Research Methodology cheat sheet

Key terms

Systematic

The steps follow a plan set out before the data is collected, not improvised as findings emerge. A systematic study can state, in advance, what evidence would change its conclusion. An unsystematic one discovers its method after it has seen the answer it wants.

Controlled

Competing explanations are ruled out deliberately rather than ignored. If a defect rate falls after a training programme, control means having some way of knowing that it was the training and not the new supplier, the new supervisor or the season.

Empirical

Conclusions rest on observation and measurement that others could repeat. A claim that cannot in principle be checked against evidence is not a research finding, however plausible it sounds.

Critical

The researcher actively looks for the weaknesses in their own work, states them, and invites others to do the same. Peer review, examiner scrutiny and replication all exist because self-deception in research is the normal case rather than the exception.

Concept

An abstract idea, such as quality culture, customer satisfaction or supplier reliability. Concepts cannot be measured directly, which is exactly why the next two terms exist.

Construct

A concept deliberately defined for research purposes so it can be observed. Quality culture becomes a construct once you specify that it consists of, say, management commitment, employee involvement and continual improvement behaviour.

Variable

A measurable property that takes different values across cases. Defect rate, training hours, years of service and ISO 9001 certification status are all variables.

Operational definition

The rule that converts a construct into a variable by stating exactly how it will be measured. Customer satisfaction operationally defined as the mean of five Likert items on a named questionnaire is researchable; customer satisfaction left undefined is not.

Population

The complete set of cases the study wants to say something about. Every tea factory registered with the Tea Board of Kenya; every batch produced in 2025; every internal auditor employed by the organisation.

Sample

The subset actually studied. The whole of sampling theory exists to make a statement about the population defensible from a sample.

Unit of analysis

The thing about which data is collected and conclusions are drawn: a person, a batch, a firm, a complaint, a shift. Confusing units is a common and fatal error. A study that samples 300 employees but concludes about 12 firms has 12 cases, not 300.

Data

Recorded observations. Primary data is collected by the researcher for this study; secondary data was collected by someone else for some other purpose and is reused.

Parameter and statistic

A parameter describes a population (the true mean defect rate of all batches); a statistic describes a sample (the mean of the 60 batches examined). Inference is the business of estimating the first from the second.

Reliability

Consistency. The same instrument applied again under the same conditions gives the same result.

Validity

Truth. The instrument measures what it claims to measure, and the conclusion follows from the evidence.