How to Compare PhD Opportunities and Make a Defensible Decision

Finding one promising PhD opportunity is difficult. Comparing two or three of them can be harder because the information is rarely presented in the same way. One page may give a clear project question but say little about fees. Another may describe a broad doctoral programme with excellent training but no single advertised project. A third may mention funding without explaining which fee category, period, or applicant group it covers.

The answer is not to find a universal ranking or to turn every part of a doctorate into a single number. A defensible decision is a decision you can explain using current sources, your own priorities, and an honest record of what is still unknown. It does not claim that one university or supervisor is objectively best. It shows why a particular option is the better next step for you, given the evidence available today.

Start with opportunities that have already passed a basic source check. Save the current vacancy, programme, funder, or provider pages. Then compare the same questions across each option. Separate facts from preferences. Keep missing evidence visible instead of giving it a comfortable score. A low score can look precise while hiding that you simply did not find the page that controls the decision.

This guide gives you a repeatable comparison method for funded STEM PhD opportunities abroad. It covers research fit, funding coverage, eligibility, application effort, supervisor and group context, training, facilities, timing, location, and personal constraints. It includes a worksheet you can copy into your own notes. It does not rank universities, judge people, predict acceptance, or tell you which offer to accept. The final decision belongs to you and, where a formal rule is involved, to the organisation that owns that rule.

The short answer: compare like with like, then explain the unknowns

Use the following sequence when two or more leads look serious.

First, define the comparison set. Name the exact project, programme, award route, host, intake, and application page for each option. If you cannot identify the current source that controls the next step, the option is not ready for a full comparison. It may still be useful as a discovery lead, but putting it beside a verified opportunity gives it more confidence than the evidence supports.

Next, make a common evidence table. Ask the same question of every option: What is the research question? Which methods and data are named? Who receives the application? What does the funding page actually cover? Which eligibility conditions are stated? What training and facilities are confirmed? Which dates control the next action? The table should contain a source link beside each important answer.

Then mark each entry as a confirmed fact, an interpretation, a personal preference, or an unknown. “The vacancy names a three-year project” is a fact if the current vacancy says it. “The topic would strengthen my computational skills” is your interpretation. “I prefer a smaller city” is a preference. “The award may cover the international fee difference” is an unknown until the provider or university confirms it.

Only after this separation should you use a weighted decision method. Weights are not facts about the opportunities. They are a way to express what matters most to you. You might give research fit more weight than location, or funding certainty more weight than a particular technique. Another applicant could reasonably choose different weights. A useful worksheet makes those choices visible; it does not turn them into an objective league table.

Finally, choose one next action for each option: apply, clarify, prepare, monitor, or remove. The next action can be more useful than a final winner. A short email that resolves a fee question may change the comparison. A check of the programme route may reveal that two apparently different listings are the same project. A source-backed decision record protects your time even when you decide not to apply.

Build a comparison boundary before you score anything

Comparison fails when the options are not really comparable. Applicants often put a live project vacancy beside a general department page, or a funded studentship beside an unfunded doctoral programme that has a separate scholarship route. These are different paths, with different owners and different decisions. The first task is to state what each option is.

Name the object you are comparing

Give every option a short label and a full identity line. A useful identity line contains the project or programme title, university or research organisation, department or group, host country, intended start period, funding route, and the URL that receives the application. Include the date you checked the page. If a page is a PDF or a portal page, save the direct link and the document name.

For example:

The identity line prevents a common error: comparing a programme name with a university name as if each were an offer. A university can host many programmes. A doctoral training partnership can recruit through several providers. A vacancy can sit inside a wider programme whose admission rules appear on another page.

If you need help collecting these links before you compare them, use the source-backed PhD search system. That guide focuses on organising discovery and evidence. This one starts when you have a set of leads that can be compared.

Decide whether the options belong in the same decision

Ask what decision you are actually making. Are you choosing where to submit an application this month? Are you choosing between two conditional offers? Are you deciding which funding route deserves a conversation with a supervisor? The right comparison fields depend on the decision stage.

At the discovery stage, compare evidence quality and next-step clarity. At the application stage, compare fit, effort, eligibility, and realistic timing. After offers, compare the confirmed terms, remaining conditions, supervision context, training, and personal constraints. Do not use a post-offer table to give a discovery listing the appearance of a confirmed place.

Two options can be related without being interchangeable. A funded project may be a direct recruitment route. A broad programme may invite your own proposal and ask you to obtain funding later. An MSCA Doctoral Network opportunity may have a defined project and mobility rules set by the current call and vacancy. An award such as the NSF Graduate Research Fellowship is an individual funding route, not a promise of admission to any university. The route type changes the questions you need to ask.

Require a complete minimum source set

Before a full comparison, require at least the following evidence for each option:

• a current page that identifies the opportunity or programme and the application owner;
• a current programme or graduate admissions page that describes the academic route;
• a funding page, call, studentship term, contract, or award notice that explains the support route;
• a research group, department, supervisor, project, or host page that makes the research setting traceable; and
• a current page for any government route you would need after admission, when immigration affects the decision.

The exact set can be larger. If the opportunity is in a doctoral network, keep the network information and the individual vacancy. If the funding is administered by a training provider, keep the provider terms and the university fee page. If a project includes work with regulated data or facilities, save the relevant institutional page or policy.

The point is not to collect a large folder of links. It is to make sure every major comparison column has an owner. If an option has only a search-result snippet, a social announcement, or an old PDF, mark it as a lead that needs verification. Do not compensate for the missing source by giving it a neutral score.

Set a source date without pretending that it freezes the page

Record when you checked each source and what page state you saw. A checked date helps you notice that a deadline, fee statement, or funding term may no longer be current. It does not make the source permanent. High-risk information needs to be checked again before submission and whenever the source says it has been updated.

Use the page's own update information when available. If no update date is shown, save the access date and the exact wording that matters. Do not create a new “freshness” date just to make an option look current. A source with no visible date can still be useful, but its uncertainty should be part of your record.

For discovery, DAAD's guide to finding a PhD position and EURAXESS Jobs can help you locate routes and vocabulary. They do not control the terms of an individual vacancy. Follow the link to the university, employer, provider, or funder, and compare from there. The discovery page belongs in your notes as a path to the source, not as a substitute for it.

Remove duplicate listings before comparing

The same project can appear on a university site, EURAXESS, a department page, and a newsletter. Check the project title, supervisor, host, funding wording, closing information, and application URL. If they describe one project, create one option with several discovery links. Do not count repeated listings as separate opportunities.

A duplicate can also appear under a project title and a doctoral programme name. Read the route instructions. One application may cover both, or the project may require a later programme application. Treat the relationship as a dependency and write it in the identity line.

Keep non-comparable options in a separate queue

Sometimes the evidence is useful but the route is not ready to compare. Put it in a “needs verification” queue. It can include an attractive project with no current vacancy, a scholarship announcement that names no application owner, or a programme whose funding is described only as “available”. The queue is not a rejection. It is a protection against spending hours scoring an option that has not yet shown how a person can apply.

Compare research fit as a set of concrete questions

Research fit is usually the most important part of a PhD decision, but it is also where applicants can become vague. “I like this topic” is a real preference, yet it does not tell you whether the proposed question, methods, data, training, and expected outputs match your goals. Compare the work itself, not only the title or a familiar keyword.

Start with the question the project will answer

Copy the current wording of the research question or aim. If the page does not state a question, summarise its stated problem without adding your own research claims. Note whether the opportunity is a fixed project, a broad field, or a proposal-led route. A fixed project gives you a defined starting point. A broad programme may offer more room to shape the question, but that flexibility can also mean more work before the scope is agreed.

Ask how the question fits your reason for doing a PhD. Do you want to develop a particular scientific method, work on a problem with social or industrial relevance, build a publication record, or move toward a research career? The page may not promise every outcome. Your comparison should show which parts are confirmed and which are your hope.

Use a table like this:

Do not make the question sound more specific than the source. A page that says “investigate mechanisms” does not confirm a particular experiment. An open project description does not guarantee that your preferred extension will be accepted.

Compare methods and the learning curve

List the methods the source names and classify your own relationship with each one: demonstrated, related, new, or not relevant. This is not an exam score. It is a way to see whether the learning curve fits the programme's training and the time available.

For Option A, you may already have experience with a method but need to learn a new instrument. Option B may use familiar data but require a statistical approach you have not used. Neither is automatically better. The useful question is whether the project provides the training, supervision, and access that the source indicates you will need.

Check how the page phrases the method requirement. “Essential” and “required” need a different response from “desirable” or “training will be provided”. If the wording is unclear, keep it as a question for the named contact. A supervisor's enthusiasm can help you judge fit, but it does not quietly remove a formal requirement.

Write one sentence for each option: “I can contribute immediately in _; I would need to learn ; the current source says training or support is __.” If the source is silent, say so. This prevents confident language from filling a gap in the page.

Check data, samples, and access

Many projects sound similar until you examine the data or samples. Compare whether the source names existing datasets, new collection, laboratory materials, field access, clinical partners, computational resources, or a host organisation. Record whether access is confirmed, planned, or dependent on permission.

Data access can shape the feasibility of a PhD more than a broad topic label. A project may rely on a partner dataset, an approval process, a field season, or a specialist facility. The source may explain the general setting without guaranteeing the exact material you would use. Do not write “data available” when the page only says that the project will work with a partner. Mark the dependency and ask who controls access.

If the project involves human participants, health data, controlled materials, or field permits, identify the institutional or government process named by the source. You do not need to solve the approval process in the comparison table, but you should not treat it as invisible. A longer access path can be acceptable if the project has a clear plan and your goals justify it. The comparison should show the path.

Compare expected work and outputs without promising publications

Read what the source says about thesis work, training, deliverables, dissemination, software, data, papers, placements, or presentations. Keep “publication expected” separate from “publication guaranteed”. A programme may support researchers to publish, but the timing and number of papers depend on the work, field, authorship, and supervisor decisions.

Compare the kind of work you want to do. One option may produce a technical system and a set of experiments. Another may produce a series of studies or a policy-facing dataset. The decision is personal. A technically impressive output is not automatically a better fit if it does not support your research question or future plans.

Avoid using journal names or impact metrics as a shortcut for quality. The San Francisco Declaration on Research Assessment explains why journal-based metrics should not stand in for an assessment of a researcher's work. If a programme uses a publication expectation, read its own wording. Do not turn a journal metric into a promise about your future career.

Compare question fit with career fit

Career fit can be useful, but it should be stated as a connection rather than a prediction. Ask what skills, methods, collaborations, or domain knowledge each option would help you build. Check whether the source names training, placements, internships, teaching, or professional development. If it does not, keep your assumption visible.

Use a simple two-column note:

This distinction protects the decision from wishful thinking. Your hope may still be a strong reason to choose an option. It should not be written as if the university or funder promised it.

Compare funding coverage and financial uncertainty

Funding is not one field called “funded”. It can include a stipend, salary, tuition support, research expenses, travel, insurance, relocation assistance, or a combination of sources. Compare what is stated and what is absent. A shorter award with clear coverage may be more useful to you than a longer award whose fee category is not explained.

Name the support type

For each option, write the support type exactly as the current source describes it. It might be a studentship, scholarship, research assistant employment contract, doctoral training award, fellowship, salary, or provider-managed grant. The label matters because it can affect tax, employment rights, fee treatment, eligibility, and the documents you receive. Do not translate “funding available” into “fully funded”.

UKRI's doctoral training framework and studentship guidance show why the provider matters. A national funder may invest through universities or training partnerships, while the provider recruits and administers the studentship. The national framework gives context. The university or provider page controls the individual route and terms. Make both visible in your source map.

Break funding into coverage rows

Use separate rows for each cost or support item:

The last wording is important. “Not stated” is different from “not covered”. The first tells you that a source gap remains. The second is a claim that needs a controlling source. Use the latter only when the terms clearly exclude the item.

Check the fee category, not just the word “tuition”

International applicants may face different fee categories. A scholarship can pay a specified rate rather than every possible rate. A provider may say that it covers “home fees” and offers no statement about an international difference. A university page may explain the fee category separately from the vacancy. Put those pages side by side.

Do not infer coverage from an advertised stipend. A living payment and a tuition obligation are separate rows. Do not infer that a fee waiver covers registration, bench fees, insurance, or other charges. If the source uses a term you do not understand, follow its definition or ask the fee office.

Compare duration and conditions

Record the stated start and end points, continuation conditions, probation or progress rules, and any link between funding and registration. A funding route can be available only for a named project, intake, or cohort. A studentship may be conditional on admission or on meeting a status requirement. A fellowship may have a host or mobility condition.

The current award terms decide the detail. Examples such as the NSF Graduate Research Fellowship, the Australian Research Training Program, CIHR training awards, and A*STAR scholarships illustrate that funding routes have their own programme rules. They are not interchangeable templates. Use the comparison table to ask the right question for your option, then read the current call or provider terms.

Compare certainty without scoring a missing page as zero

You can mark funding evidence as confirmed, partly confirmed, or open. This is not a prediction. It describes the state of your records. “Confirmed” means the current controlling source clearly states the support item for the route you are considering. “Partly confirmed” means some terms are clear and another important item is referred elsewhere. “Open” means the source does not answer a decision-critical question.

Do not let a simple score hide an unresolved fee or duration issue. Place the open issue beside the option and give it a next action. An option can remain attractive while you clarify it. It should not be described as fully funded until the evidence supports that phrase.

Compare eligibility and application feasibility

An option can be a strong research fit and still be a poor application choice if you cannot show the published requirements, meet the route conditions, or complete the required steps in time. Eligibility and feasibility belong in the comparison because they determine whether your effort can turn into a real application.

Separate formal eligibility from selection fit

Read the programme, vacancy, funder, and provider pages together. Mark formal requirements such as degree level, subject background, language evidence, nationality or residence conditions, employment status, mobility, or an application route. Keep selection preferences such as a particular method, publication, or industry experience in a different column.

The P08 guide on entry requirements and eligibility explains how to trace each requirement to its owner. Use it when one option has a complicated international qualification or language rule. This P12 comparison should show whether the rule is clear and what evidence you need. It should not decide your personal equivalency.

Count application steps, not only forms

An application may involve a vacancy form, a doctoral programme form, a funding application, supervisor confirmation, references, a language result, a proposal, or a separate mobility statement. List each step and its owner. A single portal can contain several decisions, while two separate forms may be parts of one route.

Compare effort in a way that respects your own time. An option with a shorter form is not necessarily better if its proposal requirements are unclear. An option with two applications may be reasonable if the funding and admission routes are clearly linked. Write the actual next steps, not an impression such as “easy” or “complicated”.

Check timing as a chain

Compare the date that controls the next action, not only a closing date. Ask whether the date belongs to a vacancy, a programme application, a funder call, a nomination, a reference upload, an offer condition, an award start, or an immigration stage. A project may close before the doctoral school portal opens. A scholarship may have a different deadline from the programme.

Make a timeline for each option. Include the date you need a language result, the date a referee must act, the expected decision period if stated, the intended start, and the point at which a government route could begin. If a date is absent, mark it as unknown. Do not fill the gap with a date from another cycle.

Check whether the route matches your current status

Some routes are open to applicants before graduation; others require a completed degree at a particular point. A fellowship may have a post-degree experience limit. A doctoral network may include mobility or residence conditions. An employment contract may ask for a right-to-work document at a stage that differs from a university application.

Read the current source and keep the exact status condition visible. If you need an institutional answer, ask the office that owns the rule. A supervisor cannot usually waive a funder's formal eligibility requirement. A discovery portal cannot confirm your government status.

Compare application feasibility with care

You can use a practical label such as ready, clarify, prepare, or not currently feasible. “Not currently feasible” might mean that a required document cannot be obtained before the route closes, that the source says the route is not open to your status, or that the application owner cannot be identified. It is not a judgment about your ability as a researcher.

If the issue can be resolved, keep the option in a clarification queue with one clear question. If it cannot be resolved before the relevant step, release the time and record why. A defensible decision includes the reason not to apply.

Compare supervisors and research groups without ranking people

The supervisor and group context can shape your daily research experience, but public pages rarely provide a complete picture. Compare evidence about the work and support structure. Do not create a personal ranking from publication counts, social impressions, or a single email.

Start with the supervision model

Look for the programme's stated supervision arrangements, progress reviews, doctoral committee, co-supervision, training requirements, or group structure. A project may name a principal investigator and a second supervisor. A doctoral school may set the formal supervisory process. Save both levels.

Ask who would guide the research question, methods, data access, and career development. The current source may name roles without explaining how often meetings occur. Do not invent a meeting frequency. If that detail matters, ask during a fit conversation or use the programme's formal guidance.

Compare what is confirmed:

The table avoids turning “seems supportive” into a fact. Your conversation can add useful evidence, but write what was actually said and keep its scope narrow.

Check current research activity, not only a historic profile

Read the supervisor's current university profile, group page, and recent project information. Publications can help you understand methods and topics. They do not guarantee that a person will remain available, have capacity, or supervise the exact project. Check whether the current opportunity names the same person and group.

ORCID can help you discover works and affiliations, but its records may include information added by the researcher or other sources. Use it to find publications, then verify the current role and project on the university or research-organisation page. Crossref metadata can help locate a DOI and publication record, but it does not describe supervision quality or future availability.

Avoid a publication-count comparison. A group with more papers is not automatically a better fit for your question. A smaller group may offer the methods and access you need. The comparison should be about the evidence relevant to your work.

Look for capacity and continuity questions

A public page may not say how many doctoral researchers a supervisor currently supports, whether a grant is ending, or whether a project will change. Those facts can matter. Ask the supervisor or programme a focused question when the source does not answer a decision-critical issue.

For example: “The vacancy names X as supervisor. Could you confirm whether the advertised project will be supervised by the named team for the intended start period, and whether co-supervision is part of the programme?” This asks for a factual clarification without requesting a promise about your success.

Do not treat an unanswered message as a negative score. It is evidence that the item is unresolved. If the route has a formal contact office, use it. If supervisor contact is optional, do not delay the application indefinitely while waiting for a reply.

Compare group fit with your working needs

Write down the conditions that help you work well. You may need a strong methods community, access to a laboratory, a quiet writing routine, regular contact, interdisciplinary collaboration, or a group that understands your data context. These are personal needs, not universal rankings.

Then look for public evidence: group methods, current projects, training pages, facilities, seminars, collaborations, or the programme's support structure. Mark assumptions as assumptions. A group page that lists a weekly seminar does not prove that the seminar fits your research needs, but it gives you a concrete question to explore.

The goal is a clear conversation and a realistic choice. You do not need a perfect public picture. You need to know which parts of the picture are visible and which need confirmation.

For a deeper check of research-group and supervisor fit, explore the current STEM fields and research pages and the named institution's own group and supervisor pages. Keep that assessment separate from the comparison total so that a detailed group review does not become an unsupported ranking of people.

Compare training, facilities, and research environment

The environment is more than a university name or a ranking. It includes the training route, facilities, data and software access, research partnerships, integrity support, and the practical conditions under which the project can be completed. Compare the features that affect your research rather than collecting prestige signals.

Read the training offer as a structure

Look for a doctoral training plan, required courses, methods workshops, transferable-skills provision, cohort activities, placements, or individual development reviews. UKRI's statement of expectations describes principles for doctoral training and the role of providers, but it does not guarantee the exact experience in a named programme. The provider page and programme handbook are still needed.

For each option, record what is required, what is available, and what is optional. “Training offered” is not enough. “The programme lists a methods module and an annual development review” is a usable fact. “I will become an expert in every method” is not.

Compare training with the gaps you identified in the research-fit section. If your project requires a technique that is new to you, find the named route for learning it. If a programme has broad training but no stated access to the specific method, keep that as an open question.

Check facilities and access conditions

Read the department, lab, institute, core-facility, library, computing, or research-infrastructure pages that are relevant to the project. Record whether the facility is named as part of the project, available through a partnership, or merely located at the university.

Access can depend on approval, training, booking, cost, location, or project budget. A facility page may describe capacity but not your personal access. Do not treat a nearby instrument as a confirmed resource. Ask how the project plan will use it and who pays any required charges when that affects feasibility.

For computational work, distinguish public software availability from the local computing support named by the project. For fieldwork, distinguish a research location from confirmed permission and travel support. For human data, distinguish a research partnership from an approved dataset. Keep the practical dependency visible.

Compare research culture without lifestyle marketing

You can learn about a research environment from seminars, working groups, open-science practices, data policies, mentoring, and collaboration pages. UNESCO's Recommendation on Open Science gives a broad framework for sharing knowledge and supporting open practices. It does not certify that every department follows the same practice, so look for the institution's own policy and evidence.

Ask whether the environment supports the way you plan to work. Open methods may matter for one project, while a specialist clinical partnership may matter more for another. The comparison should name the connection to your research.

Do not use a city guide, social media impression, or generic “student life” page as a proxy for a research environment. Location can be important, but it belongs in the personal-constraints section. A famous university can have a weak fit for a particular question; a less famous institution can have the exact facility and group you need.

Treat rankings and metrics as limited signals

Rankings compress complicated institutions into a few measures. They may be interesting for context, but they cannot decide research fit, supervision, funding coverage, or personal feasibility. DORA's declaration explains why journal-based metrics and similar shortcuts should not stand in for a meaningful assessment of research.

If you use a ranking as one contextual signal, label it as such and give it very little decision weight. Better still, replace it with evidence about the project, group, facilities, training, and terms you will actually experience. Do not include a “prestige score” in the matrix. It encourages false precision and can reward a signal that does not answer your question.

Compare timing, location, and personal constraints

The best research fit on paper may not be feasible under your current time, health, family, financial, or immigration constraints. These constraints do not make your decision less academic. They define whether you can start and complete the work. Include them explicitly, without turning them into a ranking of countries or cities.

Build a timeline for each option

Use the current source to identify the application deadline, reference deadline, funding decision, expected decision period, offer conditions, start date, enrolment point, and any government step that affects arrival. If the source does not publish a date, mark it as unknown. Do not copy a date from the previous cycle.

Represent the timeline as dependencies:

The government pages for Canadian study permits and the UK Student route illustrate why immigration evidence belongs in a separate source set. Your university may issue a document that connects to the route, but the government page controls the application. Do not give a visa certainty score based on a university webpage.

Include personal constraints without exposing private information

You can record a constraint in plain terms without putting sensitive details in a shared application record. For example: “Need a start period that allows current employment notice,” “must confirm dependent travel implications,” or “need accessible laboratory and housing arrangements.” These are planning facts for you, not claims about an institution.

Mark which constraints are fixed, flexible, or unknown. A fixed constraint may be a required date or a medical access need. A flexible preference may be a preferred city or travel time. An unknown may be the cost or availability of suitable housing. Do not convert a preference into a universal quality claim about a place.

Check location through the research route

Location matters when it affects field access, laboratory presence, travel to a partner, attendance requirements, or your personal ability to live there. Read the project and programme pages for the actual working location. A university may have more than one campus. A group may split work between a host institute and a partner site.

Record what is confirmed about presence, travel, remote work, and placement. Do not assume that a city label on a directory describes where the research takes place. If a project requires mobility, read the current programme or funder terms. MSCA routes, for example, have specific call and project conditions; the network page is not a substitute for the individual vacancy.

Make a private affordability check

Do not use an online cost-of-living estimate as a final funding answer. Make a private planning budget with the current award terms, fee information, likely housing, travel, insurance, and other personal costs. Keep the source for each known number and label estimates as estimates.

If the funding is not yet confirmed, do not spend money or make an irreversible move based on a hopeful interpretation. The comparison can still show that an option is worth clarifying. Financial caution is part of a defensible decision, not a sign that the research is less important.

Use a weighted decision method with limits

A decision matrix can help when several options are close. It is useful only when the categories are meaningful to you and the evidence behind each entry remains visible. The score is a summary of your priorities, not a prediction of admission or success.

Choose categories that answer your real decision

Start with four to seven categories. Too many categories create a false feeling of detail. A balanced set might include research fit, funding certainty, eligibility and application feasibility, supervision and group context, training and facilities, timing, and personal constraints.

Define each category in one sentence. “Research fit” might mean “how well the named question, methods, data, and expected work match my goals and preparation.” “Funding certainty” might mean “how clearly the current award and fee sources describe the support I would receive if selected.” Definitions prevent you from scoring the same fact twice.

Set weights from your own priorities

Give each category a weight that reflects your situation. The weights should add to 100 percent or to a simple total such as 10. Do not copy a universal set from a blog. If a funding gap would make a route impossible, you may set a minimum condition rather than trying to compensate for it with a high research score.

Example:

Another applicant may choose a different balance. The point is to show the reasoning.

Use an evidence rating, not a confidence prediction

For each category, you can use a small evidence scale such as:

• 2 means the current controlling sources answer the key question and the fit is strong for your stated need;
• 1 means the source answers part of the question or the fit needs clarification; and
• 0 means the option does not currently meet the stated need, or the evidence is not enough to continue.

Do not make “unknown” equal to 0 automatically. Keep a separate evidence state. A missing fee category is not the same as a confirmed fee exclusion. You can place a provisional 1 while the question is open, but only if the table clearly says that the score is provisional and a next action is required.

One possible format is:

If you calculate a weighted total, show the evidence notes beside it. A total without its inputs is not a defensible record.

Apply hard conditions before the total

Some conditions should not be traded against a high score. If the current route is not open to your applicant status, if a formal degree condition cannot be met, or if the funding terms make the route impossible for your fixed budget, the option may need to stop or clarify. A strong research score does not cancel a formal condition.

Write these as gates before the weighted table:

If a gate is open, the next action is clarification, not a confident score. This keeps the method from rewarding an attractive but unworkable option.

Test whether your decision changes too easily

After calculating a provisional result, change one weight or one unresolved assumption. If the winner changes immediately, your decision is close and you should not present it as obvious. Ask which evidence would resolve the difference. If the outcome remains similar across reasonable weights, you may have a more stable preference, but it is still your preference under the evidence available.

The purpose of this sensitivity check is not to produce a statistical claim. It reminds you that weights are choices. A matrix can organise a difficult decision, but it cannot remove uncertainty or tell you what you will experience in the future.

Do not score people, prestige, or acceptance chances

Do not add columns for “supervisor personality”, “university prestige”, or “chance of acceptance” unless you have a narrow, evidence-based definition that answers a real need. In practice, these columns invite guesswork. A supervisor's public profile cannot predict how your relationship will develop. A ranking cannot measure the fit of your project. An acceptance prediction has no reliable basis in a public comparison table.

Use observable questions instead: Is a supervision structure stated? Is the group working on the relevant method? Is the application route open? Does the current award document describe the fee category? This is more modest and more useful.

Run a Comparison Truth Check before you act

The Truth Check is a final audit of the matrix. Its purpose is to make each column traceable and to stop a preference, old summary, or hopeful assumption from becoming a fact. Complete it for each serious option, even if one currently looks much stronger.

Check the identity

Can you identify the exact project or programme, host, department, route, intended start, and application owner? Does each link lead to a current page rather than a search result or an old announcement? If two listings are duplicates, have you merged them into one option?

If the identity is incomplete, do not compare it as a confirmed opportunity. Move it to the verification queue and use the opportunity verification guide to complete the first check.

Check the source behind every material fact

For each row, ask:

If you cannot answer the source question, remove the fact from the score and write the open issue. The table becomes more honest when it has blank or unknown cells.

Check the funding language

Search your notes for words such as “fully funded”, “salary”, “tuition covered”, “international”, “guaranteed”, and “visa”. For each word, ask whether the controlling source supports the exact meaning you gave it. Replace a broad claim with a row-by-row description when needed.

The funding coverage guide can help you inspect the support and fee questions. A funding label should not be used as a shortcut for an unknown fee category, an award condition, or a government requirement.

Check the eligibility and timing language

Confirm that each formal requirement belongs to the same route and intake. A general university rule may not answer a project condition. A funder call may have a separate status or mobility rule. A closing date may control a nomination rather than the application itself. If you discuss immigration, keep the government page separate and current. Do not state that a visa will be issued or that a route is guaranteed.

Check the personal assumptions

Circle any sentence that begins with “I expect”, “I think”, or “this should”. These sentences may be useful preferences, but they should not appear in the fact columns. Change them to a preference note or a question. A personal preference is not a weakness. It becomes a problem only when it is disguised as institutional evidence.

Check the decision record

Can someone who did not build the table understand why you chose the next action? The record should show the options considered, the date, the main evidence, your weights or priorities, the open questions, and the action. It does not need every browsing step. It needs enough evidence to explain the decision and to update it if a source changes.

Keep a facts, preferences, and unknowns board

The matrix summarises a decision. The board protects the detail that a matrix can hide. Create three columns for each option and move every important statement into one of them.

Do not force an unknown into the facts column because the option feels exciting. Do not erase a preference because it cannot be verified by a university page. The three-column structure gives both types of information a proper place.

Use the board in a conversation

When you speak with a supervisor or programme office, take only the unknowns that the person can answer. A supervisor may explain the project's methods and expected scope. The graduate school may explain admission and supervision rules. The funding office may explain the award and fee category. The government authority controls immigration questions.

After the conversation, write the answer with its scope. “The supervisor said the team expects access to the facility” is different from “the facility access is guaranteed”. If the answer points to a policy, save the policy and let it control the row.

Keep the board after you decide

If you apply, the board becomes an application checklist. If you withdraw, it becomes a record of why. If you receive an offer, update the rows with the offer's exact conditions. If a source changes, you can see which preference or decision it affects. This is more useful than a screenshot of a final score.

Make the decision record and choose one next action

Write the conclusion in a few plain sentences. Name the option or options you will continue, the main reasons tied to your priorities, the facts that remain open, and the next action with an owner and date. If you are still choosing between two options, say that clearly and name the evidence that will decide the next step.

An example record might read:

I will prepare the application for Option A because its research question and method match my goals, and the current provider page gives clearer fee and support terms. Option B remains a strong research fit, but its international fee treatment and application sequence are not stated. I will ask the named fees office one question before deciding whether to submit there. I will recheck both current programme pages before submission.

This is a defensible decision. It does not claim that Option A is the best PhD in general. It explains the action you can justify now.

Keep the links, dates, and source wording in your private opportunity record. The methodology page explains how AcademicWings separates discovery pages from controlling sources. When your comparison is ready for the next step, browse opportunities and open the current university, employer, provider, or funder page before acting.

Official sources

• DAAD: finding a PhD position in Germany. A discovery workflow example that helps applicants find routes and research contacts. The individual university or funder page controls a specific opportunity.
• EURAXESS Jobs. A European researcher vacancy discovery service. Follow each listing to the employer, host, or funder page before comparing terms.
• MSCA Postdoctoral Fellowships. An official fellowship model that shows why current call documents and host information matter for route-specific conditions.
• UKRI doctoral training investment framework. Explains how doctoral training investment can be organised through providers and partnerships.
• UKRI studentships and doctoral training. Describes the studentship route and the role of providers in recruitment and support.
• UKRI statement of expectations for doctoral training. Provides context for doctoral training environments. It does not decide an individual programme or applicant.
• SNSF Postdoc.Mobility. An official mobility-fellowship example whose current call material controls conditions.
• Canada: apply for a study permit as a graduate student. Government guidance that illustrates source ownership for a graduate immigration route.
• Canada study permit. The official Canadian route page to read for current conditions, documents, and instructions.
• UK Student visa: your course. Official UK guidance that connects a course and its provider information to the government route.
• UK immigration rules: Appendix Student. The official rules source for the UK Student route. Read the current rules for your own case.
• DORA: read the declaration. Explains why journal-based metrics should not replace a meaningful assessment of research and researchers.
• ORCID: getting started. A researcher-record resource that can help discover works and affiliations, which should then be checked on the current institutional page.
• ORCID: source of record information. Explains how information sources appear in an ORCID record and why an institutional page may still be needed.
• Crossref metadata enrichment. Helps explain publication metadata discovery. It is not evidence of supervision quality, funding, or programme fit.
• UNESCO Recommendation on Open Science. A broad open-science framework that can help you read institutional research-culture policies without treating the framework as a programme promise.

These sources explain discovery routes, funding structures, research information, assessment limits, training environments, and government source ownership. They do not decide an individual application. For a real comparison, add the current direct pages for each exact opportunity, programme, provider, fee category, group, and government route.