HS

Hasara Research Guide

Your final year research project, explained from the beginning. Enter the password to open it.

It only asks once on this device.

The viva
The day itself

Getting ready for the viva

What happens, what to bring, what to say in the first two minutes, every question they are likely to ask, and the honest answer to each one.

The one thing to understand about a viva

The examiners are not trying to catch you out. They are checking one thing: that you understand your own document and could have produced it.

Which means the worst answer is a confident wrong one, and the best answer is often "the data showed X, and here is what I concluded from it". You are never expected to defend a result. You are expected to explain it.

The two minutes that set the tone

You will almost certainly be asked to introduce your project. Have this ready and say it without notes.

Your opening, roughly 90 seconds

My research is about hotel-hosted social events in Sri Lanka. Hotels, guesthouses and beach villas host beach parties, live music nights and cultural shows, but they coordinate them manually, through chalkboards, flyers and WhatsApp groups.

That fails on both sides. Organisers cannot predict turnout, so they over-order or under-order food, drink and staff. Travellers either find out too late, or arrive to find tickets, packages or parking already gone. And 92.5 percent of the organisers I surveyed had run an event short-staffed or short-stocked, so this is recurrent rather than exceptional.

I tested three candidate fixes against Coordination Effectiveness: real-time event visibility, pre-booking and resource coordination, and centralized information access. I surveyed 64 travellers and 40 organisers with parallel instruments, so the same model could be tested twice independently.

All three correlate strongly with the outcome. But in simultaneous regression, only centralized information access contributes uniquely, at beta 0.665 for travellers and 0.461 for organisers, and it does so in both samples. So visibility and pre-booking work through unified information rather than beside it.

I then built EVENTIA, a web platform whose architecture is that finding expressed as a schema, and evaluated it against the requirements the survey produced.

A plan for the demo

  1. Have it open before you startTwo browser windows, already logged in. One organiser dashboard, one guest view of the same event. Do not spend viva time typing passwords.
  2. Show the event page firstPoint at the availability board. Everything a traveller needs to decide, in one view.
  3. Show the dashboard and name the repetitionSay: this is the same component reading the same rows. The hotel and the guest are looking at one record, not two copies.
  4. Book from the guest window and let them watchThe number moves on both screens with no refresh. This is the whole thesis in one moment.
  5. Show the QR, then scan it twiceFirst scan admits. Second is refused as a replay with the original time shown. That is TC-07.
  6. Mention the one thing you cannot show liveTwenty simultaneous bookings. Say the transcript is in Appendix D, exactly 10 committed and 10 refused, remaining landed on zero.
Check this the day before, not on the day

Open the deployed link and confirm it loads with the seed data present. Free hosting can sleep or expire.

Have a fallback either way. Every screen was captured at three widths and those screenshots are in the project evidence folder. A prepared screenshot is a professional fallback. A blank page is not.

The questions, and what to say

About the research itself

In one sentence, what is your contribution?

That for small hotel-hosted events, unifying the information is not one feature among several. It is the mechanism through which the others deliver value at all, demonstrated statistically in two independent samples and then expressed as an architecture and tested.

What is your research gap?

Three parts. First, no study has examined a unified platform combining real-time visibility, multi-resource pre-booking and a centralized organiser dashboard for small hotel-hosted events. Second, nobody has measured whether small Sri Lankan hotel operators would actually adopt such a platform. Third, the operational consequences, over-procurement, under-sold capacity and travellers missing events, have never been quantified in this context.

Each gap maps onto one of my objectives, which is how the objectives were chosen.

Why did you only study Mirissa and Matara?

I did not. The scope is national, and the target population is hotels, guesthouses and villas across Sri Lanka.

Collection was concentrated in the Southern Province coastal belt for three reasons stated in Section 3.5.1: it has the densest concentration of small and medium coastal accommodation, the event frequency there is high enough to produce a usable sampling frame of organisers who face the problem, and prior research documents an operational profile there that matches small operators nationally, which supports analytical generalisation.

Why Design Science Research?

Because the question has two halves that neither a purely behavioural nor a purely technical project would cover. A behavioural study would have explained the coordination problem without producing anything that addresses it. A software project would have produced an artefact with no evidence it addresses the right thing.

Design Science Research requires both, and Peffers' six-phase DSRM gave the execution order, mapped chapter by chapter in Table 3.1.

About the statistics

Your correlations support all three hypotheses. Why report the regression instead?

Because correlation asks whether two things move together, and regression asks what each one contributes once the others are accounted for. My three predictors overlap heavily, with IV2 and IV3 correlating at 0.738 among travellers, so at the bivariate level each one gets credit for variance it shares with the others.

The regression distributes that credit correctly. Reporting only the correlations would have supported all three hypotheses handsomely and would have been misleading, and Section 7.3 says explicitly that the temptation was resisted.

Your organiser alpha was 0.159. How is that acceptable?

It is not, and it was not accepted. That is the figure before instrument purification. One reverse-coded item, item 20, returned a corrected item-total correlation of minus 0.623 after reverse-scoring. Removing it by the standard 0.30 threshold lifted the scale to 0.690.

Table 4.3 reports both the initial and the final value for every construct, because showing only the final figure would conceal the diagnostic step that made the scale trustworthy.

Removing items until the alpha looks good sounds like fishing.

The rule was fixed in advance rather than chosen afterwards. Items below the conventional 0.30 corrected item-total threshold are removed, worst first, recomputing after each removal, and no scale may drop below 3 items.

Only 2 items in the whole instrument crossed that threshold, and they were exactly the 2 reverse-coded items, in both independent samples. No forward-worded item was touched and no further removals were made. If I had been fishing, the removals would have been scattered across the instrument rather than confined to the two items with a documented failure mode.

Is 64 responses enough for a regression with three predictors?

It is enough for what the study claims and not enough for what it explicitly does not claim. With 64 cases and 3 predictors there are roughly 21 cases per predictor, which exceeds the commonly applied minimum of 10 to 15 for stable estimates but falls short of the 50 to 100 per predictor recommended for optimal stability.

So the model is adequately powered to detect a large effect and underpowered for small ones. That is exactly why the non-significant coefficients are reported as a failure to detect a unique effect rather than as evidence that no effect exists.

If H1 and H2 were not supported, why build those features into the system?

Because the finding is not that they do not matter. Both correlate strongly with the outcome. The finding is that their contribution operates through unified information rather than independently of it.

So the architecture reflects exactly that. Real-time visibility is built as a read surface over the unified record, and pre-booking is built as the write path into it. They exist, but as expressions of the unified layer rather than as separate features beside it.

You defined the dependent variable as three sub-dimensions but scored one. Why?

Because the evidence decided. The principal component analysis showed the 4 surviving items forming a single dimension rather than three separable factors.

It is therefore scored as one unidimensional composite, and the three conceptual sub-dimensions are retained as facets rather than as separately scored subscales. Establishing them as distinct measurable subscales would require a longer instrument, which is a named future recommendation in Chapter 7.

About the system

Why did you put tickets, rooms and add-ons in one table?

Because they differ in what they represent to a guest but behave identically to the system. Each carries a price, a total quantity and a remaining quantity, and each must be decremented atomically.

One table gives a single query returning the complete resource position for an event, a single realtime subscription keeping every surface consistent, and a single locking discipline governing all concurrent decrements.

Modelling them separately would have fragmented availability across three code paths and reintroduced, inside the system, the same inconsistency between information sources that the research identifies as the primary failure of the arrangements it replaces.

Why PostgreSQL rather than a document store?

Because a booking must decrement several independent resource counts and either succeed completely or leave no trace, which requires transactional guarantees a document store does not provide by default.

Chapter 2 identifies concurrency control over multi-resource inventory as the central algorithmic problem for platforms of this type, and one reviewed source finds that platforms on eventually consistent stores expose stale availability to users. Stale availability is the exact failure this research exists to eliminate.

How do you know your concurrency test really ran concurrently?

Because the test asserts peak overlap as well as outcome, and it recorded 20 requests in flight at once.

That assertion is what makes the rest of the result meaningful. If the requests had been serialised by the runtime rather than held by the database lock, the outcome would have looked identical while proving nothing about the locking.

Why is access control in the database rather than the interface?

Two reasons. A defect in the interface then cannot expose one property's booking data to another, because the interface is not what withholds it. And it makes the guarantee testable independently of the interface, which is what TC-09 does by issuing requests directly against the API with one user's credentials against another user's data.

It also answers the data. Security and data protection appeared in 46.7 percent of substantive organiser adoption responses.

You report seven defects. Does that weaken your evaluation?

The opposite. Two of them would otherwise have produced a false pass. The most instructive is that the responsive test passed against 36 screenshots of an error page, because an error page does not overflow its viewport.

That was caught by opening the images, not by reading the test output, and the capture script now fails the run if any screen rendered an error. An evaluation that reports only its final state conceals the work that made the final state trustworthy.

Why does your system not handle payments?

It is declared out of scope in Section 1.10. The research question concerns information coordination rather than transactions, so a payment gateway would add regulatory and financial complexity without contributing to the thing being tested.

The system still records the server-computed total and marks the booking as pending payment, so the booking record is complete and a payment step could be added without changing the data model.

About the limits

Why did you not run a usability test?

The artefact reached a testable state too close to the submission deadline to recruit and run a System Usability Scale panel with adequate rigour, and running it badly would have produced a number with no evidential value.

It leaves NFR-06 specified but unverified, which Section 6.6 names as the weakest point in the evaluation and Section 7.5 carries into the future recommendations. The evaluation effort that was available went to the two claims that could not be established any other way, concurrency integrity and access control, rather than being spread thinly.

Has this actually improved coordination at a real hotel?

No, and the thesis does not claim it. The evaluation establishes that the artefact conforms to a specification derived from the evidence, and that the concurrency and access guarantees hold. It does not establish realised coordination effectiveness in the field.

Demonstrating that would require deploying across a season and measuring procurement variance against actual attendance, the proportion of events running under capacity, and advance turnout prediction accuracy, each against a pre-deployment baseline. That is the first future recommendation in Section 7.5.

Your tourist sample is 97 percent domestic. Does that not undermine the findings?

It bounds them rather than undermining them. The traveller findings generalise to the domestic young-adult traveller segment, not to the international visitor population that Chapter 1 also identified, and the thesis states that explicitly.

Replicating the traveller survey with international visitors is a named future recommendation. The prior expectation is that centralized information matters more for travellers with less local knowledge of a destination, not less, so the direction of the finding is unlikely to reverse.

What would you do differently?

Four things, all of them already in Chapter 7. Start traveller collection earlier and push much harder on the international segment. Reword the reverse-coded items or replace them with a different attention check. Build the artefact earlier so a usability panel fits inside the window. And use a longer instrument for the outcome so the three sub-dimensions can be scored separately instead of collapsing into one.

What did you personally learn?

That the most valuable result was one I did not expect. I assumed real-time visibility would dominate, because it is the most visible failure to anyone standing in Mirissa on an event night. The data said otherwise, twice, in two independent samples, and learning to treat that as the finding rather than as a problem with the data was the most useful shift in the project.

The second thing is that instrument design, statistical inference and software architecture turned out to be one continuous piece of work. The decision to put three resource kinds in one table is not a technical preference, it is the regression result written as a schema.

The weak spots, and the honest answer to each

All seven of these are already in your thesis. Naming one before an examiner finds it converts an attack into evidence of judgement.

1. 104 responses against a planned 400

Say: the organiser half holds up at roughly half the qualifying population in the collection belt, so it keeps its near-census character. The traveller half is the real shortfall, which is why non-significant results are reported as a failure to detect rather than as evidence of absence.

2. The traveller sample is narrow

Say: 96.9 percent domestic and all aged 18 to 34. The findings generalise to the domestic young-adult segment, not to international visitors. Replication with international visitors is a named recommendation.

3. The outcome is perceived, not realised

Say: the system did not exist at survey time, so nobody could report a real improvement. Behavioural intention is a theoretically grounded proxy under the Technology Acceptance Model. It is defensible and it is a proxy, and the thesis says both.

4. Both reverse-coded items failed

Say: a documented measurement artifact, not bad data. It hit exactly the two reversed items on each instrument and nothing else, in two independent samples. The standard remedy was applied and reported in full.

5. No usability study

Say: do not minimise it. Running it badly would have produced a number with no evidential value. It is named as the weakest point of the evaluation in Section 6.6 and carried into 7.5.

6. The outcome is one dimension, not three

Say: the evidence decided. It is scored as one composite and the three sub-dimensions are kept as conceptual facets. A longer instrument would be needed to separate them.

7. Three organiser scales just under 0.70

Say: 0.691, 0.690 and 0.665. Values between 0.60 and 0.70 are conventionally acceptable for exploratory work and short scales, because the formula penalises brief scales. These are 4-item scales at n equals 40, and the organiser results are reported as exploratory throughout.

One thing in the document to be ready for

A small internal inconsistency an examiner could spot

Table 3.2 in Chapter 3, the variable operationalisation table, lists the item counts as 5 for IV1, 5 for IV2, 5 for IV3 and 6 for the outcome.

Table 4.2 in Chapter 4, the item mapping table, gives the actual allocation: items 5 to 9 for IV1, which is 5 items, items 10 to 13 for IV2, which is 4, item 14 as a supplementary frequency measure, items 15 to 18 for IV3, which is 4, and items 19 to 23 for the outcome, which is 5.

The two tables do not match. Table 3.2 was written at the planning stage in Chapter 3, before the instrument was finalised, and Table 4.2 reports what was actually administered and analysed.

If asked, say: Table 3.2 states the planned allocation from the design stage. Table 4.2 reports the instrument as it was actually administered, and every reliability, correlation and regression result in the thesis is computed on the Table 4.2 allocation. Appendix A has the instrument itself, which confirms it.

Do not be defensive about it. Acknowledge it plainly and point at where the real numbers come from.

How to handle a question you cannot answer

Three safe moves, in order of preference

One, answer from the evidence. "The regression gave a beta of 0.665 for that predictor, and it was the only significant one in either sample." Facts you know are always a safe landing.

Two, point at where it lives. "That specific figure is in Appendix B, which carries the full statistical output for both samples." Knowing where something is counts as knowing your document.

Three, say what you would need to answer it. "I cannot answer that from this data. It would need a post-deployment study measuring procurement variance against actual attendance, which is my first future recommendation." That is a strong answer, not a weak one.

What never to do

Never invent a number. If you cannot remember it, say you cannot remember it and say where it is. A wrong figure stated confidently is far worse than an admitted gap, because it makes every other figure you gave suspect.

Never argue with a limitation. If an examiner names a weakness that is real, agree with it, then say what you did about it and what you would do next.

Never overclaim. The system has not improved a real hotel yet. Saying it has, even loosely, undermines everything else.

The night before

  1. Read the numbers page onceJust the twelve key numbers. Not all of them.
  2. Say your opening out loud twiceNot in your head. Out loud, so your mouth knows it.
  3. Open the deployed system and confirm it worksLog in as both an organiser and a traveller. Make one test booking.
  4. Have the fallback readyScreenshots on your machine, in a folder you can open in three seconds.
  5. Re-read Chapter 7 Section 7.3The self-reflection. It is where the best answer to "what did you learn" lives, and it is in your own voice already.
  6. SleepYou know this material. You did the work.
The last thing to remember

Your strongest asset is not any single number. It is that your thesis is honest.

Two of your three hypotheses failed and you reported it. Two of your survey questions failed and you reported it. Your sample fell short and you reported it. Seven bugs appeared during testing and you reported them. One measurement sample was dropped and you said so.

Very few undergraduate theses do that. Every one of those admissions is a place where an examiner can see that the work is real. Lean on it rather than hiding from it.