Research Basics for Nurses
On this page
Direct answer
Evidence-based nursing practice begins with a well-built question — the PICOT format: Population, Intervention, Comparison, Outcome and Time frame — and proceeds through a fixed research process: identify the problem, review the literature, state the hypothesis, choose the design, sample ethically, collect and analyse data, and disseminate findings. Quantitative designs test hypotheses with numbers (experimental, quasi-experimental, descriptive); qualitative designs explore meaning and experience (phenomenology, grounded theory, ethnography). Sampling divides into probability methods — simple random, stratified, cluster, systematic — and non-probability: convenience, purposive, quota, snowball. A p-value below 0.05 is the conventional significance threshold, and evidence climbs from expert opinion through case reports to cohort, case-control, randomised trials and finally systematic reviews with meta-analysis.
What you must remember
- PICOT dissected: Population (which patients), Intervention (what you do), Comparison (against what alternative), Outcome (what you measure), Time (over what period) — the difference between a searchable question and a vague wonder.
- Quantitative versus qualitative: numbers and hypothesis-testing versus meaning and lived experience; in a mixed-methods study both run, sequenced to answer different parts of one question.
- Research process in order: problem identification → literature review → objectives and hypothesis → design and tool development → sampling → ethical clearance → pilot → data collection → analysis → interpretation → dissemination and utilisation; within it, the independent variable is the cause or intervention, the dependent variable the measured outcome, and extraneous variables the confounders design and randomisation must control.
- Probability sampling: simple random (lottery, random tables), systematic (every kth case), stratified (within strata for representation) and cluster (whole groups such as villages).
- Non-probability sampling: convenience (whoever is available — most student projects, weakest generalisability), purposive (information-rich cases), quota (fixed proportions) and snowball (participants recruit peers — the method for hidden populations such as injection drug users).
- Validity and reliability: validity is whether the tool measures what it claims; reliability is whether it measures consistently — a bathroom scale that is always two kilograms wrong is reliable but not valid.
- Significance and evidence: p below 0.05 means a result this extreme is unlikely by chance alone (conventionally, under 5 per cent); the evidence hierarchy peaks with systematic reviews and meta-analyses, then RCTs, cohort, case-control, cross-sectional and case reports.
- Ethics non-negotiables: ethics committee approval before any participant is approached, informed documented consent, confidentiality, the right to withdraw without penalty, protection of vulnerable groups — under the ICMR National Ethical Guidelines (2017) and the Declaration of Helsinki.
From a ward irritation to a defensible study
A nurse notices that ICU patients on standard two-hourly repositioning still develop sacral pressure injuries, and wonders whether an individualised schedule with skin assessment works better. PICOT organises it: in adult ICU patients (P), does an individualised repositioning and skin-inspection protocol (I) compared with fixed two-hourly turning (C) reduce pressure injury incidence (O) over three months (T)? The literature review finds trial evidence and a gap in Indian ICUs; a quasi-experimental comparison between two units is chosen, since randomising within one ward is impractical. Sampling and sample-size follow, a validated risk scale is selected, and ethics committee approval with consent precedes any data. Dissemination means a manuscript and — the step that matters most — a rewritten unit protocol on the ward noticeboard.
The Indian context
Every MSc Nursing student in India completes a dissertation under INC requirements, and post-basic students a smaller study, which makes this page the syllabus of thousands of vivas each year — expected to name the ICMR 2017 guidelines, define snowball sampling and read a p-value correctly. India-specific resources strengthen real projects: NFHS data for community questions, HMIS and IDSP data for health-system questions. Ward-level research literacy has a second job: reading evidence critically enough to change practice — questioning a ritual dressing change, checking the trial behind a new protocol, appraising a drug advertisement backed by a study of twelve patients. Publication ethics applies to nurses as to anyone: no plagiarism, no gift authorship, no fabricated data.
Frequently asked questions
What does PICOT stand for?
Population, Intervention, Comparison, Outcome and Time — the five elements that convert a clinical question into one that the literature can actually be searched for.
What does a p-value below 0.05 mean?
That a result at least this extreme would occur by chance alone less than 5 per cent of the time if there were truly no effect — the conventional threshold for statistical significance.
When is snowball sampling the right choice?
When the population is hidden or hard to reach — injection drug users, sex workers, homeless persons — with existing participants recruiting others from within their networks.
How does probability differ from non-probability sampling?
Probability sampling gives every population member a known chance of selection, permitting generalisation; non-probability methods select by availability or judgement and cannot.
Which guidelines govern research ethics in India?
The ICMR National Ethical Guidelines for Biomedical and Health Research Involving Human Participants (2017), operationalised through ethics committee approval, alongside the Declaration of Helsinki internationally.