Find the work question behind a low survey score.

Worked example: a survey score drops in two sites

In this fictional example, new hires at two of six distribution sites score “I can get help when I need it” lower than the previous intake. The evening shift is underrepresented in the survey. Invite a small mix of recent hires and supervisors on the affected shifts and include one site with a stable score to see how its process works. Ask everyone, “Think about the last time you needed help on shift. What happened from the moment you asked?”

One new hire says a supervisor was easy to find until 7 p.m. Another describes a coworker solving the same problem on a later shift. Check the rota, the handoff to the evening lead and who could reach the support channel. Test a named evening contact for two weeks. Record the owner and ask the next group whether they could find help. Keep the original score and the follow-up accounts visible as two kinds of evidence.

Check the item before explaining the result

Record the exact wording, response scale, field dates, eligible population and response base. Check whether the question, collection method or group changed from the comparison period. A resource score might cover equipment, staffing, information or time. Record these meanings as open questions before discussing why the measure moved.

Choose a work situation the interviews can explore

Narrow the follow-up to a decision someone can make. For example, a site lead may need to decide which part of the first-week access process to investigate. Write a neutral question about a recent occasion when a new starter needed access to perform a task. Keep room for examples where access was straightforward. Ask what happened before proposing the intervention the team already has in mind.

Recruit across the relevant work conditions

Identify roles, shifts, sites or tenure groups whose experiences may differ. Explain how the follow-up invitation relates to the survey and what will be shared from the interview. Use the participation route your organization has agreed. Where the survey was anonymous, invite people to a separate voluntary discussion without trying to connect their identity to an earlier score. Record the groups the follow-up did and did not reach.

Start by checking what the question means

Ask the participant how they understand the survey item in their work. They can discuss the topic without revealing how they answered the survey. AAPOR recommends clear, specific wording and considering whether other kinds of information are needed to answer the research question. Treat an unexpected interpretation as useful evidence about the item and the context, rather than correcting the participant toward the sponsor’s preferred meaning.

Follow a recent example in sequence

Invite the person to describe the last occasion when they needed the resource, information or support in question. Ask what they were trying to do, what was available, where they paused and what they tried next. Use neutral follow-ups to clarify a vague account. Ask which record or approved process could help someone check the sequence. The worksheet provides a compact guide for the conversation.

Ask for a contrasting case

Explore a time when the same task went more smoothly or a condition under which the issue did not occur. Ask what differed in the task, source, timing or support available. The contrasting account may reveal that the difficulty is confined to one step or that two different issues sit behind the same item. Keep those possibilities separate while the process owner checks the relevant evidence.

Worked example of a focused follow-up

In a fictional customer-service team, an item about having the information needed for work receives a low score. The interview guide explores recent complex customer cases. Some participants describe an inaccessible reference folder. Others can open it but find two conflicting instructions. One describes finding the correct instruction through a maintained team index. The findings suggest separate questions about permission, document ownership and retrieval. They do not establish that any one issue explains every survey response.

Combine the evidence without changing its meaning

Keep the survey result in its original definition. Summarize the interview accounts with the groups reached, contrasting examples and unanswered questions. Separate a reported difficulty from a checked process condition. The interview participants may be a small, self-selected group, so describe their accounts as such. A broad conclusion about the whole population needs evidence that matches that broader claim.

Agree a check or a bounded change

In the fictional example, the access owner can examine the folder permissions and the content owner can resolve the conflicting instructions. Choose a step within an owner’s control and decide what evidence will show whether it was implemented and usable. Give participants a date for an update. Explain what was heard, what will be checked and what remains open so the follow-up produces a visible next action.

Use the next measurement deliberately

If you repeat the survey item to examine change, preserve its definition and document any change in the participating population or collection method. Look at implementation evidence and new accounts alongside the score. Several conditions may have changed between waves. A higher score after an intervention is a result to examine in context. Keep the action and the interpretation tied to what the evidence can actually support.