02 Section A, Key question 2 - Designing sociological research (part 1)

Choosing a research topic

The choice of a research topic depends on what sociologists want to understand and what they can responsibly investigate.

Social inequality means unequal access to resources, opportunities or rewards between people or groups. A sociologist might notice that residents of one area seem to have fewer opportunities to find paid work. That concern is too broad to investigate all at once: it might involve qualifications, discrimination, caring responsibilities or transport. Choosing a topic means selecting a manageable focus within the concern.

Throughout this lesson, Borough L is an invented teaching scenario, with no reported research findings. Its researchers choose transport barriers and access to paid work among adult residents who are seeking jobs. They want to investigate barriers rather than assume that people without jobs lack motivation.

Interests, values and resources

Theoretical interests direct attention towards different explanations. An approach concerned with material inequalities might ask whether low income restricts travel to work. An interpretivist approach, which seeks to understand people's meanings and experiences, might ask how job seekers judge whether a vacancy is realistically within reach. These interests help shape a question; they do not establish which explanation is correct.

Researchers' values are their beliefs about what matters. Concern for fairness might motivate a study of excluded residents, while personal experience of unemployment might make the topic feel urgent. That motivation can reveal neglected issues. It can also create a temptation to notice only evidence supporting the researcher's beliefs. Researchers therefore need to make their assumptions visible and remain willing to find that their initial explanation was incomplete.

Funding can make some topics possible and leave others unsupported. If a council funds a study of bus access, researchers may be able to investigate transport in depth but lack resources for a study of workplace discrimination. Funding can influence the agenda without making every funded finding biased. A sponsor's demand that the study must conclude its services are successful would threaten the enquiry's independence. The British Sociological Association advises against accepting contracts dependent on a particular research outcome. BSA Statement of Ethical Practice, 2017, paragraphs 42 and 48.

Time, skills and access also narrow the design. A short project may examine current experiences in one borough rather than track employment across the UK for several years. Researchers need skills suited to their planned collection and analysis, such as interpreting survey responses or conducting sensitive interviews. A local organisation may help introduce researchers to job seekers; permission to approach an organisation does not guarantee that residents will agree to take part. An online-only study may fail to reach residents who have limited internet access.

Ethics concerns researchers' responsibilities towards participants and others affected by the study. Asking about financial hardship can expose sensitive information or cause embarrassment. Researchers must consider freely given informed consent, privacy and avoidable harm while designing the enquiry. In Borough L, a person should be able to decline an interview without fearing that their access to employment support will be affected. These responsibilities may change how a topic is investigated or make a proposed approach unacceptable. BSA Statement of Ethical Practice, 2017, paragraphs 12 and 18–24.

Choosing a manageable topic therefore involves a reasoned compromise: the study should matter sociologically, generate relevant evidence and be feasible without sacrificing participants' interests.

Aims, hypotheses and research questions

An aim states the overall purpose of a study, while a hypothesis or research question makes that purpose more precise.

For Borough L, the aim is to investigate how transport barriers shape adult job seekers' opportunities to apply for paid work. This identifies the issue and population but still leaves decisions about what evidence to collect.

A research question asks what the study seeks to answer. A narrower question could be: “What transport difficulties do adult job seekers in Borough L report when deciding which vacancies to apply for?” This guides the collection of evidence about people's decisions and the obstacles they encounter. It is different from an individual interview question: the research question guides the whole enquiry, while several interview questions may contribute to answering it.

A hypothesis is a proposition that can be tested against evidence, often predicting a relationship between variables, characteristics that can vary. For example: “Adult job seekers in Borough L who cannot afford travel fares submit fewer job applications in a four-week period than those who can afford the fares.” Affordability and number of applications are the variables. Evidence could support this prediction, fail to support it or suggest that the relationship is more complicated. A hypothesis is neither an already established fact nor a promise to produce a particular result.

To test that hypothesis, the researchers would need evidence from both groups using comparable measures. Evidence of a difference alone would not establish that fares caused it: qualifications, caring commitments or available vacancies could also affect applications. Even an apparently clear hypothesis requires further design decisions.

Not every study needs a hypothesis at the outset. An interpretivist enquiry might begin with the open question: “How do job seekers describe deciding whether a vacancy is reachable?” It seeks to discover meanings rather than test a fixed prediction. Interviews might reveal an unanticipated concern, such as fear of being stranded after a late shift, and lead researchers to refine their questions. Such research can still be systematic: it needs a clear purpose, justified evidence collection and careful interpretation.

The following situation is another original example, not a real study.

Primary and secondary data

Primary and secondary describe the origin and use of data, while quantitative and qualitative describe their form.

Primary data are collected specifically for the current research project, whether by its researchers or by people collecting on their behalf. New survey responses and interviews commissioned for the Borough L study would be primary data. Secondary data already exist and are reused for the present enquiry, having originally been produced for another study or purpose. Existing transport records or archived interviews could be secondary data for the Borough L researchers.

Quantitative data express quantities, such as counts, proportions or numbers of applications. Qualitative data convey meanings or descriptions, such as interview accounts of why a journey feels unsafe. Neither form tells us whether the data are primary or secondary. The UK Data Service explicitly supports the reuse of both quantitative and qualitative research data. UK Data Service, New to using data.

These four original examples show the two distinctions working independently:

The same material can have a different status in different enquiries. The archived interviews were primary data for the project that originally collected them; they are secondary data when Borough L's researchers reuse them. A survey can also collect both numbers and written explanations, so “survey” is not automatically a synonym for quantitative data.

The choice should follow the question. Primary collection allows researchers to ask directly about the journeys and decisions they need to understand, but takes time and resources. Secondary material may allow a wider comparison without new fieldwork. However, records counting daytime services might not capture access to early or late shifts. Researchers must inspect how existing material was produced, what it covers and whether reuse is permitted. Secondary data are not automatically poor quality, and primary data are not automatically well designed.