Simple Study Design Sheet

Turn the methods you chose into a plan someone else could follow: who is in the sample and why, what you collect and when, how you will analyze it, and what could go wrong.

Settled textbook method rather than tested intervention: sampling strategies come from survey statistics and qualitative practice (Cochran, 1977; Patton, 2014), and the qualitative quality checks are established criteria (Lincoln & Guba, 1985). Survey design rules have mixed support (Dillman et al., 2014). Experiment guardrails come from the field's methodologists (Kohavi et al., 2020). Timings and the contingency share are rules of thumb.

Use it when:

• You have chosen your methods (see 2.4 Methods Menu) and need to plan the fieldwork.
• More than one person will collect or analyze data, and they need the same plan.
• The results will feed a decision someone may later challenge, so the sample has to be defensible.

Skip it when:

• You are running one quick conversation to get a feel for a topic: use 2.7 Data Capture Templates directly.
• You have not yet chosen methods: go back to 2.4 Methods Menu.

A study is a chain of choices, and the weakest link sets the limit on what the results can claim. Most of the vocabulary on this sheet is settled textbook material. The sampling strategies come from classical survey statistics (Cochran, 1977) and qualitative practice (Patton, 2014), and snowball sampling was formalized as a way to reach people who are hard to find (Goodman, 1961). Each label is a claim about how far results travel: a purposive sample can give insight but not prevalence, and calling a convenience sample purposive is the most common quiet mistake in applied work. Questionnaire craft has mixed support, and many of its rules are convention (Likert, 1932; Dillman et al., 2014). The qualitative checks, such as asking participants to review findings or hunting for cases that contradict them, are accepted criteria rather than tested methods (Lincoln and Guba, 1985). The experiment guardrails come from the people who run tests at scale (Kohavi et al., 2020).

So Step 2 asks for the sampling method you will actually use and the criteria before recruitment. Step 4 fixes the analysis before collection, because it changes what you need to collect.

Summarize the study in four lines someone outside the work would understand.

Strong answer: "Why small cafes stop ordering. Lead: sales lead. 3 March to 11 April. Interviews with former and current cafe customers, then two years of order data."

Study question (from 2.1):

Lead (role):

Start and end dates:

Methods (from 2.4):

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

The Gothenburg bakery from 2.1 plans its study of small cafes that stop ordering.

Step 1: Lead: the sales lead. Six weeks, ending before the April board meeting. Interviews first, then order data.

Step 2: The population is about forty small cafes. The sample is eight that stopped and eight that still order, drawn at random within each group (stratified). Cafes owned by staff relatives are excluded. The sample can show the reasons cafes give, not how common each one is.

Step 3: Interviews by phone or visit in weeks 1 to 3, recruited through the drivers; order data pulled from the till system in week 4.

Step 4: Codes for reasons to stop are agreed after four interviews. The order data compares cafes that ordered at the minimum with the rest. Checks ticked: triangulation, negative cases, pilot test and consistent protocol. Member checking was left out for lack of time, which the report will say.

Step 5: Main risk: former customers do not reply. Fallback: driver interviews. The owner and head baker get a two-page brief at the board meeting.