Theory-based Artificial Intelligence Generated Digital Nudges and Generation Z Female Self-Esteem, Well-Being, and Purchase Intention In The Context of Beauty

Date of Award

2026

Document Type

Dissertation

Degree Name

Doctor of Organization Development & Change (D.O.D.C.)

Department

Organization Development

First Advisor

William Sawaya (Committee Chair)

Second Advisor

Carol Heckman (Other)

Third Advisor

Margaret Brooks (Committee Member)

Fourth Advisor

Fei Weisstein (Committee Member)

Abstract

This dissertation examines how theory-based AI-generated digital nudges influence emotional well-being, state self-esteem, and purchase intention among Generation Z females in the context of beauty. The dissertation is structured as a two-step research study. Step 1, the Brand Trust Survey, established the empirical basis for selecting relevant high-trust and low-trust beauty brand anchors. Step 2, the Nudge Experiment, examined the efficacy of theory-based AI-generated digital nudges across multiple controlled conditions in the context of beauty. In the Nudge Experiment, subjects completed a pre-test survey measuring state self-esteem using the State Self-Esteem Scale (SSES) and emotional well-being using the Positive and Negative Affect Schedule (PANAS). Subjects were then exposed to one of 10 experimental conditions: Control Group; SSES Sub-Scale Nudge Conditions; SSES Sub-Scale Nudge + Multi-Category Brand HT LT Conditions; and SSES Sub-Scale Nudge + Skincare Category Brand HT LT Conditions. Each condition consisted of a series of seven short, locked messages delivered via Qualtrics. Following exposure to the messages, subjects completed post-test SSES and PANAS measures, as well as a three-item purchase intention scale. Two brief attention check items were embedded in the survey to support data quality. For the Nudge Experiment, the final dataset analyzed comprised N = 614 subjects across the 10 experimental conditions. The hypotheses tested whether theory-based AI-generated digital nudges produced stronger post-exposure SSES and PANAS outcomes than the Control Group, whether the assigned brand trust level strengthened the effect of the brand-framed nudge conditions on state self-esteem, and whether higher post-test SSES was associated with higher purchase intention. A supplemental bootstrapped mediation analysis further indicated that post-test state self-esteem functioned as a significant indirect pathway linking digital nudge exposure to purchase intention, although the total and direct effects of digital nudge exposure on purchase intention remained negative. This study contributes causal evidence to the emerging intersection of ethical AI messaging, behavioral science, and beauty consumer decision-making. More specifically, it identifies a practical, ethical, and commercially sustainable path for testing theory-based AI-generated digital nudges in a controlled digital communication format.

Link to Full Text

Share

COinS