• Home
  • Search
  • Cluster trimming: Improving segment homogeneity in product testing with consumers
  • https://doi.org/10.1016/j.foodqual.2025.105782Copy DOI Icon

Cluster trimming: Improving segment homogeneity in product testing with consumers

Show More
  • Abstract
  • Literature Map
  • References
  • Similar Papers
Abstract

Sensory and consumer researchers are often required to identify clusters of consumers based on product liking. A widespread, if informally held, belief is that many studies contain a group of consumers whose data differ from the product means in their assigned cluster. The implication could be less discrimination between samples within a cluster, leading to poorer decision-making. This research presents a variant of existing cluster analysis methods to potentially improve insights by excluding subjects with non-typical response behaviour. The approach is intuitive, easy to implement, and cluster algorithm agnostic. After removing non-discriminators (i.e., those who like all samples equally), the principle is to conduct cluster analysis and then calculate the correlation of each subject's liking ratings with the mean in their assigned cluster, setting a fixed “trim out” threshold for the correlation coefficient. Consumers with correlations below the threshold are classified as non-typical and excluded. After introducing cluster trimming in a case study, the method is applied to another 29 data sets representing category appraisal and product optimisation applications. On average, the cluster-trimming process resulted in around 1/3 of study participants being “trimmed out.” These consumers tended to discriminate less between samples than the participants who were retained following cluster trimming. While cluster trimming was found to add some value regarding product discrimination, this benefit was somewhat obscured by the reduction in cluster size. Exploring within-cluster heterogeneity, made possible by examining histograms of correlation coefficients (r-values), is recommended. Because the cluster trimming approach is correlational, it should not be used in product studies with too few samples ( n < 6). • A correlational approach for refinement of consumer segments based on liking. • Subjects are retained or trimmed out based on similarity with the cluster product means. • Improves cluster homogeneity but has little impact on sample discrimination. • Cluster trimming is easy to perform and does not bias mean liking ratings.

Similar Papers
  • Research Article
  • Citations2

Accompanying the Development of Sidetapa‘s Fruit Brem as a Leading Tourism Product in Buleleng Regency

  • Jun 14, 2023
  • Jurnal Kewirausahaan dan Bisnis
  • Ni Made Ary Widiastini +3
  • Research Article
  • Citations3

The effects of comparative product testing on industry and trade: Findings of a research project

  • Dec 01, 1984
  • Journal of Consumer Policy
  • Hans Raffée +1
  • Conference Article
  • Citations3

Finding Gas Using Underbalanced Drilling in a Tight Reservoir

  • Feb 24, 2010
  • John Bowling +2
  • Research Article
  • Citations331

Generalizing test-enhanced learning from the laboratory to the classroom

  • Apr 01, 2007
  • Psychonomic Bulletin &amp; Review
  • Mark A Mcdaniel +2
  • Conference Article
  • Citations2

Multi-gb/s ic test challenges and solutions

  • Sep 01, 2003
  • B.G West
  • Research Article

Assessing a dryer state using the parameters of a spent drying agent

  • Oct 01, 2019
  • IOP Conference Series: Earth and Environmental Science
  • V F Sorochinsky +1
  • Research Article
  • Citations2

Developing Scientific Article Writing Materials Using a Hybrid Learning Approach: Insights from an Indonesian Language for Foreign Speakers (BIPA) Program

  • Jul 30, 2022
  • Jurnal Pendidikan: Teori, Penelitian, dan Pengembangan
  • Ah Rofiuddin +3
  • Research Article
  • Citations3

Development of Connecting Extending Review (CER) Learning Model to Improve Student’s Mathematical Reasoning Ability

  • Sep 11, 2021
  • Asian Journal of Applied Sciences
  • Andi Muhammad Iqbal Akbar Asfar +4
  • Research Article
  • Citations19

Does the product test really make a difference? Evidence from the launch of a new wine

  • Aug 20, 2018
  • Food Quality and Preference
  • Alba García-Milon +3
  • Research Article
  • Citations1

New trends in product testing

  • Oct 01, 1975
  • Business Horizons
  • Joseph Nemec
  • Front Matter
  • Citations15

Rhinitis with negative skin tests and absent serum allergen-specific IgE: More evidence for local IgE?

  • Nov 01, 2009
  • The Journal of Allergy and Clinical Immunology
  • Louisa K James +1
  • Conference Article

Oil Well Transient Testing by Injection of Miscible Fluid

  • May 05, 2008
  • Joseph James Voelker
  • Conference Article

Maximum short quantity in Japanese and finish in two perception tests with F0 and db variants

  • Oct 04, 2004
  • Toshiko Isei-Jaakkola
  • Research Article
  • Citations29

Geochemistry and sources of hydrate-bound gas in the Shenhu area, northern south China sea: Insights from drilling and gas hydrate production tests

  • Jan 01, 2022
  • Journal of Petroleum Science and Engineering
  • Qianyong Liang +15
  • Research Article
  • Citations2

The Acquisition of Derivational Suffixes by Students of English at the University of Human Development

  • Oct 13, 2018
  • Journal of University of Human Development
  • Inaad Mutlib Sayer +1
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.