Research Methodology
How you were going to find the answer, written down before you had any data. This is the chapter that makes the result trustworthy.
Chapter 3 is your plan, written down before you collected a single answer. That ordering is the entire point. It is what stops anyone claiming you went looking for data that suited a conclusion you had already picked.
Pages 34 to 45.
The five decisions in this chapter
A methodology chapter is really five choices, each with a reason. Learn them as a sequence, because that is how they are usually asked about.
| No | The choice | Yours |
|---|---|---|
| 1 | Paradigm, what you believe knowledge is | Post-positivism |
| 2 | Approach, which direction reasoning runs | Deductive |
| 3 | Strategy, the overall kind of study | Design Science Research |
| 4 | Methods, how facts were collected | Two structured questionnaires |
| 5 | Execution, the order the work happened in | Peffers' DSRM, six phases |
1. Post-positivism
There is a real world out there and it can be measured. But we can only ever get close, never perfect, because instruments have error and researchers have biases.
The stricter position, positivism, claims perfect objectivity is possible. Post-positivism does not.
Why it fits your study: you are measuring relationships between defined variables using structured surveys and statistics, which is a measurement-based approach. But you also openly report that your sample fell short, that two questions failed, and that your outcome is a proxy. That combination, measure seriously and admit the limits, is exactly post-positivist.
2. Deductive approach
You started from existing theory, wrote three hypotheses, then collected data to test them.
The theory came from two places. The Technology Acceptance Model, which says people adopt a system when they think it is useful and easy. And the Smart Tourism Destination perspective, which says destinations improve when digital networks deliver real-time information.
H1. Real-time Event Visibility has a statistically significant positive effect on Coordination Effectiveness.
H2. Pre-booking and Resource Coordination has a statistically significant positive effect on Coordination Effectiveness.
H3. Centralized Information Access has a statistically significant positive effect on Coordination Effectiveness.
A hypothesis counts as supported if its beta coefficient is significant at p below 0.05.
All three hypotheses were written in Chapter 3, before any data existed. H1 and H2 were not supported. That is a legitimate finding produced by a fair test, not a failure of the study.
If you had written the hypotheses after seeing the data, the result would be worthless. The order is the value.
3. Design Science Research
Design Science Research is the strategy where you build an artefact to address a real problem and then evaluate it.
Chapter 3 justifies it by elimination, which is the strongest way to justify a choice:
A behavioural study alone
Would have explained why coordination fails, and produced nothing that addresses it. No artefact.
A pure software project alone
Would have produced a working system with no evidence that it solves the right problem. No validation.
Design Science Research
Requires both. Build the thing, and prove with data that it addresses something real. That is why your thesis contains statistics and code without looking like two projects glued together.
4. How the facts were collected
Two structured questionnaires, distributed through Google Forms, to two different groups.
| Survey | Who | Target | Actual | Sampling |
|---|---|---|---|---|
| Tourist survey | Local and international travellers, recruited mainly at Southern Province coastal destinations | 350 | 64 | Convenience with a screening question |
| Organizer survey | Hotel managers, event coordinators and hospitality staff responsible for events | 50 | 40 | Purposive |
Where the target numbers came from
It comes from the Krejcie and Morgan table, a published lookup table that gives the minimum sample size for a given population size. For a population of roughly 1,000 to 10,000, it says 350.
It also gives enough statistical power for a regression with three predictors at a medium effect size.
Only 64 came in. That shortfall is reported without adjustment, and it is why your thesis says H1 and H2 were not detected rather than do not exist.
Different logic entirely. The number of properties in the collection belt that actually host social events regularly is estimated at 60 to 80, from field observation and tourism authority registration data.
So 50 would be a near-census, meaning almost the whole qualifying population rather than a sample of it.
You got 40. That is still roughly half the qualifying population, which is why the organiser sample holds up much better than the tourist one.
Both surveys use a 5-point Likert scale, from strongly disagree to strongly agree.
Two things were added after the supervisor rejected the first draft: two open-text questions on each survey, and reverse-coded items inside each construct to catch people agreeing with everything without reading.
The open questions turned out to be one of the strongest parts of the thesis. The reverse-coded ones both failed, which is discussed in Chapter 4.
5. The execution workflow, DSRM
Peffers' Design Science Research Methodology gives six phases. Table 3.1 in your thesis maps each one onto a chapter. This is a very common viva question, so learn the six words in order.
- Problem identificationIdentify the challenges in event promotion and resource booking.Chapters 1 and 2
- Define objectivesEstablish what the platform must support.Chapter 1
- Design and developmentBuild the prototype with Next.js and Supabase.Chapter 5
- DemonstrationExercise the prototype in scenarios covering high turnout and booking conflict.Chapters 5 and 6
- EvaluationAnalyse the survey data statistically and test the artefact against its requirements.Chapters 4 and 6
- CommunicationCompile the findings into the thesis, and defend them.All chapters, and the viva
The analysis plan, decided in advance
Chapter 3 also fixes the order of the statistical work before any data arrived. This matters, because deciding the rules afterwards would let you pick whichever rule gave the nicest answer.
- Reliability firstCronbach's Alpha on every construct. The stated threshold is 0.70 for inclusion in the inferential analysis.
- Then correlationPearson bivariate correlation to see the direction and strength of each relationship, and to screen for predictors that overlap too much.
- Then regressionSimultaneous multiple regression to estimate each predictor's unique contribution, judged significant at p below 0.05.
Ethics
Four commitments, and they are short enough to memorise. If asked about ethics, give all four.
Voluntary participation
Consent is obtained on the first page of each questionnaire. Anyone can close the browser and leave at any point with no penalty.
Anonymity
The surveys collect no names, no email addresses and no phone numbers. All analysis is on aggregated data.
Secure storage
Responses stored on secure servers, extracts kept in encrypted folders, and all raw data deleted on graduation.
Academic honesty
All literature cited, and the prototype genuinely built and tested as described.
The ethics commitment is not just a paragraph. When the system was seeded with demonstration data, no real person's name, phone number or email appears in it. Chapter 6 states that explicitly and ties it back to this section.
The words in this chapter
Your underlying belief about what knowledge is and how it can be obtained. It comes first because it determines everything after it.
Post-positivist. Measurable reality, approached approximately.
Not random. You choose who to ask, either because they qualify (purposive) or because you can reach them (convenience).
Both your samples are non-probability, and Chapter 3 justifies why. For organisers the population is small, bounded and directly accessible, so a purposive near-census is more representative than a random subset would be.
A question at the start that filters out people who do not qualify.
Travellers were screened on whether they had attended, or planned to attend, a hotel-hosted event. Without it you would be surveying people with no view on the subject.
A way of managing work in short repeating cycles, usually two weeks, with a review at the end of each one.
Your project management method. Weekly progress reviews with the supervisor acted as sprint reviews, so feedback could change the design before too much had been built on it.
Numbers from this chapter
If they ask you
Why post-positivism and not interpretivism?
Because the research question asks about measurable relationships between defined variables, and about how strongly each one predicts an outcome. That needs quantitative measurement across a sample, not deep interpretation of individual meaning.
Post-positivism rather than strict positivism, because the study openly accepts that the instrument carries error, that the sample is limited, and that the researcher has assumptions. Those are acknowledged in the text rather than assumed away.
Why non-probability sampling? Does that not weaken the findings?
For organisers it strengthens them. The population of properties actively hosting social events in the collection belt is estimated at 60 to 80, so it is small, bounded and directly accessible. A purposive near-census of roughly half that population is more representative than a random sample of the same size would be.
For travellers, convenience sampling with a screening question was the practical option for intercepting visitors. It is a real limitation and it is reported as one. The finding is bounded to the domestic young-adult traveller segment rather than claimed for all visitors.
You set a reliability threshold of 0.70, but three of your organiser scales are below it. Explain.
They are at 0.691, 0.690 and 0.665, so marginally below. Values between 0.60 and 0.70 are conventionally accepted for exploratory research and for short scales, because Cronbach's Alpha is directly sensitive to the number of items and penalises brief scales. These are 4-item scales at n equals 40.
They are treated as acceptable on that basis, and the organiser results are reported throughout as exploratory relative to the tourist results. The decision and the reasoning are both in the text rather than hidden.
Walk me through your research methodology.
Post-positivist paradigm, deductive approach, Design Science Research as the strategy, and Peffers' six-phase DSRM as the execution workflow.
Data came from two structured Google Forms questionnaires on a 5-point Likert scale, one for travellers and one for organisers, sharing a common item structure so the same model could be tested twice independently. Organisers were sampled purposively as a near-census of the qualifying properties, travellers by convenience with a screening question.
The analysis sequence was fixed in advance: reliability with Cronbach's Alpha, then Pearson correlation, then simultaneous multiple regression, with significance at p below 0.05.
Chapter 3 is where I wrote the plan down before I had any data. I adopted a post-positivist paradigm and a deductive approach, so my three hypotheses were formalised before collection began. The strategy is Design Science Research, because a behavioural study alone would not have produced an artefact and a software project alone would not have validated the need. Execution followed Peffers' six-phase DSRM, mapped chapter by chapter. Facts came from two structured questionnaires, and the analysis sequence, reliability then correlation then regression, was fixed in advance rather than chosen once I saw the results.