Abstract
Today marketing researchers are faced with many challenges in the pursuit of high quality, affordable survey data collection. Frequently researchers would like to measure the perceptions of multiple products or brands and then gather detailed information about each. However, when respondents qualify to answer questions about many of the entities (i.e., brands or products), there may not be enough time to answer questions about all of them. It becomes even more challenging to achieve the necessary sample size per entity when invariably some of the entities are of low incidence. Therefore, the researchers typically assign the respondents to answer questions about a subset of the entities using one of the three most common strategies. Given that incentives are frequently used to address non-response, the assignment strategy is used to minimize the sample size and cost as well as ensure the quality of the data. This paper examines the efficiency and effectiveness of the Random Assignment (randomly assigning respondents to one of the entities for which they qualify), Least-Filled Sampling (assigning respondents to the entity for which they qualify which has the fewest completed responses at the time) and Lowest-Incidence-First Sampling (assigning respondents to the lowest incidence entity for which they qualify) using a study from the Home Improvement Research Institute which evaluated products from fifteen different home improvement projects. The results will be discussed, including an evaluation of bias and error, and two possible solutions to address the error are presented.
Recommended Citation
Norvell, Timothy and Gillespie, Erin
(2022)
"Common Error in Brand Perception Survey Research: Cell Assignment Strategies,"
Journal of Business, Industry, and Economics: Vol. 27, Article 6.
Available at:
https://roar.una.edu/jobie/vol27/iss1/6