Abstract

The proliferation of network-connected consumer electronics has created a sustained challenge for sustainability and data security at the End-of-Current-Use (EoCU) stage. In the United States, devices accumulate in storage rather than entering reuse or recycling pathways, and a substantial share of devices change hands without proper data sanitization. This dissertation characterizes and quantifies consumer behavior at EoCU of consumer electronics in the U.S., with the goal of supporting the design of interventions that improve sustainability and data security outcomes. A survey of 4,000 U.S. respondents (3,747 after data cleaning) was developed and deployed across 10 categories of network-connected devices, capturing stated knowledge, attitudes, past disposition, intended disposition, and data wiping practices. The resulting dataset is analyzed through three complementary methods: k-means clustering, binomial and multinomial logistic regression, and Random Forest simulation, each addressing a distinct facet of the same overarching question: how do stated attitudes and knowledge connect to consumer choice among storing, reselling, recycling, donating, trading in, or throwing away in the trash? First, k-means clustering on 34 stated attitude and knowledge variables produced three consumer segments with weak linkage to age, income, or home location type. Cluster 1 (36.6%) is less knowledgeable of disposition options and more concerned about data, and stores more often than the general population. Cluster 2 (34.7%) is more knowledgeable about disposition options and resells and trades in more often. Cluster 3 (28.7%) is less concerned about data and less knowledgeable of options, and resells, recycles, and throws away more, and stores less. The implication is that intervention design should be customized to segment rather than applied uniformly. Second, modeling the EoCU decision in two stages:  a binomial choice between storing and acting, and a multinomial choice among the active options with recycling as baseline, separated the factors that drive initial retention from those that drive disposition once a consumer has decided to act. The largest factors favoring storing are sentimental value of device, data security concerns when recycling and reselling, and the usefulness of a stored device as a data backup. The largest factors favoring not storing are considering resale value, knowing and trusting recycling, and finding recycling convenient. Knowledge of recycling location strongly drives recycling, while knowledge of online or local resell markets strongly drives reselling. Third, Random Forest simulation reveals that behavior change is non-linear, with critical tipping points typically occurring at neutral rather than extreme attitude positions. Moving the perception of reselling as worthwhile from "Neutral" to "Agree" increases the probability of resale by 26%, with stronger agreement yielding minimal additional gains. The simulation also identifies behavioral trade-offs that linear models obscure, with reselling increases corresponding to decreases in giving away (-14%) and recycling. An additional analysis on data wiping, published at COMPASS '25, finds that 31% of U.S. respondents do not wipe data when dispositioning a device. Combined with the regression finding that data security concerns drive storage rather than active disposition, this identifies the data security paradox at the heart of the EoCU problem: the same concerns that should motivate careful disposition are instead driving devices into indefinite hibernation in consumers' homes. Together, these results provide a comprehensive characterization of U.S. consumer behavior in EoCU electronics, identifying the segments, the linear factors, and the non-linear thresholds, and establish the empirical and methodological foundation for designing targeted interventions that move U.S. consumers toward more sustainable and more secure End-of-Current-Use behavior.

Publication Date

8-4-2026

Document Type

Dissertation

Student Type

Graduate

Degree Name

Sustainability (Ph.D.)

Department, Program, or Center

Sustainability, Department of

College

Golisano Institute for Sustainability

Advisor

Eric Williams

Advisor/Committee Member

Stacey Watson

Advisor/Committee Member

Nathan Williams

Campus

RIT – Main Campus

Download

Share

COinS