A neighbour wants to visit the library on Saturday. She finds the council homepage on her phone, follows a link to events and ends up looking at last month’s reading group. You suspect the opening hours are hard to find. One afternoon gives you a reason to investigate. What would it take to find out where other people get stuck?
To see how to start a research project, follow that observation through to a proposal for a tutor or library manager. The library, participants and illustrative results are hypothetical. You can borrow the decisions for a different subject, whether you are investigating how people read a bus timetable or how students find places to study.
1. Give the question an ending
“How good is the library website?” could occupy a team for months. Try a question with a visible endpoint. “Where do adults unfamiliar with this council website encounter difficulty when finding one branch’s Saturday opening time on a smartphone?” Name the branch, the starting page and a particular Saturday in your protocol. Check the correct answer with the library, including any temporary change for that date.
Write down what would count as evidence. A person opening an events calendar while looking for hours gives you a route to examine. A participant arriving directly at the correct page is equally informative. Your question leaves room for the possibility that everybody succeeds. It also rules out claims about desktop access, other branches or whether a new design would work better.
2. Find what others have already tried
Split the question into concepts. In a library database try “public library” alongside “usability”, then try “library website” with “navigation”. Look for studies of people attempting realistic tasks. Add “mobile” if results mainly concern desktop sites. If results dry up, drop that term and check the methods sections for device information. The University of Leeds search guide explains how to vary terms and adapt searches across databases.
Keep a small search log with the database, date, terms and reasons for keeping a paper. For each useful study, note who took part, what they tried and what the researchers measured. An academic library study using experienced students might offer a good observation sheet, yet tell you little about adults using a council website for the first time. Preserve that distinction in your notes and save the citation while the paper is open.
Finish the reading with a sentence that changes your design. Perhaps a study recruited only regular borrowers. You can explain why a notice outside the library might reach people it missed. An earlier project earns its place in your proposal when it helps you make a decision or recognise an assumption.
3. Choose a method that shows the difficulty
A questionnaire could establish what individual volunteers remember. Here you want to see the moment a link takes someone away from their goal. Sit beside a volunteer or use an agreed screen sharing arrangement and observe one attempt. Ask afterwards what they expected a particular link to contain. The Government Service Manual describes how to design tasks that give participants a goal without giving away the route.
Prepare one instruction. “Imagine you want to visit this branch on Saturday morning. Starting here, find out when it opens.” Show everyone the branch name and date. Decide how long to allow, what counts as a correct answer and when help may be offered. Silence while someone reads through a long menu can feel long enough to tempt you into pointing.
4. Decide whose experience you can study
Plan six short sessions with adult smartphone users who have never visited this council website. Invite volunteers through a community noticeboard and a library notice, with permission from the organisers. Ask about prior use before booking. Offer different session times and let people use familiar accessibility settings. Record the recruitment routes so a reader knows whose experiences might be missing.
Six volunteers can show you six routes through a task. People answering a public invitation may be more confident, have more spare time or be more interested in libraries than their neighbours. Explain that limitation when proposing the sample. A claim about what proportion of residents can locate opening times would require a different recruitment and measurement plan.
5. Prepare for consent and a rehearsal
Check your institution’s review process before recruiting anyone. Give participants a short explanation of the purpose, what you will record, who can read the notes and when you will delete them. Explain how they can stop and the deadline for removing their contributions. The Economic and Social Research Council’s guidance on consent asks researchers to consider participants’ choices throughout the study.
You can record link choices and outcomes without filming screens or collecting browsing histories. Keep contact details separate from notes labelled with participant codes. Use approved storage and restrict access. Rehearse with a consenting volunteer after any required approval, then revise unclear instructions. Keep the rehearsal separate from the six planned observations, particularly if you change the wording or time limit.
6. Make the record intelligible tomorrow
Use one row per attempt. Include a participant code, session date, starting page, sequence of pages, answer given, help offered and outcome. Define “unassisted success” as giving the verified opening time before the limit without a hint. Define separate labels for assisted answers and unfinished attempts. The UK Data Service guide to documenting data explains how definitions and missing values help another person interpret a file.
Separate what you saw from your interpretation. “Opened Events twice” describes an action. “Expected branch information under Events” needs supporting evidence, perhaps an answer to your follow-up question. Date your observation sheet and keep a clean copy of the original notes before coding them. If the website changes during collection, record the change and consider whether the sessions are still comparable.
7. Analyse the attempts and write the proposal
Imagine four volunteers find the opening time unaided, one succeeds after a hint and one stops. Report those counts together. Trace the routes individually, including successful detours. If three people visit Events, check whether that choice actually prevented progress. One might have found a branch link there. Use their explanations to identify an improvement worth testing and preserve examples that challenge your first impression.
Decide how to handle an interrupted session. If a connection fails before the first page loads, record that separately and explain it when reporting the denominator. Treat a request to withdraw according to the agreement you made at consent.
Write a one-page proposal before the real study begins. Give the question and its practical purpose, summarise relevant reading, then specify recruitment, consent, the task, the outcome rules and your intended analysis. Budget for travel, accessible materials and participant expenses where applicable. Allow time for approval, a rehearsal, six sessions and writing. End with the decision the library could take after reading your findings.
For example, you might propose testing a clearer branch link on the events page if observation supports it. Give the library a short findings sheet showing the routes, where each attempt ended and which design change deserves a second trial. Someone else can then repeat the task after an update and see what changed.