Steps of research
Research moves in a fixed order: identify the problem, review literature, frame hypotheses, design and sample, collect and analyse data, then interpret and report.
The research process
Research is cyclic: the report usually raises the problem for the next study.
Kinds of hypothesis
| Type | Meaning | Example |
|---|---|---|
| Null (H0) | No difference or no relationship; the one tested statistically | There is no difference in the scores of online and offline learners |
| Alternative or research (H1) | A difference or relationship exists | Online and offline learners differ in scores |
| Directional | States the direction of the difference | Online learners score higher than offline learners |
| Non-directional | States a difference without its direction | Online and offline learners differ in scores |
Sampling techniques
| Family | Technique | Key idea |
|---|---|---|
| Probability | Simple random | Every member has an equal chance (lottery, random numbers) |
| Probability | Systematic | Every kth member from a list |
| Probability | Stratified | Divide into strata, then sample randomly from each |
| Probability | Cluster | Randomly select whole groups |
| Non-probability | Convenience | Whoever is easy to reach |
| Non-probability | Purposive (judgement) | Chosen for a reason that fits the study |
| Non-probability | Quota | Fixed numbers per category, filled non-randomly |
| Non-probability | Snowball | Each participant refers the next; for hidden groups |
Errors in hypothesis testing
- Type I error (α): rejecting a true null hypothesis. The level of significance (0.05 or 0.01) is its probability.
- Type II error (β): accepting (failing to reject) a false null hypothesis.
- Power of a test = 1 − β, the chance of rightly rejecting a false H0.
- A good hypothesis is testable, clear, specific, consistent with known facts and limited in scope.
NET gives jumbled steps to sequence, asks you to classify sampling techniques as probability or non-probability, and tests the null hypothesis and Type I/II errors in statement form.
Typical MCQ: “Which is the first step of research?” Identifying and defining the research problem. Everything else, including the literature review, follows from it.
Typical MCQ: “Rejecting a null hypothesis when it is actually true is which error?” A Type I error. Its probability is the level of significance (α).
Order hook: 'Please Read Hard, Do Some Careful Analysis, Interpret': Problem, Review, Hypothesis, Design, Sample, Collect, Analyse, Interpret. Type I = an innocent convicted (a true H0 rejected).