Data Capture Templates

Prepare one sheet per interview or observation: open with consent, ask about real past events, write down what happened apart from what you think it means, and log the session the same day.

Mostly practice, not research: asking about past behavior is practitioner craft (Fitzpatrick, 2014), resting on a well-documented gap between what people intend and what they do (Sheeran & Webb, 2016). Watching work where it happens comes from contextual design (Beyer & Holtzblatt, 1998), also practitioner method. Survey rules have mixed support (Likert, 1932; Dillman et al., 2014). The quality scale in the log is a rule of thumb.

Use it when:

• You are about to run interviews, observations or a document review and want every session captured the same way.
• More than one person collects data and you need notes that can be compared.
• You have caught yourself remembering a conversation differently from how it went.

Skip it when:

• You have not yet decided who to study or how: plan it first in 2.5 Simple Study Design Sheet.
• You need structured feedback from a large group: use the survey layout under Optional below, after checking 2.4 Methods Menu.

Data quality is decided when you capture it, not when you analyze it, and most of what follows from that is practitioner method rather than tested science. The interviewing stance used here, asking about the last time something happened instead of what someone would do, is practitioner craft (Fitzpatrick, 2014). It rests on a gap that is well documented: what people intend and what they then do often differ, so stated intentions are weak evidence (Sheeran and Webb, 2016). Watching work where it happens comes from contextual design (Beyer and Holtzblatt, 1998), also practitioner method. For surveys the evidence is mixed: many rules about scales and question order are closer to convention than to findings (Likert, 1932; Dillman et al., 2014). The consent wording in the opening script applies an ethical standard you follow rather than weigh.

So the probes in Step 3 push for specific examples, and the opening in Step 2 is consent made routine, not small talk. Step 4 keeps what happened apart from your interpretation, so the raw record survives when your view changes. Step 5, the log, is the audit trail everything later rests on.

Give each session a code so notes, recordings and the log can be matched without using names.

Strong answer: "S07. 12 March, 25 min. Interview by phone. Sales lead. Participant C-F04 (former cafe customer, fourth contacted)."

Session code:

Date and length:

Method (interview, observation, document review):

Collected by (role):

Participant or source code (no names):

Illustrative example: the business, the roles and all numbers are made up.

The Gothenburg bakery from 2.1 interviews a former cafe customer by phone.

Step 1: Session S07, 12 March, 25 minutes, by the sales lead. Participant code C-F04.

Step 2: The sales lead explains that the bakery wants to understand why small cafes stop ordering, that it takes twenty minutes and that no name goes on the notes. The owner agrees; consent is noted.

Step 3: First question: "Tell me about the last week you ordered from us." Probe: "What happened to the bread you didn't sell?" A second question asks what the cafe orders now and from whom.

Step 4: Record: "10:05. 'On Mondays we threw away half a crate.'" Interpretation, in its own column: "Minimum order too large early in the week?" A later line: "Now buys from a supermarket, small amounts daily." Theme: waste and flexibility, not price. Surprise: price came up only once, when asked.

Step 5: Logged the same afternoon, quality 4. Follow-up: ask the driver how often bread came back from this cafe.