Introduction to Social Research Methodology

The nature and purpose of social research

Ben Stanley

Department of Social Sciences, SWPS University

October 13, 2026

Course outline

Course outline: session 1 (today) Course outline: 15 stations. 1 — Nature and purpose of social research (today); 2 — Research goals and research questions (to come); 3 — Hypotheses and research design (to come); 4 — Research ethics (to come); 5 — Conducting a literature review (to come); 6 — Secondary data analysis (to come); 7 — Qualitative approaches (to come); 8 — Quantitative approaches and measurement (to come); 9 — Mixed methods research (to come); 10 — Sampling and participant recruitment (to come); 11 — Questionnaire design and survey fieldwork (to come); 12 — Interviews and focus groups (to come); 13 — Quantitative data analysis (to come); 14 — Qualitative data analysis (to come); 15 — Reporting research; course review (to come). 1 Nature and purpose of social research TODAY 2 Research goals and research questions 3 Hypotheses and research design 4 Research ethics 5 Conducting a literature review 6 Secondary data analysis 7 Qualitative approaches 8 Quantitative approaches and measurement 9 Mixed methods research 10 Sampling and participant recruitment 11 Questionnaire design and survey fieldwork 12 Interviews and focus groups 13 Quantitative data analysis 14 Qualitative data analysis 15 Reporting research; course review

Today’s lecture

  • Foundations — what social research is, what it is for, and the main types
  • Qualitative and quantitative approaches — the central methodological divide, and how to bridge it
  • The research process — the stages of a study, from choosing a topic to acknowledging its limits
  • By the end, you should be able to read any study and ask: what were they trying to find out, and how did they go about it?

Foundations of social research

What is social research?

What makes social research different from everyday sense-making At the top, the definition: social research is the systematic investigation of human behaviours, attitudes, relationships and social systems, to produce reliable knowledge and insight. Below, a comparison in four rows. Procedure: everyday sense-making is casual, with its steps left unstated; social research is explicit, transparent and repeatable. Questions framed by: personal experience, against theory, meaning existing concepts and explanations. Claims settled by: intuition, anecdote and authority, against systematically gathered evidence. Aim: a personal impression, against describing, then explaining, then informing change. At the bottom: organisations run on it too; employee surveys, customer research and market analysis are social research applied to managerial decisions, and the same standards apply. Social research is the systematic investigation of human behaviours, attitudes, relationships and social systems, to produce reliable knowledge and insight Everyday sense-making Procedure Questions framed by Claims settled by Aim casual; steps left unstated personal experience intuition · anecdote · authority a personal impression Social research explicit · transparent · repeatable theory: concepts and explanations systematically gathered evidence describe → explain → inform change Organisations run on it too — employee surveys, customer research and market analysis are social research applied to managerial decisions, and the same standards apply

Objectives of social research

The objectives of social research as a ladder of increasing ambition Four steps rising from left to right. Step 1, describe: what is happening, to whom, and under what conditions. Step 2, explain: causes and consequences, moving from what to why. Step 3, predict: use established patterns to anticipate what comes next. Step 4, inform policy: the evidence on which governments, organisations and communities act. An arrow along the bottom reads: increasing ambition; each step builds on the one below. increasing ambition: each step builds on the one below 1 Describe what is happening, to whom, under what conditions 2 Explain causes and consequences: from what to why 3 Predict use established patterns to anticipate what comes next 4 Inform policy the evidence on which governments, organisations and communities act

Types of social research: two axes

Types of social research: two axes, four combinations A two-by-two grid. The columns are the first axis, the kind of data and reasoning: qualitative (deep, interpretative, concerned with meaning and experience) and quantitative (numerical, statistical, concerned with measurement and pattern). The rows are the second axis, whether the researcher intervenes: experimental (controls and varies conditions) and descriptive (observes only, without intervening). Every cell is a possible study. Qualitative and experimental: a new civic-education workshop is tried in some schools, and focus groups explore how pupils experienced it. Quantitative and experimental: half of voters, chosen at random, get a text-message reminder, and turnout in the two groups is compared. Qualitative and descriptive: in-depth interviews explore how first-time voters experienced an election. Quantitative and descriptive: a large survey maps how turnout varies with age, without intervening. Axis 1: what kind of data and reasoning? Qualitative deep, interpretative: meaning and experience Quantitative numerical, statistical: measurement and pattern Axis 2: does the researcher intervene? Experimental intervenes: controls and varies conditions Descriptive observes only: no intervention A new civic-education workshop is tried in some schools; focus groups explore how pupils experienced it Half of voters, chosen at random, get a text-message reminder; turnout in the two groups is compared In-depth interviews explore how first-time voters experienced an election A large survey maps how turnout varies with age, without intervening

Inductive and deductive logic

Deductive and inductive logic as two halves of one cycle Theory, the general, sits at the top; observations, the specific, at the bottom. On the left, the deductive path runs down: from theory to a hypothesis, then to data gathered to test it; it moves from the general to the specific and is typical of quantitative work. On the right, the inductive path runs up: from observations to patterns, then to theory built up from them; it moves from the specific to the general and is typical of qualitative work. In the middle: most research cycles between the two, and today's inductive insight becomes tomorrow's deductive test. Theory the general Observations the specific Hypothesis DEDUCTIVE general → specific start from theory, derive a hypothesis, gather data to test it typical of quantitative work Patterns INDUCTIVE specific → general start from observations, look for patterns, build theory up from them typical of qualitative work most research cycles between the two: today's inductive insight becomes tomorrow's deductive test

Qualitative and quantitative approaches

Qualitative research

Qualitative research: its focus, three methods, and the data it produces At the top: the focus of qualitative research is human experiences, meanings and interpretations, that is, how people make sense of their social world. Three pictograms show its common methods. In-depth interviews: a researcher and one participant in conversation, exploring one perspective in detail. Focus groups: a moderator and five participants around a table, with lines of interaction between them, showing how views form and shift through interaction. Ethnography: the researcher, notebook in hand, among the people of a setting over weeks or months, observing behaviour in context. At the bottom: the data are rich, detailed and contextual, usually from small samples; the approach is best for new, sensitive or poorly understood topics, where the right questions are not yet clear. Focus: experiences, meanings and interpretations — how people make sense of their social world ? In-depth interviews one perspective explored in detail Focus groups how views form and shift through interaction one setting, over weeks or months Ethnography immersion in a setting: behaviour observed in context Data rich, detailed and contextual — usually from small samples Best for new, sensitive or poorly understood topics, where the right questions are not yet clear

Quantitative research

Quantitative research: its focus, three methods, and the data it produces At the top: the focus of quantitative research is to quantify variables, that is, how they are distributed and the patterns and relationships among them. Three pictograms show its common methods. Surveys: one standardised questionnaire sent to a large sample, the same questions for everyone. Experiments: a treatment group and a control group, with a variable manipulated under controlled conditions and the groups compared. Statistical analysis: a scatterplot with a fitted line, drawn from existing numerical data. At the bottom: the data are standardised, so they can be summarised, compared and tested for statistical significance; with appropriate sampling the findings generalise to the population, at a cost in depth. Focus: quantify variables — how they are distributed, and the patterns and relationships among them Surveys the same standardised questions for a large sample treatment control compare Experiments a variable manipulated under controlled conditions Statistical analysis of existing numerical data Data standardised: summarised, compared, tested for statistical significance Pay-off with appropriate sampling, findings generalise to the population — at a cost in depth

Qualitative vs quantitative: a comparison

Qualitative Quantitative
Data Words, images, meanings Numbers, measurements
Logic Inductive — builds theory Deductive — tests theory
Samples Small, purposively chosen Large, ideally representative
Methods Interviews, focus groups, ethnography Surveys, experiments, statistics
Strength Depth, context, discovery Breadth, comparison, generalisation

Mixed methods: combining the two

Two ways of combining qualitative and quantitative research in one study Two sequences. First: interviews first, to design a better survey (the exploratory sequential design of session 9). Qualitative in-depth interviews to understand the phenomenon come first; their insights shape the questions of a quantitative survey that measures it across a large sample. Second: a survey first, to find cases for in-depth study (the explanatory sequential design of session 9). A quantitative survey establishes the pattern; its results pick the cases for qualitative in-depth case studies, which explore the mechanism. Interviews first, to design a better survey the exploratory sequential design (session 9): the inductive phase feeds the deductive one QUAL In-depth interviews understand the phenomenon first insights shape the questions QUAN Survey measures it across a large sample A survey first, to find cases for in-depth study the explanatory sequential design (session 9): breadth first, then depth where it is needed QUAN Survey establishes the pattern results pick the cases QUAL In-depth case studies explore the mechanism behind it

Discuss: which approach?

  • For each question, would you reach for a qualitative or a quantitative design — or both?
    • Why has turnout among under-25s fallen over the last decade?
    • How do first-time voters describe their experience of an election?
    • Does sending a text-message reminder increase the likelihood of voting?
    • Why are employees leaving a company at twice last year’s rate?
  • There is rarely a single right answer — the point is to match the method to the question
  • Managers face this choice constantly: a falling satisfaction score tells you that something is wrong; finding out why usually needs a different kind of evidence

The research process

The stages of a study

The stages of a study as a chain of seven links Seven interlocking chain links, numbered in order: 1, choose a topic (what is worth studying, and can it be studied?); 2, design the study (turn the topic into answerable questions); 3, sample (decide who or what to study); 4, collect data (gather the evidence); 5, analyse and interpret (make sense of the evidence); 6, present (communicate what was found); 7, reflect (acknowledge the limits). Each decision constrains the ones that follow, so the chain is only as strong as its weakest link. An orange thread runs through every link: the running example followed through all the stages, why do young people vote less than older people? 1 Choose a topic what is worth studying, and can it be studied? 2 Design the study turn the topic into answerable questions 3 Sample decide who or what to study 4 Collect data gather the evidence 5 Analyse and interpret make sense of the evidence 6 Present communicate what was found 7 Reflect acknowledge the limits RUNNING EXAMPLE, FOLLOWED THROUGH EVERY STAGE Why do young people vote less than older people?

Choosing a research topic

A good research topic meets three criteria at once Three overlapping circles; a good topic sits where all three overlap. Relevant: it speaks to a current issue or a gap in knowledge. Feasible: it is achievable with the time, resources, expertise and access available. Ethical: its benefits outweigh any risk of harm. The running example, why do young people vote less, passes all three: turnout among the young is falling; survey data on turnout and age already exist; and asking people about voting is low-risk. Relevant speaks to a current issue or a gap in knowledge Feasible achievable with the time, resources, expertise and access available Ethical its benefits outweigh any risk of harm a good topic RUNNING EXAMPLE Why do young people vote less? ✓ turnout among the young is falling ✓ survey data on turnout and age already exist ✓ asking people about voting is low-risk

Research design and planning

Research design as the blueprint between a topic and the evidence On the left, the topic: what to study. On the right, the evidence needed to answer it. Between them, the design, drawn as a blueprint with three tasks: 1, research questions or hypotheses, clear and realistically answerable; 2, scope, meaning what is studied, who is included, when and where; 3, methods and tools, qualitative, quantitative or a mixture, chosen to fit the questions. Running example: the topic "why do young people vote less?" is sharpened into the testable question "does political interest explain the age gap in turnout?" Topic what to study Design: the blueprint Evidence needed to answer it 1 Research questions or hypotheses clear, and realistically answerable 2 Scope what is studied, who is included, when, where 3 Methods and tools qualitative, quantitative or a mixture RUNNING EXAMPLE Why do young people vote less? a topic Does political interest explain the age gap in turnout? a testable question: the first task of the design

Variables and operationalisation

Variables, their roles, and operationalisation Top: a variable is anything that varies across cases, such as age, income, turnout or political interest. An arrow runs from the independent variable, the presumed cause, to the dependent variable, the presumed effect. In the running example, the independent variable is political interest, measured as a 0 to 10 self-rating, and the dependent variable is turnout in the last national election. Bottom: operationalisation turns an abstract concept into something measurable. The concept political engagement, which cannot be observed directly, is linked to three observable indicators: did you vote; how often do you discuss politics; are you a party member. A bracket around the indicators asks the question of validity: do the indicators really capture the concept? Variables — anything that varies across cases: age, income, turnout, political interest Independent variable the presumed cause shapes? Dependent variable the presumed effect Operationalisation — turning an abstract concept into something measurable Political engagement abstract: cannot be observed directly observable indicators Did you vote? How often do you discuss politics? Are you a party member? Validity: do the indicators really capture the concept? RUNNING EXAMPLE political interest: a 0–10 self-rating RUNNING EXAMPLE turnout in the last national election

Sampling techniques

Probability and non-probability sampling from the same population On the left, a population of 64 people, 16 of them younger voters (dark dots) and the rest older voters (light dots); the younger voters are concentrated in one corner, the part of the population that is easiest to reach. Probability sampling: every member has a known, non-zero chance of selection, for example simple random sampling; its 10 selections are scattered across the population, and the sample, with 3 younger voters, mirrors the population; it supports generalisation and statistical inference, but needs a complete list of the population. Non-probability sampling: participants are chosen by specific criteria or convenience, for example convenience, quota or snowball sampling; its 9 selections all come from the easy-to-reach corner, and the sample, with 7 younger voters, is skewed; it is quicker and cheaper, but carries a higher risk of bias. At the bottom: the central concern is representativeness, how well the sample reflects the population it is meant to describe. Running example: the population is all eligible voters, and a random sample of 1,500 lets us compare turnout across age groups. Population the whole group the study is meant to describe younger voters older voters Probability sampling every member has a known, non-zero chance of selection e.g. simple random sampling sample ✓ mirrors the population: supports generalisation and statistical inference, but needs a complete list of it Non-probability sampling participants chosen by specific criteria or convenience e.g. convenience, quota or snowball sampling sample ✗ skewed towards whoever is easiest to reach quicker and cheaper, but a higher risk of bias Representativeness: how well does the sample reflect the population it is meant to describe? ALL ELIGIBLE VOTERS 1,500 AT RANDOM

Data collection methods

Four methods of data collection, each with a strength and a limitation Four cards. Surveys: standardised questions for many respondents, via a structured questionnaire; strength, breadth and comparability; limitation, weak on depth and context. Observation: behaviours or events recorded as they occur, with or without participating; strength, what people actually do, not what they say; limitation, labour-intensive and hard to generalise from. Interviews: direct questioning, from tightly structured to open and conversational; strength, depth and nuance; limitation, time-consuming and shaped by the relationship between interviewer and interviewee. Archival sources: pre-existing records, documents and datasets; strength, unobtrusive, so suited to the past and to sensitive topics; limitation, limited to the records that happen to exist. No method is best in the abstract. Running example: a structured survey, outlined in orange, is the natural fit for measuring turnout across a large, representative sample. Surveys standardised questions for many respondents, via a structured questionnaire + breadth and comparability − weak on depth and context Observation behaviours or events recorded as they occur, with or without participating + what people actually do, not what they say − labour-intensive; hard to generalise from Interviews direct questioning, from tightly structured to open and conversational + depth and nuance − time-consuming; shaped by the relationship Archival sources pre-existing records, documents and datasets + unobtrusive: the past, sensitive topics − limited to the records that happen to exist RUNNING EXAMPLE: A STRUCTURED SURVEY

Ethics in social research

Three ethical principles, and ethical review before data collection Three cards. Informed consent: participants understand what the research involves and agree freely, without coercion. Privacy and confidentiality: personal information is safeguarded and identities are protected. Avoiding harm: anticipate and minimise physical, psychological or social discomfort. Below, the safeguard as a sequence: design the study, then ethical review by an independent committee, then collect data. No data are collected until the study has passed review. Informed consent participants understand what the research involves, and agree freely, without coercion Privacy and confidentiality personal information safeguarded; identities protected Avoiding harm anticipate and minimise physical, psychological or social discomfort The safeguard: no data are collected until the study has passed review Design the study Ethical review by an independent committee Collect data

Data analysis and interpretation

Two routes from raw data to an answer to the research question Two routes converge on one box, the answer to the research question. Qualitative analysis starts from texts such as transcripts and identifies themes and patterns of meaning, through thematic or content analysis. Quantitative analysis starts from numbers and summarises and models them with descriptive statistics, regression, t-tests and related techniques. A warning under the answer: it is only as trustworthy as the measures behind it. Running example, in two steps: first compare turnout across age bands, which describes the pattern; then use regression to ask whether political interest accounts for the gap, which explains it. “ QUAL Qualitative analysis themes and patterns of meaning: thematic or content analysis 739582 QUAN Quantitative analysis summarise and model the data: descriptive statistics, regression, t-tests An answer to the research question ! only as trustworthy as the measures behind it RUNNING EXAMPLE 1 Compare turnout across age bands — describes the pattern 2 Regression: does political interest account for the gap? — explains it

Validity and reliability

Validity and reliability pictured as an archer's target Two definitions at the top. Validity means accuracy: are we measuring what we intend to measure? Reliability means consistency: would the same procedure give the same result again? Below, three archery targets. First, reliable but not valid: six arrows tightly clustered, but in the lower left, off the bullseye; consistently wrong. Second, valid but not reliable: six arrows scattered around the bullseye, centred on it on average, but none of them on it. Third, reliable and valid: six arrows tightly clustered on the bullseye. Validity = accuracy are we measuring what we intend to measure? Reliability = consistency would the same procedure give the same result again? Reliable, not valid tightly clustered, off the bullseye: consistently wrong Valid, not reliable scattered around the bullseye, none of them on it Reliable and valid tightly clustered on the bullseye

Presenting research findings

The same finding presented two ways Two bar charts of the same illustrative data: turnout by age group, rising from 51 per cent among 18 to 24 year olds through 59, 66, 71 and 75 per cent to 78 per cent among those aged 65 and over. Left, hard to grasp: a jargon title (Fig. 3: self-reported electoral participation, per cent, by age cohort, weighted, n equals 1,500), heavy gridlines, bars in six unrelated colours, no labels on the bars and a separate legend to decode them, and an axis titled with a questionnaire code. Right, easy to grasp, the running example: a plain-language headline (under-25s are the least likely to vote), one colour with the youngest group highlighted, values written on the bars and age groups under them, and one sentence relating the chart back to the question: the age gap our question asks about is real; next, we ask whether political interest explains it. Three numbered callouts mark the three habits: 1, a plain-language headline; 2, a clean, directly labelled chart; 3, one sentence relating it back to the question. Hard to grasp Fig. 3: Self-reported electoral participation (%) by age cohort, weighted (n = 1,500) 0 10 20 30 40 50 60 70 80 90 100 % Q17 = 1 18–24 25–34 35–44 45–54 55–64 65+ Easy to grasp RUNNING EXAMPLE · ILLUSTRATIVE DATA Under-25s are the least likely to vote turnout at the last national election, by age (%) 51 18–24 59 25–34 66 35–44 71 45–54 75 55–64 78 65+ The age gap our question asks about is real; next, we ask whether political interest explains it. 1 2 3 1 plain-language headline 2 clean, directly labelled chart 3 one sentence relating it back to the question

Limitations and critiques

Acknowledging limitations, and how they move knowledge forward Three practices at the top: 1, acknowledge the weaknesses of the design, data and analysis, openly; 2, invite critique from peers, reviewers and the wider community; 3, point forward, since the limits show what future research should address. Below, how knowledge accumulates: a study leads to the next study, and that to the next, because each study's limits set the next study's questions. Running example: our survey's limitation is that self-reported turnout is over-stated, because voting is socially approved; the next study validates self-reports against official voting records. 1 Acknowledge the weaknesses of the design, data and analysis, openly 2 Invite critique from peers, reviewers and the wider community 3 Point forward the limits show what future research should address How knowledge accumulates: each study's limits become the next study's questions A study its limits set the questions The next study its limits set the questions … and the next RUNNING EXAMPLE Our survey's limitation: self-reported turnout is over-stated, because voting is socially approved The next study: validate self-reports against official voting records

Conclusion

Conclusion

  • Methodology matters: the quality of our conclusions depends on the quality of our methods
  • Social research is an evolving practice — new data sources, tools, and techniques continually reshape what is possible
  • The choices made at each stage — topic, design, sampling, collection, analysis, presentation — are connected and consequential
  • We followed one question — why young people vote less — from a vague idea to a defensible, if imperfect, answer; every study makes that same journey
  • Questions and discussion are welcome

Exercise

Today’s exercise: Research or not?

QR code linking to the exercise worksheet

bdstanley.netlify.app/social-research-methodology-1-exercise

Model answers: research or not?

Model answers to Task 1: research or not? Eight cards, one per scenario, each with a verdict and the criterion that decides it. Scenario 1, store visits: not research. Missing: an explicit procedure. Stores and staff picked haphazardly; "basically fine" is an impression, not evidence. Scenario 2, engagement survey: research. Same questionnaire, same month, all 450 staff, documented: explicit, transparent, repeatable. Scenario 3, ceo's memo: not research. Missing: evidence. Two articles and the CEO's authority settle the claim; no data were gathered. Scenario 4, warehouse log: research. A predefined protocol and a structured daily log: systematic observation in context. Scenario 5, checkout a/b test: research. Random assignment and recorded outcomes: an experiment anyone could repeat. Scenario 6, four-day-week poll: research, but bad. A procedure and real data, but a self-selected sample: 78% of readers who chose to click, not of employees. Scenario 7, turnover figures: not research. Missing: theory. Orderly, repeatable record-keeping, but no question it was designed to answer. Scenario 8, team interviews: research. Same open questions, recorded with consent, coded for themes: a small qualitative study. 1 Store visits NOT RESEARCH Missing: an explicit procedure. Stores and staff picked haphazardly; “basically fine” is an impression, not evidence. 2 Engagement survey RESEARCH Same questionnaire, same month, all 450 staff, documented: explicit, transparent, repeatable. 3 CEO's memo NOT RESEARCH Missing: evidence. Two articles and the CEO's authority settle the claim; no data were gathered. 4 Warehouse log RESEARCH A predefined protocol and a structured daily log: systematic observation in context. 5 Checkout A/B test RESEARCH Random assignment and recorded outcomes: an experiment anyone could repeat. 6 Four-day-week poll RESEARCH, BUT BAD A procedure and real data, but a self-selected sample: 78% of readers who chose to click, not of employees. 7 Turnover figures NOT RESEARCH Missing: theory. Orderly, repeatable record-keeping, but no question it was designed to answer. 8 Team interviews RESEARCH Same open questions, recorded with consent, coded for themes: a small qualitative study.

Model answers: what kind of research?

Model answers to Task 2: the research scenarios on the two axes The two-by-two grid from the lecture: columns qualitative and quantitative, rows experimental and descriptive. Quantitative and descriptive: scenario 2, the engagement survey (standardised items, compared year on year), and scenario 6, the four-day-week poll (a bad sample, but still this type). Quantitative and experimental: scenario 5, the checkout A/B test (random halves see version A or B, so conditions are deliberately varied). Qualitative and descriptive: scenario 8, the team interviews (open questions, coded for themes). Scenario 4, the warehouse log, sits across the boundary between qualitative and quantitative in the descriptive row: it is the odd one out, combining a predefined protocol whose entries can be counted with immersion in context aimed at meaning. The qualitative and experimental cell holds no scenario. Axis 1: what kind of data and reasoning? Qualitative deep, interpretative: meaning and experience Quantitative numerical, statistical: measurement and pattern Axis 2: does the researcher intervene? Experimental controls and varies conditions Descriptive observes only: no intervention 2 Engagement survey 2: standardised items, compared yearly 5 Checkout A/B test random halves see version A or B: conditions deliberately varied 6 Four-day-week poll 6: a bad sample, but still this type 8 Team interviews open questions, coded for themes 4 Warehouse log The odd one out: a predefined protocol (countable) and immersion in context (meaning) no scenario here

Model answers: record-keeping or research?

Model answer to Task 3: the line between record-keeping and research An arrow runs from record-keeping on the left to research on the right. Left of a dashed vertical line, scenario 7, the turnover figures: systematic, documented and repeatable, but compiled for a reporting template, not to answer a question. Right of the line, scenario 2, the engagement survey: designed to measure a concept, engagement, and track it, asking how it is changing and where. The dashed line is labelled: the line is a research question the data are designed to answer. Underneath: the turnover figures cross the line when the same figures are used to answer a question, such as why turnover is higher in some stores than others; the engagement survey sits close to the line, so it is defensible either way: if HR only files the results it is monitoring, and if it asks why scores fell it is research. Record-keeping Research 7 Turnover figures systematic, documented and repeatable — but compiled for a reporting template, not to answer a question 2 Engagement survey designed to measure a concept (engagement) and track it: how is it changing, and where? The line: a research question the data are designed to answer 7 crosses the line when the same figures are used to answer a question: why is turnover higher in some stores than others? 2 sits close to the line, so it is defensible either way: if HR only files the results, it is monitoring; if it asks why scores fell, research

Model answers: fixing scenario 6

Model answer to Task 3: why scenario 6 is still bad research, and how to fix it Left, why it is still bad research. The claim, "78% of employees want a four-day week", is based on a poll of the consultancy's own newsletter subscribers who chose to click through. The main problem is who answered: self-selected readers, not a sample of employees, so the 78% describes them, not the population the claim is about. Also: who asked, since the consultancy may gain from the result, and what was asked, since the wording is not published. Right, what it would take to fix it, in four steps: 1, define the population, the group the claim is about, for example all employees in Poland; 2, draw a probability sample, at random, from a list of that population (a sampling frame); 3, ask a neutral question and publish its exact wording; 4, report the method and its limits: sample size, response rate, and who commissioned it. Why it is still bad research “78% of employees want a four-day week” based on a poll of its own newsletter subscribers who chose to click through ✗ Who answered: self-selected readers, not a sample of employees. The 78% describes them, not the population the claim is about. Also: who asked (the consultancy may gain from the result) and what was asked (the wording is not published) What it would take to fix it 1 Define the population the group the claim is about, e.g. all employees in Poland 2 Draw a probability sample at random, from a list of that population (a sampling frame) 3 Ask a neutral question and publish its exact wording 4 Report the method and its limits sample size, response rate, and who commissioned it

Study guide

Full summary of this session, for revision: The nature and purpose of social research

QR code linking to the session handout

bdstanley.netlify.app/social-research-methodology-1-handout