From research questions to survey items
Start from your research questions and hypotheses, not from a list of questions you think of. For each variable, decide how you will measure it, and then write or adapt items for it.
Whenever possible, adapt validated scales from published studies and cite them, because they already have evidence of reliability. Write each item as a single idea, in plain language and without leading or double-barrelled wording ("Is the lecturer knowledgeable and friendly?" asks two things). Keep the questionnaire as short as the research allows.
Scales and question types
Likert scales (for example 1 = strongly disagree to 5 = strongly agree) are common for attitudes and perceptions. Label all points clearly and keep the direction consistent. Other types include multiple choice, ranking, yes/no and open-ended questions.
Use demographic questions at the end unless they are needed for screening, and make sensitive questions optional where you can. Keep layout clean, and offer the survey in the language your respondents understand, with a proper translation check if you translate.
Sampling methods
Probability sampling gives every member of the population a known chance of selection: simple random, systematic, stratified and cluster sampling. It supports generalising to the population but needs a sampling frame.
Non-probability sampling is used when a frame is not available: convenience, purposive, quota and snowball sampling. It is common in student research but limits how far you can generalise. Match the method to your objectives, explain your choice and acknowledge its limitations.
Sample size basics
Required sample size depends on the population, the analysis and the precision you want. For descriptive surveys, tables or formulas such as Krejcie and Morgan (1970) or Cochran's formula give a number for a given population. For inferential tests, power analysis tools such as G*Power estimate the sample needed for a chosen effect size, power and significance level.
For regression or SEM, there are rules of thumb that depend on the number of predictors or paths. Check what your supervisor and method sources recommend, and plan for non-responses and incomplete answers.
Pilot testing and reliability
Before the main survey, test the questionnaire with a small group similar to your target respondents. Ask whether items are clear, how long it takes and where people hesitated. Revise unclear or biased items.
For multi-item scales, check reliability, often with Cronbach's alpha, where values of about 0.70 or higher are commonly treated as acceptable for research. Reliability is only part of quality, so also consider whether items cover the construct (validity), and have an expert review the questionnaire.
Ethics and consent
If you collect data from people, check whether your university requires ethics approval. Provide a short information sheet that explains the purpose, what participants will do, how long it takes and how data will be used and stored.
Participation must be voluntary and informed, and participants should be able to withdraw. Keep responses anonymous or confidential, collect only the data you need and store it securely. Do not include data that identifies people unless it is necessary and approved. Malaysia's Personal Data Protection Act may apply, so ask your university for guidance.