Understanding Research Designs Based on Data Collection: Survey, Case Study, and Content Analysis

When you set out to do research, one of your earliest and most important decisions is not what you want to study, but how you will gather the information to study it. Different questions demand different ways of collecting data. A scholar who wants to know what thousands of public library users feel about digital services needs a very different toolkit from one who wants to understand the inner workings of a single university library, or one who wants to trace how the idea of “information literacy” has been written about in journals over the last fifty years. These three needs map neatly onto three classic research designs based on data collection methods: the survey approach, the case study approach, and content analysis. Understanding when and how to use each one is a core skill for anyone working in research.

Why data collection methods shape research design 🔗

A research design is essentially a blueprint. It decides how you will collect, measure, and analyse data so that your conclusions are credible. When we classify designs by data collection method, we are asking a practical question: where does the evidence come from, and how do we capture it? The answer changes everything downstream, from sample size to the kind of analysis you can perform.

Broadly, the survey approach gathers a little information from many people. The case study approach gathers a great deal of information from a few cases. Content analysis steps away from people altogether and gathers data from documents and recorded communication. Each design trades breadth against depth in its own way, and each suits a particular type of research question. The skill lies in matching the method to the question rather than forcing a question into a familiar method.

The survey approach 🔗

The survey approach is built for scale. Survey methodology studies how to sample individual units from a population and then collect data from them using tools such as questionnaires. Because surveying an entire population is rarely possible, the method depends on choosing a smaller group that fairly represents the whole. This is why sampling sits at the heart of every good survey.

Sampling: the foundation of a survey 🔗

A sample is a selection of individuals drawn from a larger population so that you can study the sample instead of everyone. The success of a survey depends heavily on how representative this sample is. The larger and more representative the sample, the more valid the conclusions, and researchers must guard against problems like self-selection bias, non-response bias, and undercoverage bias. Sampling methods fall into two families: probability sampling, where every unit has a known chance of selection and results can be generalised, and non-probability sampling, which is quicker but harder to generalise from.

For a topic like reading habits among college students across a state, a researcher might use random or stratified sampling to make sure different types of institutions and regions are fairly included. Getting the sample wrong undermines everything that follows, no matter how well-designed the questionnaire is.

Questionnaires and interviews 🔗

Once the sample is set, data is collected mainly through two instruments. Depending on how the data is collected, survey research divides into questionnaire surveys and interview surveys. Questionnaires are self-administered and written, distributed by post, online, or in person, and respondents fill them in themselves. Interviews are more personal, with a trained interviewer asking a standardised set of questions by phone or face to face and recording the answers.

The two methods differ in important ways. An interview script may carry special instructions for the interviewer that respondents never see, and it can include space for the interviewer to note personal observations. Questionnaires are cheaper and faster for large groups but offer less flexibility. Within both, questions can be closed-ended, offering fixed options that are easy to analyse, or open-ended, which yield richer detail but take far longer to interpret.

Designing and testing the instrument 🔗

A questionnaire should never go straight into the field. Pretesting the questionnaire before fielding it is essential, often through cognitive interviews with people similar to the eventual respondents, followed by a pilot test to check that recruitment, administration, and data cleaning all work as intended. Pretesting exposes ambiguous wording, unclear instructions, and hidden bias that would otherwise corrupt the data.

Strengths and limits of surveys 🔗

Surveys have real advantages. Large samples can detect small effects and allow comparison between subgroups, and survey research is economical in terms of a researcher’s time, effort, and cost compared with experiments. But it carries unique weaknesses too. It is vulnerable to non-response bias, sampling bias, social desirability bias, and recall bias. A survey can tell you what a large group reports, but it rarely captures the depth and context behind those reports.

The case study approach 🔗

Where the survey goes wide, the case study goes deep. It is a research design built to study one case, or a small set of cases, in exhaustive detail. A case study is an in-depth design that primarily uses qualitative methodology to investigate an individual, group, organisation, or event, and it is especially suited to answering “how” and “why” questions.

What counts as a “case” 🔗

A case can be almost any bounded social unit. Case study research has no single dominant design, and one common distinction, following the work of Robert K. Yin, separates single-case studies from multiple-case studies. The “case” might be a person, a family, an institution, a cultural group, a programme, or even a whole community. The defining feature is that the unit is studied as an integrated whole within its real-world context, not isolated in a laboratory.

Depth over breadth 🔗

The case study is, at its core, a study in depth rather than breadth. It places emphasis on the full analysis of a limited number of events or conditions and the relationships between them. Instead of collecting a little from many, the researcher collects a great deal about a few, examining the internal features of each case alongside its surrounding situation. The study often extends over a long period to capture the natural history of the unit and gather enough information to draw correct inferences.

For a Library and Information Science student, a case study might examine how one rural community library transformed its services after introducing digital access, tracing the decisions, struggles, and outcomes in fine detail. This kind of insight simply cannot emerge from a tick-box questionnaire sent to hundreds of libraries.

Ensuring rigour through triangulation 🔗

Because case studies rely on the researcher’s interpretation, they must work hard to stay credible. A key tool here is triangulation, the use of multiple sources of evidence to confirm findings. Triangulation helps ensure trustworthiness and lets the researcher accurately reflect what is being investigated. Strong case study work also draws on multiple data sources, reflexivity, and systematic analysis to build confidence in its conclusions.

Strengths and limits of case studies 🔗

The case study’s great strength is the richness and contextual depth it produces. Its weaknesses are equally well known. It is resource-intensive and time-consuming, it can be vulnerable to the researcher’s subjective bias, and its findings are difficult to generalise to a wider population. A single library’s experience may not predict what happens elsewhere. The case study buys depth at the price of breadth, which is exactly the opposite trade-off from the survey.

Content analysis 🔗

The third design steps away from living participants entirely. Content analysis is a method for studying documents and recorded communication. Content analysis determines the presence of certain words, themes, or concepts within qualitative data such as text, and lets researchers quantify and analyse their meanings and relationships. From these, they can make inferences about the messages in the text, the people who produced it, the audience, and even the culture and time around it.

Working with documents, especially historical ones 🔗

The raw material for content analysis is existing content rather than freshly collected responses. Document analysis is especially useful for studying historical records, policies, and communications without needing direct interaction with participants. This makes it invaluable for historical studies, where the people involved may be long gone but their letters, reports, newspapers, and official records survive.

The range of material is enormous. Archival and historical works, media coverage, blogs, films, advertisements, and photographs can all be the content analysed. For an LIS researcher, this could mean studying how library annual reports framed their mission across decades, or how newspapers covered the role of public libraries during a particular era. A major practical advantage is accessibility, because documents often exist before the study even begins.

Quantitative and qualitative content analysis 🔗

Content analysis is flexible enough to work in two modes. Content analysis can be both quantitative, focused on counting and measuring, and qualitative, focused on interpreting and understanding. In both, the researcher categorises or “codes” words, themes, and concepts and then analyses the results. The quantitative side answers questions of “how much” and “how often” by systematically measuring frequency. The qualitative side answers “why” and “how” through close, interpretive reading of the text.

Reliability and the coding process 🔗

The credibility of content analysis rests on consistent coding. The classification procedure must be reliable, meaning different people should code the same text in the same way. This is why researchers pay close attention to inter-coder and intra-coder reliability, often training multiple coders and comparing their results. Software tools such as NVivo or ATLAS.ti help manage large volumes of text, apply codes, and document the analysis. A notable strength of the method is that it is unobtrusive: studying a document does not change it, unlike a survey or interview that might influence the very people being studied.

Choosing the right design 🔗

The three designs answer fundamentally different questions. Choose a survey when you need to measure attitudes, behaviours, or characteristics across a large population and want results you can generalise. Choose a case study when you need to understand a complex phenomenon in depth, in its real-world setting, and you care more about “how” and “why” than about “how many”. Choose content analysis when your data already exists as text or recorded communication, particularly for historical or comparative work where direct contact with participants is impossible or unnecessary.

In practice, strong research often combines them. A study of a library system might use a survey to map broad user satisfaction, a case study to explore one branch in depth, and content analysis of policy documents to understand the institutional backdrop. Knowing the distinct logic of each design is what lets a researcher mix them wisely rather than at random.

What do you think? If you were studying how reading habits have changed among students over the past twenty years, which of these three designs would you reach for first, and what would you lose by ignoring the other two? Could a single research question ever be fully answered by just one of these methods alone?

References

1. https://en.wikipedia.org/wiki/Survey_methodology
2. https://www.scribbr.com/methodology/survey-research/
3. https://courses.lumenlearning.com/suny-hccc-research-methods/chapter/chapter-9-survey-research/
4. https://aapor.org/standards-and-ethics/best-practices/
5. https://resources.nu.edu/researchtools/casestudy
6. https://en.wikipedia.org/wiki/Case_study
7. https://www.publichealth.columbia.edu/research/population-health-methods/content-analysis
8. https://lumivero.com/resources/blog/the-basics-of-document-analysis/
9. https://open.oregonstate.education/qualresearchmethods/chapter/chapter-17-content-analysis/
10. https://www.scribbr.com/methodology/content-analysis/