MKR 7000 Module 3 Questionnaire, Scales and Sample Plan Example

Measuring Intent Without Inflating It: The Meal-Subscription Survey

Student Name

Master of Business Administration, Thomas Edison State University

MKR-7000: Marketing Research

Mentor Name

Month Day, Year

Measuring Intent Without Inflating It: The Meal-Subscription Survey

Purpose

This document specifies the measures and sample for the descriptive survey in Hearthstone Market's meal-subscription study. Babin (2019) treats measurement and sampling as the stages where research most often goes wrong: concepts measured imprecisely, or samples that do not represent the population, produce confident answers that are wrong. The goal is a survey whose numbers can be trusted when the launch decision is made.

Constructs and Measures

Each research question requires a defined construct and a measure. Churchill (1979) recommended a sequence for developing marketing measures: specify the domain of the construct, generate items, collect pilot data, purify the measure and assess reliability and validity. The survey applies that logic to five constructs.

Two choices guard against inflated intent. The plan is described concretely, with price, number of meals and pickup or delivery terms, because vague descriptions overstate interest. And price is randomly assigned rather than asked as a ladder, so each respondent reacts to one price, which avoids anchoring and gives clean comparisons across three groups.

Reliability and Validity

Reliability is consistency; validity is whether a measure captures what it claims to. In a pilot with 40 loyalty members, the four convenience items produced a Cronbach's alpha of 0.84, above the common threshold of 0.70, so they can be averaged into one score. Content validity rests on the focus groups, which generated the items in customers' own terms. Predictive validity will be assessed after the field test by checking whether survey intent in test stores' trade areas predicted actual sign-ups. Current behavior is measured both by self-report and by loyalty data, which allows a check on accuracy.

Questionnaire Design

The questionnaire runs about eight minutes on a phone and moves from general to specific: current weeknight dinner habits, prepared-food purchases, the convenience items, a description of the subscription, intent at the assigned price, features, expected duration and finally demographics. Podsakoff et al. (2003) describe common method bias, the inflation of relationships when the same respondent answers related questions in the same format at the same time. The design reduces it by separating the convenience items from the intent question with unrelated questions, using different response formats, and validating behavior against loyalty records rather than relying only on self-report.

Sampling Design

The population is Hearthstone's loyalty households with an email address, about 96,000. The sampling frame is the loyalty database, which excludes non-members, a limitation that matters little because the subscription would be marketed mainly to members. A stratified random sample by store ensures that every store's trade area is represented in proportion to its members, which allows estimates for store groups as well as the chain.

Sample Size

For a proportion, the sample needed for a given margin of error at 95 percent confidence is 1.96 squared times p times (1 minus p), divided by the margin squared. Using p = 0.5, the most conservative assumption, and a margin of 0.04, the result is about 600 completed surveys. Splitting the 600 completes across three prices leaves roughly 200 per price, which gives an 80 percent chance of spotting a gap of around 13 points in stated intent between any two of them. Past loyalty surveys have drawn about a 12 percent response rate, so about 5,000 invitations are needed.

Nonresponse

A 12 percent response rate means most invited households will not answer, and respondents may differ from nonrespondents. The analysis will compare respondents with the full sample on loyalty-data variables such as spending, visit frequency and store, and weight the results if differences appear. A reminder after four days and a small incentive, entry into a drawing for gift cards, should raise the response rate.

Pretesting the Questionnaire

Before fielding, the questionnaire will be pretested in two ways. Five cognitive interviews, in which loyalty members think aloud while answering, will reveal questions that are misunderstood, such as whether a week means five dinners or seven. A soft launch to 200 invitees will then check completion time, drop-off points and the distribution of answers, so that problems such as a feature list that everyone rates as extremely important can be fixed before the full sample is contacted. Pretesting costs little and protects the much larger investment in fielding and analysis. Results from the soft launch will not be pooled with the main sample if any question changes, which keeps the measures consistent across respondents.

Conclusion

Defined constructs, concrete descriptions, randomized prices, tested scales, a stratified sample of adequate size and a nonresponse check give the survey the precision the launch decision requires, while the field test provides the behavioral check that a survey cannot.

References

Babin, B. J. (2019). Essentials of marketing research (7th ed.). Cengage.

Churchill, G. A., Jr. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16(1), 64-73. https://doi.org/10.1177/002224377901600110

Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879-903. https://doi.org/10.1037/0021-9010.88.5.879