know where to act. The results will be three papers and a coding manual that other researchers can reuse for pharmacy records.

The summary drops every technical term ("interrupted time series", "multilevel") because Adelaide says a non-specialist must follow it. The methods move to section 4 of the form, at 200 words or fewer. The 19-word title stays as it is.

What changes by field

• Computer science and engineering. Methodology grows a compute line (what hardware, whose allocation), a dataset line (licence, size, whether you can share derived data) and an evaluation line: what result would show the method works. Ethics shrinks unless people or personal data are involved.
• Education. If the research takes place in schools, check whether the state or school system runs its own approval process in addition to the university's ethics committee, and name it in the ethics paragraph.
• Humanities. Data access becomes archive access. Name the archives, the languages you read and any travel the work needs. The literature section may take a third of the words, because the argument with prior work is the contribution.

Questions about Australian PhD proposals

Do I need a hypothesis?

Not on UWA's headings, which ask for aims. UNSW Arts, Design & Architecture lists "Research problems, questions or hypotheses", so either form is accepted there. Use a hypothesis when your method tests one, as an interrupted time series does.

Can I reuse my UK proposal?

Yes, after three changes: cut the literature to about a quarter, add a timeline that ends at 3.5 years, and add a sentence on how the project extends your supervisor's work. Our 1,500-word UK proposal example shows the UK version of the same moves.

Should I name the supervisor in the proposal?

Yes, once they have agreed to discuss it. The example names Dr [Surname] in the background and the significance, because fit is part of what the reader judges.

Does the reference list count?

UWA asks for "a brief reference list" and publishes no limit. The example keeps references off the word count. If your form is silent, ask your supervisor. For another two-page format with a placement purpose, see the MEXT two-page research plan.

Get a free review of your research proposal

Get a free Quick Review of your research proposal against GradPilot's PhD research proposal rubric for UK, Ireland and Australia. It reads for a research question stated early, a gap built from named prior work, methods and data that answer each question, a scope that fits the doctorate's length, and the access and ethics the design depends on. For the paragraph on why this supervisor, use the PhD adviser-fit rubric, or start from the PhD statement of purpose review page.

More Australian PhD guides:

• PhD research proposal length in Australia: what 10 universities publish.
• Supervisor approval in Australia: what proof of agreement each university accepts.
• Auckland Statement of Research Intent example: the New Zealand version, by faculty.
• PhD essays hub: every PhD statement and proposal guide in one place.

Sources

Retrieved 31 October 2026, except where noted.

• The University of Western Australia, Higher Degree by Research (no year stated): the five summary headings.
• The University of Sydney, How to apply for postgraduate research (no year stated): 2,000-word limit, written with the supervisor, supervisor-first rule.
• University of Technology Sydney, Write a research proposal (no year stated): two-page floor.
• UNSW Sydney, 2025 Higher Degree Research Guide (July 2025): one-page research description, 3.5-year scholarship, supervisor agreement.
• UNSW Arts, Design & Architecture, Expression of Interest in Higher Degree Research form (January 2026): 1,000 to 3,000 words; "Research problems, questions or hypotheses".
• The University of Queensland, How to write a research proposal (published 23 June 2026): proposal as first contact.
• Adelaide University, Graduate Research Proposal Form (no year stated): section limits, non-specialist summary.
• University of Auckland, Doctor of Philosophy (no year stated): provisional enrolment, full thesis proposal, submission window.
• University of Auckland, Auckland Bioengineering Institute proposal form (no year stated): data, methods and ethics questions.