DATA 5013 Module 2 Research Design and Sampling Plan Example

Fifteen Strata, 3,000 Invitations and 600 Answers: A Research Design and Sampling Plan for a Bookstore Membership Study

Student Name

American College of Education

DATA5013: Market Research

Module 2 Assignment

Instructor Name

May 8, 2028

Introduction

Module 1 set five research objectives for Northstar Books, the composite five-store chain in Minneapolis and St. Paul where I manage marketing, as it considers a $45 paid membership: estimate demand at three prices, rank five benefits, identify segments, estimate added spending and test effects on the free loyalty card. This paper chooses the research design to meet those objectives. It explains the mix of exploratory and descriptive research, defines the populations and sampling frames, calculates sample sizes, describes the sampling method, plans for nonresponse and sets incentives, a timeline and a budget.

Choosing the Design

Marketing research distinguishes exploratory designs, which generate insight and hypotheses, from conclusive designs, which measure and test (Malhotra, 2019). This study needs both. Four focus groups, held first, will explore why customers value certain benefits, how they would react to a paid tier and what language they use, insight that will sharpen the survey questions in Module 3. A cross-sectional descriptive survey of loyalty card members will then put numbers on how many would join, which perks matter and how buying habits might shift. A short in-store intercept survey of customers without a card will extend the estimates beyond existing members. Purchase data from the point-of-sale system will supply segmenting variables without asking respondents to recall spending.

Matching Methods to Objectives

Each objective has a primary method. Demand at three prices will be measured in the member survey by randomly assigning respondents to see one of the three prices, so that each price is judged on its own rather than in comparison with the others, which would push answers toward the middle. Benefit importance will be measured with a ranking exercise in the survey and explored in the focus groups. Segments will be built from purchase data, then compared on survey answers. Added spending will be estimated from survey questions on expected purchases and checked against the actual behavior of members of the free card who received occasional discount coupons last year. Effects on the free card will be explored in the focus groups and measured with attitude items in the survey.

Populations and Frames

The main population is active loyalty card members, defined as the 41,000 cardholders with at least one purchase in the past twelve months. The sampling frame is the loyalty database, limited to the roughly 29,000 active members who have agreed to receive email from the chain, a limitation discussed below. The second population is customers who shop at the stores without a card. There is no list of these customers, so they will be sampled at checkout. Focus group participants will be recruited from both groups, with at least one group drawn from frequent buyers and one from occasional buyers.

Sample Size

For the member survey, the key estimate is the share likely to join at each price. For a proportion near 50 percent, a simple random sample of about 385 gives a margin of error of about five percentage points at 95 percent confidence, and about 600 gives about four points. The plan targets 600 completed responses, enough for reasonably precise chainwide estimates and for comparisons between frequent and occasional buyers, though store-level estimates will be less precise. Past email surveys by the chain had response rates of about 20 percent, so 3,000 members will be invited. The intercept survey targets 300 completed responses, 60 per store. These calculations assume simple random sampling; stratification usually improves precision slightly, so the estimates are conservative.

Sampling Method

Members will be selected by stratified random sampling. The frame will be divided into fifteen strata, by home store, five, and purchase frequency over the past year, three groups: one to three purchases, four to nine and ten or more. Invitations will be allocated in proportion to each stratum's size, then drawn at random within each stratum, ensuring that every store and frequency group is represented in its true share. Because frequent buyers are fewer but central to the decision, the analysis will report their results separately as well. For the intercept survey, interviewers will approach every fifth customer leaving the register during set shifts across weekdays, evenings and weekends, a systematic approach that limits interviewer choice.

Nonresponse

Nonresponse is a concern, but its effect is often misunderstood. Groves and Peytcheva (2008), in a meta-analysis of studies that could measure nonresponse bias directly, found that response rates were only weakly related to the size of the bias; the real danger is that respondents and nonrespondents may hold different views on the very questions being asked. The plan therefore checks for bias rather than only chasing a high rate. Because purchase histories exist for every invited member, respondents and nonrespondents can be compared on spending, frequency and store. If respondents prove to be heavier buyers, results will be weighted by stratum to match the population.

Coverage and Mode

Two further sources of error need attention. Coverage error arises because members without email permission, about 29 percent of the active base, cannot be invited; they may be older or less engaged. A small random sample of them, 200 members, will receive a mailed postcard with a link and a short paper option, to test whether their answers differ. Mode effects arise because people may answer differently online and face to face. Dillman et al. (2014) recommend designing questions to work similarly across modes, so the intercept questions will be a shortened version of the online survey with identical wording for the key likelihood and price items.

Analysis Plan

The analysis will be planned before data arrive. For each price, the share very or extremely likely to join will be reported with a confidence interval, weighted by stratum. Because people overstate purchase intentions, a common adjustment will also be shown, counting only a fraction of those who say they are very likely, so owners see a cautious estimate beside the raw one. Benefit rankings will be summarized as average ranks and the share ranking each benefit first. Differences between frequent and occasional buyers will be tested with chi-square tests for proportions and t-tests for averages. Focus group transcripts will be coded for themes, as Module 4 describes, and the themes will be used to explain the survey numbers rather than stand alone.

Incentives, Timeline and Budget

Each completed member survey earns a $5 store credit, and focus group participants receive a $75 gift card and a signed book. Fieldwork runs for six weeks: focus groups in the first week, questionnaire revision in the second, the email survey in weeks three to five with two reminders, the intercept survey in weeks four and five and the postcard follow-up in week five. The $14,000 budget covers about $3,000 in survey credits, $3,000 for focus group incentives and facilitation, $1,500 for intercept staffing, $800 for printing and postage and the rest for analysis time and contingencies.

Conclusion

The design combines exploratory focus groups with a descriptive member survey and an intercept survey, supported by purchase data. A stratified random sample of 3,000 members across fifteen strata targets 600 responses and a margin of error of about four points, with checks for nonresponse, coverage and mode effects. Module 3 will build the questionnaire.

References

Dillman, D. A., Smyth, J. D., & Christian, L. M. (2014). Internet, phone, mail, and mixed-mode surveys: The tailored design method (4th ed.). Wiley.

Groves, R. M., & Peytcheva, E. (2008). The impact of nonresponse rates on nonresponse bias: A meta-analysis. Public Opinion Quarterly, 72(2), 167-189. https://doi.org/10.1093/poq/nfn011

Malhotra, N. K. (2019). Marketing research: An applied orientation (7th ed.). Pearson.