Thesis Supervision Guide: Proposal, Methods and the Defense

Most thesis topics begin as a wide interest: digital marketing, student mental health, river water quality. Something that broad cannot be researched in a single semester, so the student's first job is to filter it. The first filter asks whether the data is within reach, meaning respondents can be found, documents can be accessed and any institutional permission can realistically be obtained in time. The second asks whether a theory exists to build on. The third asks whether the problem still leaves room for new work, which the student checks by reading five to ten recent theses and journal articles on similar themes.

A topic that clears all three filters is then written as one sentence containing the variables, the subjects and the setting. An interest in digital marketing, for instance, narrows into the effect of short-form video content on the purchasing decisions of university students in one city. A sentence like that tells the supervisor at a glance what is being measured, among whom and where, so the first supervision meeting can move straight into substance.

Data access is where students most often get caught. Plenty of them discover that company or agency data is hard to obtain only after the proposal has been approved, and end up changing topic halfway through. A habit that helps is contacting likely data sources before submitting the title, and bringing the draft topic to the supervisor early so its direction fits the supervisor's own field.

Research questions are what the study sets out to answer, phrased as clear and measurable questions. Two to four of them are generally enough for an undergraduate thesis. Each question has a partner among the research objectives, written as a statement with parallel content. If a question asks about the effect of two variables, the matching objective names the testing of that effect, and the significance section explains who benefits from the answer, both academically and in practice.

Mismatches here are among the most frequent revision notes at the proposal seminar. An exercise that prevents them is a three-column table: research question, objective and planned analysis. Every row should be complete. A row with an empty analysis column marks a question that has no way of being answered yet, and an objective with no matching question marks a promise the study will never keep.

The verbs in each research question deserve attention from day one. Phrases like "how strongly does X affect" or "is there a relationship between" point toward measurement, while "how do participants experience" or "why" point toward exploring meaning. That choice of wording later decides the approach in Chapter 3, so students who write it deliberately save themselves several rounds of revision.

Chapter 3 opens with the type and approach of the research. A quantitative approach rests on positivism and works with numbers, populations, samples and statistical tests, which suits a student who wants to measure effects or relationships between variables. A qualitative approach is naturalistic, treats the researcher as the key instrument and suits deep exploration of meaning, process or experience. Mixed methods combine the two when a research question needs figures and explanation together.

The clearest clue sits in the verb of each research question. "Measure", "test" and "how much" lead to quantitative designs such as correlational studies or experiments. "Understand", "describe the experience of" and "how does the process unfold" lead to qualitative designs such as case studies or phenomenology. A quantitative title might read: the effect of learning motivation and campus environment on student achievement. A qualitative one: what independence means to first-year students living away from home. A mixed design: a large survey on e-learning effectiveness followed by interviews with lecturers.

A common mistake is borrowing the research type from a senior student's thesis because it looks easier, then realising later that the approach fails to answer one's own question. A safer habit is writing one sentence that justifies the chosen approach, backed by a methods reference such as Sugiyono or Creswell. That sentence later becomes the core of the answer in the defense room, when an examiner asks why this method was chosen.

Operational definitions translate abstract concepts from Chapter 2 into something measurable. For each variable, the student writes a conceptual definition, breaks it into indicators and then sets the scale. Learning motivation, for example, might be measured through persistence, interest and goals on a one-to-five Likert scale. This section reads best as a table listing variables, indicators and item numbers, and each indicator should be covered by more than one item.

Once drafted, the instrument is piloted before it reaches the real sample, ideally with around 30 respondents who resemble that sample. A validity test checks whether each item measures what it claims to, usually by correlating item scores with the total score. A reliability test checks consistency, most often with Cronbach's alpha. Values above 0.70 are widely accepted as adequate internal consistency, and values of 0.80 and above indicate stronger reliability.

Items that fail the validity test are reworded or dropped, and the operational definitions table is updated to match. The pilot results go into Chapter 3 as evidence that the instrument is fit for use. In qualitative work, the equivalent is an interview guide derived from the research focus, a plan for source or method triangulation to protect the trustworthiness of the data, and a log of the date and length of every interview as an audit trail.

The usual defense format at Indonesian universities is a 10 to 15 minute presentation followed by a question session that typically runs longer, in front of two or three examiners other than the supervisor. Examiners come to the manuscript with fresh eyes, so the student needs to be ready to explain the research from the beginning. Assessment generally covers three components: command of the material, soundness of the methodology and conduct during the defense. Preparation fits into seven days, starting with a full reread of the manuscript and a one-page note that condenses each chapter into three to five points.

Slides follow a rule of roughly one per minute, moving through problem, method, findings and conclusion. Three questions come up almost every time: why this method, what the research contributes and where its limits lie. Answers to all three are written down and rehearsed until each fits into 60 to 90 seconds. Running the presentation at least three times with a timer and a listener who fires off unexpected questions, and recording it for review, helps the student spot habits such as speaking too fast.

When criticism comes, a strong answer follows a simple pattern: acknowledge the point, refer to the data in the manuscript, then name the fix. For a question outside one's preparation, the student asks for clarification and answers within the limits of the data. After the defense comes a revision window with a fixed deadline. Examiner feedback is written down in the room, the deadline and approval procedure are confirmed with the academic office the same day, and the work schedule is planned backwards from that date.