research methods and evidence-based practice

If you are a nursing, health, or social work student — or any practitioner who has to defend a decision with more than 'my supervisor said so' — this is your map. It shows how to turn a vague worry into a focused question, find the evidence, judge its quality, and apply it to a real person. It also teaches you to read research critically, so conflicting advice stops paralysing you.

1. 1Chapter 1

   Conflicting advice is the problem evidence-based practice was built to solve

   A woman is diagnosed with breast cancer. One specialist tells her one thing, another tells her something different, and she is left holding two contradictory treatment plans with no way to tell which is right. The book opens with the story of a woman diagnosed with breast cancer who received conflicting treatment advice, illustrating the need to evaluate evidence critically. The way out is a definition — evidence-based practice means 'the integration of best research evidence with clinical expertise and patient values' (Sackett et al., 2000) — and a five-step routine: ask a focused question, acquire the evidence, appraise it, apply it, assess the outcome.

   The idea

   The breast cancer story works because it strips away the fantasy that experts always agree. Two competent clinicians can read the same field differently, and a frightened patient cannot arbitrate between them. Evidence-based practice is not a promise that research has one right answer; it is a discipline for finding, judging, and combining what the evidence says with the clinician's skill and the patient's own values. That last part surprises people — patient values are not a soft add-on, they are inside the definition.

   Try this: Take one clinical question from your own work this week and run it through the five steps on paper: ask, acquire, appraise, apply, assess.

2. 2Chapters 2 and 4

   PICO turns a vague worry into a question research can answer

   The book presents the PICO framework for formulating research questions. Fill in the four boxes and the fuzzy concern becomes a sentence a database can actually hunt for. The book introduces the PICO framework for formulating clinical questions: Patient, Intervention, Comparison, Outcome. The PICO framework is used for formulating research questions.

   The idea

   Most failed literature searches fail before they start, because the question was never a question. PICO forces you to commit: which population, which intervention, compared to what, measured how. The discipline is uncomfortable — you have to name your comparison and your outcome, which vague questions let you dodge. Once the four parts are fixed, searching, appraising, and applying all have a target. It is the smallest tool in the book and possibly the one you will use most.

   Try this: Write your current professional question in four lines — Population, Intervention, Comparison, Outcome — and rewrite it until each line is specific.

3. 3Chapter 3

   Before any data exists, an ethics committee decides if you may collect it

   The book discusses the role of Human Research Ethics Committees in approving research. The book discusses informed consent, voluntary participation, and the right to withdraw at any time without penalty. The book discusses vulnerable groups in research, including children and people with cognitive impairment, and the need for special ethical consideration when researching them. Underneath all the checks sits one principle, beneficence — the book highlights the ethical principle of 'beneficence', meaning the obligation to maximise possible benefits and minimise possible harms to participants.

   The idea

   Ethics is not a signature page at the back of a thesis; it is a gate at the front. The committee's job is to imagine harms the researcher might be too enthusiastic to see, which is why approval comes before recruitment, not after. Beneficence is a double duty — you must actively seek benefit and actively limit harm, not merely avoid malice. And vulnerability is situational: the same consent form that is fine for a healthy adult may be inadequate for a child or a person with cognitive impairment.

   Try this: Draft a participant information sheet that states, in plain words, that withdrawal is allowed at any time without penalty — then check whether you would sign it yourself.

4. 4Chapter 4

   The evidence pyramid: systematic reviews on top, expert opinion at the bottom

   The book explains the hierarchy of evidence, with systematic reviews and meta-analyses at the top and expert opinion at the bottom. The book explains the hierarchy of evidence, with systematic reviews and meta-analyses at the top and expert opinion at the bottom. To understand what earns a trial its place, the book covers experimental designs including randomized controlled trials (RCTs), quasi-experimental designs, and non-experimental designs such as cross-sectional, cohort, and case-control studies. The ranking is not snobbery; it reflects how well each design protects against being fooled.

   The idea

   The hierarchy answers the breast cancer patient's problem in structural terms: not all advice carries equal weight, and the design behind a claim tells you how much to trust it. Randomization is what separates an RCT from a quasi-experiment; observing groups over time (cohort), looking backward (case-control), or taking a snapshot (cross-sectional) each trade away some certainty. Expert opinion sits at the bottom not because experts are foolish but because experience alone cannot rule out coincidence, bias, or luck.

   Try this: Next time you read a health claim, find the study design behind it and locate it on the hierarchy before deciding how much it should move you.

5. 5Chapters 5, 6, and 8

   Five qualitative designs, five different lenses on human experience

   The book describes qualitative research designs including phenomenology, grounded theory, ethnography, case study, and narrative research. Pick the wrong lens and you answer a question nobody asked; pick the right one and the design itself starts telling you what data to gather and how.

   The idea

   Qualitative design is not one method with five names — each design carries its own logic about what counts as understanding. Phenomenology and narrative stay close to individual experience; ethnography and case study widen to context and setting; grounded theory aims at a theory that emerges from the data rather than preceding it. The surprise for newcomers is that the choice is made before any interviewing happens, and it constrains everything downstream: sampling, questions, analysis.

   Try this: For a topic you care about, write one sentence per design explaining what a phenomenological version of the study would ask that a grounded theory version would not.

6. 6Chapters 5, 6, and 8

   In qualitative research, you stop recruiting when the data stops teaching you

   The book explains sampling in qualitative research, including purposive sampling, snowball sampling, and theoretical sampling, and the concept of data saturation. The book explains sampling in qualitative research, including purposive sampling, snowball sampling, and theoretical sampling. The book explains the concept of data saturation.

   The idea

   Saturation flips the usual logic of sample size. In a survey, bigger is generally better; in a qualitative study, the goal is depth and the endpoint is redundancy — when the last three interviews echo the ones before, more data adds weight, not insight. Snowball sampling is especially useful for hard-to-reach groups, because trust travels through the chain. The trade-off is honest: these are non-random choices, so what you gain in depth you give up in statistical generalisability.

   Try this: If you run interviews, keep a running memo of what each new session adds; when a session adds nothing new, you have your saturation point.

7. 7Chapters 6 and 8

   Probability or convenience: how you sample decides who your results can speak for

   The book covers probability sampling and non-probability sampling. The book covers probability sampling (simple random, systematic, stratified, cluster). The book covers non-probability sampling (convenience, quota, purposive). The book discusses sample size and response rates.

   The idea

   The probability-probability line is the line between results that can generalise statistically and results that cannot. Stratified sampling protects small subgroups from being drowned out; cluster sampling trades precision for feasibility when a population is scattered. Response rates deserve more respect than they get: a 90% plan executed at a 40% response rate is a convenience study wearing a probability costume. Sample size is not a formality — it is part of the claim your data can support.

   Try this: Find one published survey in your field and work out which sampling method it used and what its response rate was — then judge how far its conclusions can travel.

8. 8Chapters 6 and 7

   A questionnaire is an instrument, and instruments can be faulty

   The book discusses data collection methods for quantitative research including questionnaires, surveys, structured observation, Likert scales, and physiological measures. The book discusses data collection methods for quantitative research including questionnaires, surveys, structured observation, Likert scales, and physiological measures. The book discusses validity and reliability of instruments.

   The idea

   Validity and reliability are different failures, and confusing them is a classic beginner's error. Reliability is consistency; validity is truth. A bathroom scale that reads two kilos heavy every morning is reliable but invalid; one that wanders randomly is neither. Likert scales and structured observation earn their place in the toolkit only when someone has checked both properties. Physiological measures look objective but still need the same interrogation — a device can be precise and still not capture the construct you care about.

   Try this: Take any scale you use at work and ask two questions of it: what exactly does it claim to measure, and has anyone shown it measures that consistently?

9. 9Chapters 7, 8, and 10

   Interviews, focus groups, participant observation: hearing people in their own words

   The book covers in-depth interviews, focus groups, and participant observation as primary qualitative data collection methods, with discussion of interview guides, open-ended questions. The book discusses recording and transcribing data.

   The idea

   The three methods capture different data: interviews give depth on individual experience, focus groups give the dynamics of shared talk, participant observation gives what people do rather than what they say they do. Open-ended questions are the point — a good interview guide is a map, not a script. Transcription is not clerical work; it is the moment talk becomes data, and choices made there shape every analysis that follows.

   Try this: Rewrite three closed questions from a survey you know as open-ended interview questions, and notice what each rewrite invites the person to add.

10. 10Chapters 5, 7, and 8

    Mixed methods: neither number nor narrative alone is enough

    The book defines mixed methods research as the combination of quantitative and qualitative approaches in a single study or series of studies. The book defines mixed methods research as the combination of quantitative and qualitative approaches in a single study or series of studies, and describes designs such as convergent parallel, explanatory sequential, and exploratory sequential.

    The idea

    The sequencing is the design. Run the strands together and you get two views of the same moment; run one after the other and the second phase exists to serve the first. Explanatory sequential suits the moment your survey finds a pattern it cannot explain; exploratory sequential suits the moment you do not yet know enough to write good survey items. Mixed methods is not a compromise — it is a deliberate architecture where each strand covers the other's blind spots.

    Try this: Sketch a mixed methods design for one of your own questions: which strand goes first, and what exactly will the second strand explain or explore?

11. 11Chapters 9, 10, and 12

    Statistics: from mean and p-value to what actually differs

    The book covers descriptive statistics and inferential statistics. The book covers descriptive statistics (mean, median, mode, standard deviation). The book covers inferential statistics (t-tests, ANOVA, chi-square, correlation, regression). The book discusses levels of measurement and statistical significance (p-values).

    The idea

    Descriptive and inferential statistics answer different questions: what does my data look like, and can I say something about a population I never fully observed? Levels of measurement decide which tests are even legal — a mean of ranked categories is nonsense. The p-value is the most used and most misused number in the toolkit: it flags how surprising your result would be if there were truly no effect, nothing more. Choosing the test is not maths anxiety; it is matching the test to the question and the measurement level.

    Try this: Take one dataset column and identify its level of measurement, then name the one descriptive statistic and one test that fit it.

12. 12Chapters 9 and 14

    Trustworthiness: the qualitative answer to 'but is it valid?'

    For qualitative research, the book discusses trustworthiness. The book discusses trustworthiness, including credibility (e.g., member checking), transferability (thick description), dependability (audit trail), and confirmability (reflexivity). The book maps them to quantitative concepts of internal validity, external validity, reliability, and objectivity.

    The idea

    The mapping is the argument: qualitative work faces the same four threats as quantitative work, it just answers them differently. Member checking is credibility's version of a validation study; thick description lets readers transfer findings themselves instead of claiming statistical generalisation; an audit trail makes the study's path inspectable; reflexivity replaces the fiction of a neutral observer with an honest account of the observer. Rigour is not the absence of a human researcher — it is the documentation of one.

    Try this: Start an audit trail for your next project: a dated log of every design and analysis decision, written as you make it, not reconstructed later.

13. 13Chapters 10, 11, and 13

    Thematic and content analysis: turning transcripts into findings

    The book covers thematic analysis, content analysis, and grounded theory analysis. The book covers thematic analysis, including coding and theme development. The book covers content analysis. The book covers grounded theory analysis. The book discusses the use of software such as NVivo.

    The idea

    Coding is the hinge: the first pass labels the material, and themes are built upward from those labels, not imposed from above. The three approaches differ in destination — themes describe, content analysis systematises, grounded theory theorises. NVivo speeds the mechanics of retrieval and sorting, which is genuinely valuable with large datasets, but the interpretive work stays human. A theme is not a topic that came up often; it is a pattern that says something about the question.

    Try this: Code one page of an interview transcript yourself, then group your codes into candidate themes and check whether each theme says something your research question cares about.

14. 14Chapters 11 and 15

    The systematic review: a protocol, not a pile of papers

    The book describes the steps of a systematic review: formulating the question, searching the literature, appraising studies, synthesizing findings, and reporting results. The book mentions the Cochrane Collaboration and PRISMA checklist. Done this way, the review is reproducible — another researcher could follow the same path and check the work.

    The idea

    A systematic review differs from a narrative review the way a protocol differs from an essay: the steps are declared in advance, the search is documented, and the appraisal of each study is explicit. That is why systematic reviews and meta-analyses sit at the top of the evidence hierarchy — they are research about research, with the bias of cherry-picking designed out. PRISMA matters because transparency is the product; a review you cannot audit is just a well-read opinion.

    Try this: Take a question from your field and draft the search string you would use — then check whether it covers the population and intervention your PICO named.

15. 15Chapters 12, 14, and 17

    IMRaD: the shape a finished study must take

    The book discusses the structure of a research proposal and the structure of a research report. The book discusses the structure of a research proposal. The book discusses the structure of a research report (IMRaD: Introduction, Methods, Results, and Discussion). The book discusses publishing in journals and disseminating findings.

    The idea

    IMRaD is not a formatting rule; it is an argument in four moves, each of which a reader can attack separately. Methods written tightly enough let someone replicate or critique the work; results separated from discussion keep findings from being smuggled in as interpretation. The proposal and the report are different genres — one sells a plan, the other confesses an outcome. Dissemination is the step researchers skip at their peril: unpublished evidence helps no patient, and evidence-based practice runs on published evidence.

    Try this: Take a paper you admire and mark where Introduction ends, Methods begin, Results end, and Discussion starts — then notice how each section answers a different reader question.

16. 16Part 3, Chapter 16

    Case studies: evidence-based practice with its sleeves rolled up

    The final chapters present case studies of evidence-based practice in nursing, allied health, and social work settings. The final chapters present case studies of evidence-based practice in nursing, allied health, and social work settings, showing how practitioners find, appraise, and apply research evidence to clinical decisions. The final chapters present case studies of evidence-based practice in nursing, allied health, and social work settings.

    The idea

    The case studies close the loop the breast cancer story opened: the same five-step discipline that could have resolved her conflicting advice is shown running in ordinary professional life. What the cases add is texture — appraisal is not abstract when the decision is about a real person, and applying evidence means fitting research to a specific patient's circumstances. Across nursing, allied health, and social work, the method holds; only the content changes.

    Try this: Pick one decision from your own practice this month and write a one-page case study of it: the question, the evidence you found, your appraisal, and what you did.

Nursing, allied health, and social work students who must find, judge, and apply research evidence, and practitioners who want to defend clinical decisions with more than habit. Also anyone starting a research project who needs the whole pipeline — ethics to dissemination — in one book.

FAQ