The Library
Understanding individual food healthiness perceptions using a computational approach
Tools
Gandhi, Natasha (2022) Understanding individual food healthiness perceptions using a computational approach. PhD thesis, University of Warwick.
|
PDF
WRAP_THESIS_Gandhi_2022.pdf - Submitted Version - Requires a PDF viewer. Download (14Mb) | Preview |
Official URL: http://webcat.warwick.ac.uk/record=b3894940
Abstract
Is granola healthy? What about steak? What knowledge do we use when judging the healthiness of different foods? Previous research recognizes that various nonnutrient related attributes are integral to food healthiness perceptions. However, our understanding of what contributes to healthiness judgments is constrained by researcher or participant assumptions. In this thesis, we uncover these attributes through a study of how different food names co-occur with other words in large-scale language data. Inspired by previous work using language data to predict judgments and choices, we use this data-driven methodology to reveal the psychological underpinnings of healthiness perceptions across all food categories.
Chapter 2 investigates the representation of superfoods in online news articles. By comparing articles written about the same 25 foods in a superfood context versus not, we were able to identify words uniquely used to characterize a food as a superfood. Our findings show that superfoods have the strongest association with health, far outweighing associations between organic and health, and even organic and naturalness. Moreover, mentions of the medicinal properties of these foods frequently occurred in a superfood context. Overall, the findings from this paper illustrate the complex, and often biased, representation of superfoods.
Chapter 3 demonstrates the consistently high ability of knowledge representations (taken from computational models) in predicting people's healthiness judgments of foods. This is found to be the case even if people are shown front-of-pack nutrient and calorie information. We also show how this approach can be used to uncover words and concepts that are more associated with healthy and unhealthy foods. Here, our results show that people's judgments of food healthiness are largely explained by the strength of association with naturalness and rawness. Conversely, we find environmental and social contextual factors are strongly associated with the model's predictions of unhealthiness.
Item Type: | Thesis (PhD) | ||||
---|---|---|---|---|---|
Subjects: | R Medicine > RA Public aspects of medicine T Technology > TX Home economics |
||||
Library of Congress Subject Headings (LCSH): | Nutrition, Nutrition -- Evaluation, Nutrition in mass media, Food, Food -- Quality, Food -- Labeling, Food habits, Health attitudes | ||||
Official Date: | February 2022 | ||||
Dates: |
|
||||
Institution: | University of Warwick | ||||
Theses Department: | Warwick Manufacturing Group | ||||
Thesis Type: | PhD | ||||
Publication Status: | Unpublished | ||||
Supervisor(s)/Advisor: | Walasek, Lukasz ; Meyer, Caroline | ||||
Sponsors: | Engineering and Physical Sciences Research Council | ||||
Format of File: | |||||
Extent: | xiv, 145 pages : illustrations, charts | ||||
Language: | eng |
Request changes or add full text files to a record
Repository staff actions (login required)
View Item |