03 · What You Need to Know
How to Estimate What Your Research Will Actually Cost
Start with the methodology, not a budget ceiling
A common approach is to begin with the amount of money available and distribute it across broad categories. That can be useful later, but it is not the best way to determine what the study actually costs.
Begin with the research plan. What must happen for the question to be answered? What people, materials, software, equipment, travel, services, data, and facilities does each stage require?
Research-budget guidance from major funders follows this activity-based logic. The U.S. National Institutes of Health, for example, expects proposed costs to be reasonable, allowable, allocable, and directly related to the proposed work. UK Research and Innovation similarly structures project costs around the resources required to conduct the proposed research.
Your thesis budget may be far simpler than a grant budget, but the principle remains useful: the cost estimate should emerge from what the methodology requires.
Separate costs you already have from costs the study creates
Not every resource used in research creates a new expense.
You may already have access to a computer, office space, internet connection, statistical software, laboratory facilities, recording equipment, institutional cloud storage, library databases, videoconferencing tools, or research-support services through your university.
Check before purchasing anything.
A commercial software package that appears to cost hundreds of dollars may already be institutionally licensed. A laboratory instrument may be available through another department. Your university may provide transcription software, survey platforms, secure storage, statistical consultation, or interlibrary access.
Conversely, do not assume institutional availability simply because a resource exists somewhere on campus. Verify whether you can actually use it, whether access is free, whether booking is required, and whether it will remain available during your study.
Distinguish fixed costs from costs that grow with the study
This distinction is especially useful when evaluating whether changing the sample size or number of sites will change the budget substantially.
Fixed or largely fixed costs
Costs that remain similar across a reasonable range of study sizes, such as a software license, equipment purchase, ethics submission fee where applicable, or one-time setup service.
Variable costs
Costs that increase with participants, interviews, visits, samples, sites, travel days, transcription hours, tests, or other units of research activity.
If every additional participant requires a ₱500 payment, a laboratory test, and two follow-up visits, increasing the sample by 100 has substantial financial consequences. If the primary cost is one annual software license, increasing the sample may have relatively little effect on that particular expense.
Use quantities and unit costs
A useful budget should make its arithmetic visible.
This is more useful than writing "Participant expenses: approximately ₱50,000" because it shows what assumption produced the number. If the required sample changes to 180, you can immediately see how the budget changes.
Research software can be free, institutionally licensed, subscription-based, or surprisingly expensive
Software costs vary enormously.
Open-source tools such as R and many Python packages can eliminate licensing fees for some analytical workflows. Commercial statistical, qualitative-analysis, survey, reference-management, transcription, visualization, GIS, simulation, or specialized laboratory software may use annual subscriptions, monthly subscriptions, institutional licenses, student pricing, perpetual licenses, or usage-based fees.
Do not estimate software cost from memory or from what someone paid several years ago. Pricing and licensing models change. Check the provider's current official pricing and then check whether your institution already provides access.
Also identify how long you need the software. A three-month project may not require the same license arrangement as a three-year longitudinal study.
Finally, consider whether collaborators need their own access. One license on your laptop may not be sufficient if several authorized researchers must work with the files.
Free software is not automatically the cheapest option
A zero-dollar license can still have implementation costs.
If using an unfamiliar open-source package requires substantial training, programming, troubleshooting, or specialist support, those demands belong in the feasibility assessment. A commercial package may sometimes be more economical when institutional licensing, existing expertise, support, and project time are considered.
The reverse can also be true. Paying for expensive software does not make an analysis methodologically superior.
Choose software based on what the study requires, what the research team can use competently, reproducibility, access conditions, and total resource implications rather than license price alone.
Equipment costs include more than the purchase price
If the study requires equipment, ask whether you need to buy, rent, borrow, or merely book it.
Then look beyond the advertised price. Equipment may require accessories, calibration, installation, maintenance, consumables, protective equipment, software, training, technical support, shipping, insurance, or replacement parts.
A sensor costing ₱20,000 may require disposable components for every participant. A laboratory instrument may be free to use internally but require a technician. Recording equipment may require microphones, memory cards, batteries, storage, and backups.
For expensive resources, verify availability before designing the study around them. This connects directly to whether your institution actually has the facilities and resources your study requires.
Consumables can become larger than equipment costs
Reusable equipment attracts attention because of its visible purchase price. Consumables can quietly become the larger expense.
Laboratory research may require reagents, assay kits, specimen containers, gloves, pipette tips, slides, culture materials, cartridges, test strips, or other single-use items. Field research may require batteries, labels, printing, protective materials, or replacement components.
Calculate consumables per participant, specimen, test, batch, or session and multiply them by the expected volume. Include reasonable allowance for repeats, quality control, failed runs, or unavoidable wastage where appropriate.
Travel should be calculated trip by trip
"Travel: ₱20,000" is difficult to evaluate without knowing what it covers.
List where the researcher must go, how often, who must travel, and how long each visit lasts. Depending on the project, costs may include local transportation, airfare, rail or bus fares, fuel, tolls, parking, accommodation, meals or per diem under applicable policies, visas, travel insurance, and transport of equipment or materials.
For fieldwork involving repeated visits, calculate the number of trips rather than only the number of sites.
Remote data collection may reduce travel, but only if it remains methodologically appropriate for the research question and population.
Transcription should be estimated from audio duration, not interview count alone
Twenty interviews does not tell you how much transcription is required unless you know how long those interviews are.
Professional transcription services may charge by recorded minute, recorded hour, turnaround time, speaker count, language, audio quality, verbatim requirements, or other characteristics. Automated transcription tools may use subscription or usage-based pricing and still require substantial human checking.
If you transcribe the material yourself, the direct monetary cost may fall, but the time cost can be considerable. Researchers should not treat their own labor as infinitely available simply because it does not generate an invoice.
Translation and multilingual research can add another layer of cost
If participants use multiple languages, costs may include translation of recruitment materials, consent forms, questionnaires, interview guides, transcripts, quotations, or reports.
High-stakes or validated instruments may require more rigorous adaptation procedures than ordinary document translation. Depending on the methodology, this can involve forward translation, independent review, reconciliation, back translation, pilot testing, or specialist linguistic input.
Do not assume that a bilingual member of the research team can absorb unlimited translation work without consequences for time, quality, or workload.
Participant payment is not one single category
Researchers often use terms such as payment, incentive, compensation, and reimbursement interchangeably, but they can refer to different arrangements.
Reimbursement
Repayment of reasonable expenses participants incur because of research participation, such as transportation or parking.
Compensation or payment
Payment recognizing participants' time, inconvenience, burden, or contribution according to the study's approved arrangement.
Terminology and requirements vary by institution and jurisdiction. The U.S. Food and Drug Administration, for example, distinguishes reimbursement for expenses from payment for participation and states that payment to research participants is generally acceptable but should be reviewed to avoid undue influence in FDA-regulated research.
Your ethics committee or institutional review board may have its own requirements concerning the amount, timing, method, and description of participant payments. Check those requirements before promising a particular amount.
Participant payment should be multiplied by the number who will actually receive it
If payment is tied to participation, clarify who qualifies and when.
Will screened but ineligible participants receive anything? Is payment provided per visit, after each completed activity, or only at the end? Are travel expenses reimbursed separately? In a longitudinal study, are payments prorated across visits?
These details affect both ethics and budget.
For repeated visits, calculate payment per participant across the entire protocol rather than using the amount associated with one encounter.
Participant payment may affect recruitment, but it should not be treated merely as a recruitment purchase
Payment can recognize time, inconvenience, expenses, or burden and may influence willingness to participate. Ethical review is important because payment arrangements should not improperly pressure people to accept risks or remain in research against their interests.
FDA guidance notes that payment should generally accrue as participation progresses rather than being contingent entirely on completing a study, although a modest completion incentive may sometimes be acceptable under appropriate review.
The applicable rules vary across contexts. The practical lesson for budgeting is that participant-payment structures should be developed alongside the ethics and recruitment plan rather than added casually after the sample-size calculation.
Do not forget screening costs
A study may need to screen considerably more people than it ultimately enrolls.
If screening requires staff time, laboratory tests, clinical assessments, travel reimbursement, telephone calls, or other resources, those costs should be based on the number screened rather than the final sample.
Suppose you need 100 eligible participants but expect only half of screened candidates to qualify. A screening test costing ₱300 would need to be budgeted for roughly 200 candidates under that assumption, not 100.
This is another reason to connect the budget with the recruitment funnel required by the study.
Data can have acquisition and access costs
Existing data are not always free.
Some datasets require purchase, subscription, membership, application fees, administrative charges, secure-environment fees, data-extraction costs, or researcher travel to a controlled facility. Providers may charge for customized data preparation or linkage.
Even when the data themselves are free, using them may require computing infrastructure, secure storage, specialized software, or staff time.
Before building a project around secondary data, determine not only whether permission is required but whether the access arrangement creates costs that your project must absorb.
Specialist services belong in the budget
A study may require statistical consultation, qualitative methodological support, programming, database development, laboratory analysis, instrument calibration, graphic design, translation, transcription, data linkage, information-security review, or another specialist service.
Some institutions provide these services internally at no direct cost to students or investigators. Others charge hourly rates or project fees. External consultants may charge substantially more.
If the study depends on expertise you do not possess, verify availability and cost before assuming that help will be accessible later. It is also important to determine when a statistician, methodologist, or other specialist should become involved, since expertise that could affect the study design may be needed before data collection begins. Any associated consultation or collaboration costs should therefore be considered early.
Survey platforms and communication can create recurring costs
An online survey may appear virtually free, particularly if your institution provides a platform. But advanced features, higher response limits, institutional branding, multilingual functionality, complex branching, secure data handling, SMS distribution, or specialized integrations may require paid plans or services.
Participant communication can also generate costs through SMS, telephone calls, postage, courier services, or other channels.
Check which features the study actually requires rather than assuming the free version of a service will support the final protocol.
Data storage and security may cost money
Research data can require secure storage, backups, encryption, controlled access, or long-term retention.
Your institution may provide suitable systems without a direct charge. Large imaging files, genomic data, video recordings, high-frequency sensor data, or other substantial datasets may exceed ordinary storage allocations.
Restricted data may also require an approved computing environment rather than consumer cloud storage.
Data-management requirements should therefore be established before budgeting storage. The cheapest storage option is irrelevant if the applicable data-protection or provider requirements prohibit its use.
Research assistants and personnel can dominate larger budgets
For student research, the researcher often performs most tasks personally. Larger projects may require research assistants, interviewers, field workers, laboratory technicians, programmers, coordinators, or other personnel.
Personnel costs can include wages or salary, benefits, taxes, training, supervision, and other employment-related costs depending on the institution and funding arrangement.
Even when no salary is paid, ask whether the workload is realistic. Replacing paid assistance with your own labor changes the monetary budget but does not make the work disappear.
Your own time is a real feasibility resource
Researchers frequently solve budget problems by deciding to do everything themselves.
You can transcribe the interviews, write the code, clean the data, conduct every site visit, prepare every document, and analyze everything personally. This may indeed reduce direct expenditure.
It transfers the constraint from money to time.
If self-transcription saves ₱30,000 but requires weeks of work that your thesis timeline cannot accommodate, the cheaper financial option may be the more expensive feasibility option.
Budget decisions should therefore be evaluated alongside whether the complete study can fit within your deadline.
Price the study at the scale you actually intend to conduct
A pilot budget and a full-study budget can look dramatically different.
Per-participant costs multiply. Additional sites create travel and coordination expenses. Longitudinal studies repeat payments, assessments, communication, and travel. Multilingual studies may require several versions of materials. More interviews create more transcription.
Calculate the complete intended study rather than estimating the cost of the first few participants and assuming the remainder will somehow be inexpensive.
Use current quotations for consequential costs
For small routine expenses, reasonable estimates may be enough. For items that materially affect feasibility, obtain current information.
Ask the transcription service for a quotation. Check the official software price. Ask the laboratory what it charges per test. Confirm the facility-use fee. Determine current travel costs. Ask the data provider whether access involves a fee.
Prices change, currencies fluctuate, and institutional discounts expire. A research budget copied from a proposal written three years ago is a historical document, not necessarily a current estimate.
Include taxes, fees, and currency effects where relevant
An advertised international software price may not equal the amount charged to your card or institution.
Taxes, foreign-exchange conversion, bank charges, shipping, customs duties, transaction fees, and institutional procurement requirements can affect the final amount. The importance of these costs depends on the item and setting.
For expensive purchases in another currency, use a reasonable exchange-rate assumption and allow some room for movement rather than budgeting to the last centavo.
Some costs occur only if something goes wrong
Equipment can fail. Samples can require repeat analysis. Participants can miss appointments. Materials can be damaged. Travel may need rescheduling. A transcription recording may need manual correction because the audio quality is poor.
You do not need to budget for every imaginable catastrophe. You should consider predictable uncertainty, especially for expensive or failure-prone activities.
Funders commonly permit contingency only under particular rules, so grant-funded projects should follow the sponsor's budgeting requirements. For personal feasibility planning, the broader principle is simply to avoid assuming perfect efficiency.
Do not confuse contingency with an unexplained miscellaneous category
A budget containing a large line labeled "miscellaneous" is difficult to defend because nobody knows what it represents.
Where possible, identify foreseeable costs explicitly. Then maintain a modest allowance for genuine uncertainty if appropriate to the project and permitted by the funding or institutional rules.
Transparency is useful because it tells you which parts of the design create financial risk.
Check when each expense must be paid
Total affordability and cash flow are different questions.
You may have enough funding overall but not at the time an expensive item must be purchased. Participant payments may be needed throughout recruitment. A software license may require payment before analysis. Travel must often be paid before reimbursement. A data provider may invoice before access is granted.
If you are relying on reimbursement from a university or grant, determine what expenses can be reimbursed, what documentation is required, and how long reimbursement usually takes.
A study can be affordable on paper while still being impossible to finance at the moment the expense occurs.
Identify who is actually paying
Do not write "covered by the university" without confirming what that means.
Is the resource freely available? Does your department have a budget? Has funding been formally approved? Is your supervisor paying from a grant? Are you expected to pay first and claim reimbursement? Is access limited to staff rather than students?
Likewise, do not quietly assume that you will personally cover any shortfall. Determine the maximum amount you are genuinely willing and able to contribute before the project begins.
Watch Out
Do not make the study financially feasible by assigning every unbudgeted task to yourself. Doing your own transcription, programming, travel, data management, analysis, and administration may reduce invoices while creating a workload that makes the research impossible to finish on time.
Separate essential costs from desirable upgrades
When the first estimate exceeds your budget, classify expenses by whether the study actually requires them.
| Cost type |
Question to ask |
Example |
| Essential |
Can the research question be answered credibly and safely without this? |
A required laboratory assay or necessary participant follow-up |
| Method-dependent |
Would a different but defensible method remove this cost? |
Travel required for face-to-face interviews when remote interviews would also be methodologically appropriate |
| Efficiency-enhancing |
Does this mainly save researcher time rather than determine validity? |
Professional transcription instead of researcher transcription |
| Optional enhancement |
Would removing this leave the central research contribution intact? |
An additional secondary analysis requiring expensive proprietary data |
This classification helps reduce costs intelligently rather than cutting whichever line happens to look largest.
Do not cut costs in ways that make the study incapable of answering the question
A cheaper study is not automatically a feasible study.
Reducing the sample below what the design requires, eliminating necessary measurements, using an invalid instrument because it is free, removing essential follow-up, or replacing appropriate analysis with an inadequate method can lower the budget while undermining the research.
If the scientifically necessary version of the study costs more than you can afford, the next question is whether a good research question can still be a bad thesis because it is too expensive.
Sometimes the responsible response is to redesign the question rather than underfund the methodology.
