Chapter 3 is due on Friday, your title is about hotel guest satisfaction, and you have been staring at the respondents section for three days because nobody can tell you how many guests to survey. Your groupmates keep saying “just use Slovin’s.” Slovin’s formula needs a population size N. Nobody knows how many guests that hotel will have.
This is the specific wall HRM and HM groups hit, and it is not a sign you have chosen badly. It is a genuine methodological fork that your general research subject probably glossed over, and it has two clean answers depending on who your respondents actually are.
What it costs to stay stuck here
The respondents section is not one paragraph you can leave blank and come back to. It decides your sample size, which decides how long data gathering takes, which decides whether you finish before the semester does. It also decides your statistical treatment, because a sample of 40 and a sample of 385 do not support the same tests.
Groups that leave this unresolved usually discover the problem in October, three weeks into distribution, when it is too late to change the design without redoing the proposal. Meanwhile the establishment you were counting on has entered peak season and stopped answering your letters. A stalled respondents section in August turns into a repeated semester in March.
First: which population are you actually sampling?
Almost every HM and HRM title has one of two respondent groups, and they behave completely differently.
Group A — employees, staff or students. Hotel front office personnel, restaurant service crew, housekeeping staff, HM students in a specific year level, food service employees of a chain. These are a known, finite, listable population. Somebody has an HR roster with a number on it.
Group B — guests, diners, tourists or customers. Hotel guests, restaurant patrons, resort visitors, travellers at a terminal. These are an unknown, moving population. No fixed list exists, and next month’s guests are different people.
Read your title. If your respondents are Group A, Slovin’s formula applies and your problem is solved in five minutes. If they are Group B, Slovin’s does not apply, and using it anyway is the single most common methodological error in Philippine hospitality theses.
Group A: the Slovin’s computation, worked
Slovin’s formula is n = N ÷ (1 + Ne²), where N is the population size and e is the margin of error, conventionally 0.05.
Say your locale is a four-star hotel in your city with 280 rank-and-file employees, confirmed by the HR department in writing.
n = 280 ÷ (1 + 280 × 0.05²)
n = 280 ÷ (1 + 280 × 0.0025)
n = 280 ÷ (1 + 0.7)
n = 280 ÷ 1.7
n = 164.7, rounded up to 165 respondents
Three things your Chapter 3 must then say. Where N came from and who confirmed it — “as of the 2026 roster provided by the Human Resource Department.” That you rounded up rather than down. And your sampling technique, which for a Group A study is normally stratified random sampling if your comparison sub-problem groups employees by department or length of service, and simple random sampling if it does not.
Stratify when you are comparing departments. Simple random sampling across 280 employees can easily hand you six housekeeping respondents and forty from food and beverage, and a one-way ANOVA on groups that lopsided is a defense question you cannot answer.
Group B: what to use when the population is unknown
For guests, diners and tourists, use the standard sample size formula for an unknown or infinite population, sometimes called Cochran’s formula:
n = z²pq ÷ e²
Where z is 1.96 for a 95% confidence level, p is 0.5 (the most conservative estimate of the proportion), q is 1 − p = 0.5, and e is 0.05.
n = (1.96)² × 0.5 × 0.5 ÷ (0.05)²
n = 3.8416 × 0.25 ÷ 0.0025
n = 0.9604 ÷ 0.0025
n = 384.16, rounded up to 385 respondents
That number is why so many hospitality theses have a sample of 384 or 385 — it is not a coincidence or a convention, it is the arithmetic of a 95% confidence level at a 5% margin of error with maximum variability. Your sampling technique here is almost always purposive or convenience sampling with stated inclusion criteria, because you cannot randomly select from a list that does not exist.
State the criteria explicitly: Respondents were guests who had stayed at least one night at the establishment during the data gathering period and were at least 18 years of age. Then state the limitation honestly in Chapter 1, because a panel will raise generalisability whether you do or not.
If 385 is not feasible in your timeline — and for a single-establishment study during a lean season it often is not — you have two legitimate moves. Raise the margin of error to 7% or 10% and recompute, stating the choice and its consequence. Or reduce scope to a bounded population you can list, which turns a Group B study into a Group A study.
The constraint that actually decides your timeline
Neither formula matters if the establishment does not let you in. Hospitality theses live or die on the permission letter, and the sequence is usually: adviser endorsement, then a letter from your dean or department head, then a formal request to the General Manager or HR Manager of the establishment, then a scheduling conversation with the department supervisor who will actually host you.
Two things HM groups underestimate. Establishments refuse guest surveys far more often than employee surveys, because interrupting a paying guest is a service risk to them and nothing to you — expect to be redirected to a post-checkout online form or turned down outright. And peak season closes doors: December through the summer months, a busy hotel will simply not have the management time. Send your letters in the lean months and have a second and third establishment lined up before the first one replies.
Write the fallback into your plan now. A Chapter 3 that names one establishment and no alternative is a Chapter 3 with a single point of failure.
How the sample size flows into the rest of the chapter
Once n is fixed, four other sections fall into place. Your locale section describes the establishment and justifies why it fits your study. Your instrument section describes the questionnaire, its Likert scale and its validation — the item-by-item logic is the same across courses, and the step-by-step questionnaire build transfers directly to a guest satisfaction instrument. Your data gathering procedure describes distribution, which for guest respondents usually means deciding between paper cards at the front desk and an online form, a trade-off covered in the comparison of Google Forms versus paper for Philippine thesis data collection. And your statistical treatment section lists the tests each sub-problem needs, following the pattern in the explainer on statistical treatment of data in a thesis.
All of it has to stay consistent. If your sampling says 165 stratified by department and your statistical treatment forgets the ANOVA that stratification exists to support, the panel finds the gap in about ninety seconds. The full sequence of sections in the order Philippine schools require is laid out in the guide to writing Chapter 3 section by section, and the same design-first logic applies whatever your course, as the walkthrough on choosing a research design shows for a very different respondent group.
Getting the chapter written before Friday
You have now made every decision that requires your judgement: which group your respondents belong to, which formula applies, what your margin of error is, which establishment you are approaching. What is left is roughly 1,500 words of highly conventional academic prose — the design justification, the locale description, the sampling paragraph with the computation typed out, the instrument description, the procedure, the statistical treatment entries — in the exact register a Philippine panel expects.
That is the part Tesify does. You give it your statement of the problem, your respondent group and your computed sample size; it drafts the chapter around them, keeps the sampling, the instrument and the statistical treatment consistent with each other, and formats your citations in APA 7th as you write. The research stays yours — it has to, because only you know which hotel said yes. There is a free tier, so you can see what it produces for your own Chapter 3 before deciding whether to pay for anything.
Start your thesis on Tesify and have a full Chapter 3 draft to bring to your adviser this week instead of an empty respondents section.
Frequently asked questions
Can I use Slovin’s formula for hotel guests?
No. Slovin’s formula requires a known population size N, and hotel guests are a moving population with no fixed list. Use the unknown-population formula n = z²pq ÷ e² instead, which gives 385 respondents at a 95% confidence level and a 5% margin of error.
Why do so many hospitality theses use 384 or 385 respondents?
Because it is what the unknown-population formula produces at a 95% confidence level, a 5% margin of error and p = 0.5. It is arithmetic, not imitation. Show the computation in Chapter 3 rather than stating the number, so the panel can see you derived it.
What if I cannot reach 385 guest respondents?
Raise the margin of error to 7% or 10% and recompute, stating the choice and its effect on precision as a limitation. Alternatively, narrow to a listable population such as the employees of one establishment, which lets you use Slovin’s formula and a much smaller sample.
Is using an AI writing tool allowed by my school?
Policies differ by institution, so read your student handbook and ask your adviser directly. Most Philippine schools distinguish between assistance with drafting and language, which is generally accepted, and submitting work you did not produce, which is not. Using a tool to draft prose around your own data and decisions sits on the permitted side at most schools, but confirm rather than assume.
How much does Tesify cost?
There is a free tier that lets you draft and see the output before paying anything, and paid plans are billed in pesos for Philippine users. Check the current rates on the pricing page before signing up, and consider splitting a plan across your research group, which is how most Filipino thesis groups use it.
Is my thesis data safe if I use an online tool?
Read the privacy policy of any tool before uploading, and as a rule do not upload raw respondent data containing personal information. Under the Data Privacy Act your respondents’ details are your responsibility. Draft with anonymised or aggregated figures, which is all a writing tool needs anyway.
Do I need permission from the hotel before I collect data?
Yes, in writing, and the request normally goes through the General Manager or HR Manager after your adviser and dean have endorsed it. Guest surveys are refused more often than employee surveys, and peak season makes refusal more likely. Approach two or three establishments in parallel.
Should I round the sample size up or down?
Always up. A computed value of 164.7 becomes 165 and 384.16 becomes 385. Rounding down puts you below the margin of error you claimed. Type the unrounded result and the rounded figure in Chapter 3 so the panel can follow the arithmetic.