Writing Your Thesis at a Company: Here's How to Make It Work

Your topic is approved, the company has said yes and you are looking forward to real data instead of another literature review. Then you spend three weeks waiting to send your first employee survey, because nobody has told you that your university's review board has to see it first.

You write your thesis at a company by checking your program requirements first, securing one advisor on campus and one at the firm, putting the use of internal data in writing, and submitting your study to the Institutional Review Board before you collect a single response. We will walk you through the timeline, the IRB question and the part most guides skip: how you actually get usable data inside a business.

📌 The key points

• An IRB determination comes before data collection, never after.
• Students cannot declare their own study exempt under 45 CFR 46.
• An anonymous survey of adults often qualifies for exemption category 2.
• The federal minimum wage has been 7.25 dollars since July 2009.
• Thesis abstracts stay public on ProQuest even during an embargo.

What a company-based thesis actually is

A company-based thesis is a final research project completed in partnership with a business: the topic comes from a real operational problem, the data comes from the firm, and the grade still comes from your university alone. The company has an interest in the outcome but no say in your evaluation.

American terminology is worth getting right, because it is the reverse of British usage. In the United States a master's capstone research project is a thesis, doctoral work is a dissertation, and an undergraduate version is usually a senior thesis or honors thesis. British universities call both undergraduate and master's work a dissertation and reserve thesis for the doctorate, which is why guidance written abroad can seem to contradict your handbook.

The company arrangement has its own vocabulary too. A structured, alternating work period is a co-op, short for cooperative education, and at schools built around it a co-op typically runs four to eight months and is usually paid. A shorter arrangement is an internship. A capstone project is a separate option at many programs, often positioned as a practical, applied alternative to the research thesis.

Work out whether your project needs primary data at all before you commit. The difference between an empirical and a theoretical thesis decides how much access to the business you really need, and how your own data fits into a methods chapter is covered in the overview of empirical research. If your level is already settled, two guides go deeper: one on the undergraduate thesis at a company and one on the master's thesis at a company.

The timeline of a company-based thesis in six stages

A company-based thesis starts roughly a semester before you register your topic and runs through six stages, the first four of which are purely administrative. Compressing them costs you writing time at the far end, because IRB review and contracts cannot run in parallel.

Allow more time than feels comfortable. Expedited IRB review commonly takes several weeks and full board review a month or more, and none of that time counts toward your defense date.

1. Read the program requirements. Does your department allow a project hosted by an outside organization, and who approves it? Both are set out in the graduate handbook.
2. Secure your faculty advisor. Your advisor has to agree to the project before you commit to the company, and at most schools the advisor is listed as the responsible investigator on the IRB application.
3. Turn the brief into a research question. The firm's operational problem has to become a research question that stands up academically. That translation is your job, not the company's.
4. Submit to the IRB. The determination comes before recruitment, and the instrument you submit is the one you have to use.
5. Get the agreement in writing. Pay, hours, data access, confidentiality and publication all belong in a written agreement before you start.
6. Write, submit, defend. Before the defense, agree who from the company will attend and which figures may be shown.

💡 Tip

Draft your recruitment email and consent language at the same time as your questionnaire, not afterward. Most IRB applications ask for all three together, and a submission missing the invitation text is the single most common reason a student application comes back with revisions.

Two advisors, two sets of expectations: how to defuse the conflict

A company-based project gives you two points of contact with different goals: your faculty advisor judges the quality of the research, your workplace contact wants something the business can use. The grade comes from the university alone, and the company has no influence over it.

Most of the time this works. It gets uncomfortable at exactly one point, and that point arrives earlier than most students expect: when your findings are not what the business hoped for. A suggestion to trim a chapter, smooth a figure or phrase something a little more positively sounds harmless, and it is the moment your research loses its value.

A company may remove its trade secrets. It may not remove your findings.

In practice you solve this through presentation rather than argument. Company names, sites, revenue figures and product names can all be anonymized without losing a single finding: the Ohio plant becomes Site A, absolute revenue becomes an index. That is what anonymization is for, and it often removes the need for an embargo altogether.

If the objection is about substance rather than confidentiality, tell your faculty advisor right away and put it in writing. A thesis that reports an inconvenient result, explains it and places it in context is evaluated more favorably than one that confirms what the business already believed.

IRB review: the gate that surprises most thesis students

An Institutional Review Board determination is required before you recruit anyone, and a thesis based on a company survey almost always needs one. The governing federal rule is the Common Rule at 45 CFR 46, and its definition of a human subject applies to an investigator “whether professional or student”.

Stanford puts the student case plainly in its research policy handbook: independent undergraduate research projects, honors theses, master's theses and dissertations that involve human participants and systematic data collection must be reviewed and approved by the IRB before initiation. The same policy makes the second point that catches people out. A project may well be exempt, but the compliance office has to review it and issue the exempt determination. You cannot decide that yourself.

Why exemption category 2 can fail on a workplace survey

Exemption category 2 at 45 CFR 46.104(d)(2) is the one that covers survey procedures, and it is met if the information is recorded so that identities cannot readily be ascertained, or if disclosure outside the research would not reasonably place subjects at risk. Read the second branch carefully before you rely on it, because the regulation names damage to a subject's employability explicitly. A survey of your host company's staff that is not genuinely anonymous can fail on exactly that word and fall back into limited IRB review.

⚠️ Warning

Do not run a pilot “just to test the questions” before your determination arrives. Pilot responses collected from real employees are data, and an IRB cannot approve a study retroactively. Test the questionnaire on classmates or friends outside the company instead.

Practice varies more than students expect. Harvard's Department of Government tells undergraduates that every project involving human subjects needs a decision form, while noting that most senior theses are not expected to require full review. Your own institution's policy is what binds you, so ask your IRB office early. This guide is not a substitute for that determination.

Getting a staff survey approved inside the company

Company sign-off is a second gate, separate from the IRB and just as capable of stopping your fieldwork. You usually need it from three places: your workplace contact or a senior manager, human resources, and whoever owns privacy and information security.

Missing approvals are the most common reason company-based projects run late. Most guides on the subject cover applications and contracts in detail and stop exactly where the research begins. Two things make the conversation go faster.

What HR actually wants to see before saying yes

Bring the complete instrument, not a description of it. HR reviews the wording of individual questions, because a question about a named manager or about protected characteristics creates exposure for the employer regardless of your academic intentions. Bring the recruitment email too, since many organizations want to approve the invitation text separately from the questionnaire.

Explain how anonymity is produced technically, in one or two sentences. Whether responses are tied to a login, whether IP addresses are stored, and who at the university can see the raw file are the three questions that come up every time. A tool that collects responses without requiring a company account removes most of the objection before it is raised, and it also strengthens your IRB position under exemption category 2. Where a union contract is in force, ask whether it says anything about surveying members, because that is a contractual question rather than a federal one.

Privacy and anonymity: what changes inside a business

The United States has no comprehensive federal privacy law covering employee survey data, and saying so plainly is more useful than implying otherwise. The Government Accountability Office stated in report GAO-19-52 that the country has no comprehensive internet privacy law, and its recommendation to Congress remained open as of February 2026. What governs your study instead is a mix of state law and your IRB's confidentiality conditions.

California is the clearest case. The California Privacy Protection Agency confirms that the exemptions for employment-related personal information expired on December 31, 2022, so California employees now have full rights under the CCPA as amended. A genuinely anonymous survey sidesteps most of this, because data that identifies nobody is not personal information to begin with. What to check technically is covered in the guide to data protection in surveys. This section is not legal advice.

Small teams: when demographic questions identify the person

A survey is only anonymous when nobody can be identified indirectly either. In a business with 3,000 employees, asking for department, age band and length of service is harmless. In a team of eight, those same three answers are a unique fingerprint, and everyone in the room knows it. Response rates then collapse not because the topic is sensitive but because the attribution is obvious.

Two rules deal with this reliably. Ask for demographic details only where your analysis genuinely needs them, and band them generously: three age groups rather than year of birth, three business areas rather than nine departments. Then say plainly in the invitation what is stored and what is not, instead of simply asserting the word anonymous. How a workplace survey is set up end to end is covered in the guide to employee surveys.

Response rates and sample size: how many answers to expect

Your sample in a company-based project is almost always capped by the size of the workforce, and that limitation belongs in your methods chapter rather than being glossed over. Inviting everyone at one site is not random sampling: it is a census with self-selection at the response stage.

No honest guide can promise you a response rate, because it depends on the size of the business, the topic and how visibly leadership backs it. So work backward instead of trusting a number from a blog: how many completed questionnaires does your analysis need as a minimum, and how many people must you invite to get there? What sample size your chosen method requires is a question for before recruitment, not after.

Five things lift response rates inside a business noticeably, and none of them costs anything:

• The invitation visibly comes from a senior manager, participation stays voluntary.
• Responding during work hours is allowed, confirmed in writing.
• The survey takes under ten minutes and says so upfront.
• One single reminder after a week, not three.
• Results go back to employees, with a stated date.

Reporting the results back is the strongest of the five and the one most often forgotten. Anyone who has filled out a workplace survey and never seen an outcome will not bother next time. It is also worth checking what makes a study representative before you write your conclusions: your findings may represent the business you studied, but they do not represent the industry.

Pay, hours and the written agreement: what to settle before you start

Pay for a company-based thesis usually comes through a co-op, an internship or a part-time position. The amount depends on hours, industry and location, and no guide can quote you a realistic figure for your case.

Whether the position has to be paid at all is a separate question with a real answer. The Department of Labor Fact Sheet 71 sets out a seven-factor primary beneficiary test for unpaid internships at for-profit employers, weighing among other things whether the internship is tied to your formal education program, whether it accommodates the academic calendar, and whether your work complements rather than displaces paid employees. The Department describes it as a flexible test with no single deciding factor. If the analysis makes you an employee, minimum wage and overtime are owed, and the federal minimum wage has been 7.25 dollars an hour since July 24, 2009, with the higher of the state or federal rate applying.

Five points belong in the agreement or a written side letter:

Data access is the most important line and the one most often left vague. A verbal promise that you will get the figures is worthless the moment your contact leaves the company or moves to another project. Write down which report, which period and which named person you are relying on.

Embargo and restricted access: what happens to the finished thesis

An embargo delays release of your thesis in the university repository and in ProQuest so that confidential material stays out of public view for a set period. Every school sets its own options, and the periods are shorter than most students assume.

The University of Washington is a useful illustration because it publishes the exact choices. Restricting a thesis to campus access only can be set for one, two or five years, while denying access to everyone is limited to one year, after which the work becomes openly available. In ProQuest the restricted access period can be six months, one year, two years or a specified length. Accepted grounds include protecting intellectual property during a patent application and preventing disclosure of sensitive information.

The abstract stays public even under embargo

One warning from the same policy deserves more attention than it usually gets. The metadata describing your thesis, including the citation and the abstract, is openly available immediately regardless of embargo status and is indexed by search engines. If your abstract names the company and quantifies its problem, the embargo protects nothing. Write the abstract as if it will be public on the day you submit, because it will be.

⚠️ Warning

Agree the embargo with your advisor before you start writing, not at deposit. Some departments will not supervise a restricted project at all, and finding that out four weeks before your defense leaves you choosing between a new advisor and a new topic. Anonymizing the company usually solves the same problem without any paperwork.

An embargo does not stop your work being evaluated. Your advisor and committee read the full text, and the graduate school keeps its copy. Your own institution's rules are what count here, so check the graduate school page before you plan around any of these numbers.

Conclusion

A company-based thesis rarely fails on the topic and almost always fails on the approvals around it. Settle your advisor, the IRB determination and the company sign-off before you commit to a defense date, and put every promise in writing. Securing access to the data contractually means the hardest part is behind you before you write the first sentence.

Where to go next

• Planning the empirical work behind it? Empirical bachelor's or master's thesis
• Need respondents beyond the company? Finding survey participants
• Want the whole research process in order? The empirical research process

Need a staff survey that clears IRB review without a second revision?

With empirio.ai, an online survey tool from Germany, you create your survey in a few steps, share it by link with no sign-up for respondents and watch responses come in live.

Create a survey for free