Browsing by Subject "bias"

Now showing 1 - 7 of 7
  • Thumbnail Image
    Item
    Analysis of AI Models for Student Admissions: A Case Study
    (ACM, 2023-03) Van Basum, Kelly; Fang, Shaiofen; Computer and Information Science, School of Science
    This research uses machine learning-based AI models to predict admissions decisions at a large urban research university. Admissions data spanning five years was used to create an AI model to determine whether a given student would be directly admitted into the School of Science under various scenarios. During this time, submission of standardized test scores as part of a student's application became optional which led to interesting questions about the impact of standardized test scores on admission decisions. We first developed AI models and analyzed these models to understand which variables are important in admissions decisions, and how the decision to exclude test scores affects the demographics of the students who are admitted. We then evaluated the predictive models to detect and analyze biases these models may carry with respect to three variables chosen to represent sensitive populations: gender, race, and whether a student was the first in his family to attend college.
  • Thumbnail Image
    Item
    Effect of Socioeconomic Status Bias on Medical Student–Patient Interactions Using an Emergency Medicine Simulation
    (Wiley, 2017-04) Pettit, Katie E.; Turner, Joseph S.; Kindrat, Jason K.; Blythe, Gregory J.; Hasty, Greg E.; Perkins, Anthony J.; Ashburn-Nardo, Leslie; Milgrom, Lesley B.; Hobgood, Cherri D.; Cooper, Dylan D.; Emergency Medicine, School of Medicine
    Objectives Implicit bias in clinical decision making has been shown to contribute to healthcare disparities and results in negative patient outcomes. Our objective was to develop a high‐fidelity simulation model for assessing the effect of socioeconomic status (SES) on medical student (MS) patient care. Methods Teams of MSs were randomly assigned to participate in a high‐fidelity simulation of acute coronary syndrome. Cases were identical with the exception of patient SES, which alternated between a low‐SES homeless man and a high‐SES executive. Students were blinded to study objectives. Cases were recorded and scored by blinded independent raters using 24 dichotomous items in the following domains: 13 communication, six information gathering, and five clinical care. In addition, quantitative data were obtained on the number of times students performed the following patient actions: acknowledged patient by name, asked about pain, generally conversed, and touching the patient. Fisher's exact test was used to test for differences between dichotomous items. For continuous measures, group differences were tested using a mixed‐effects model with a random effect for case to account for multiple observations per case. Results Fifty‐eight teams participated in an equal number of high‐ and low‐SES cases. MSs asked about pain control more often (p = 0.04) in patients of high SES. MSs touched the low‐SES patient more frequently (p = 0.01). There were no statistically significant differences in clinical care or information gathering measures. Conclusions This study demonstrates more attention to pain control in patients with higher SES as well as a trend toward better communication. Despite the differences in interpersonal behavior, quantifiable differences in clinical care were not seen. These results may be limited by sample size, and larger cohorts will be required to identify the factors that contribute to SES bias.
  • Thumbnail Image
    Item
    Heuristics and Biases in the Intuitive Projection of Retail Sales
    (1987-08) Cox, Anthony D.; Summers, John O.
    Retail merchandise buyers are shown to exhibit a nonregressive bias when making sales projections. A quantitative model based on the principle of statistical regression is found to outperform the judgmental sales predictions of experienced buyers. Implications for the appropriate roles of intuitive and model-based decision making in retail merchandise buying are discussed.
  • Thumbnail Image
    Item
    How to more effectively determine what is true: The limits of intuition
    (Elsevier, 2020-08) Kaefer, Martin; Kalfa, Nicolas; Herbst, Katherine W.; Harper, Luke; Beckers, Goedele M. A.; Bagli, Darius; Fossum, Magdalena; Pediatrics, School of Medicine
    The plethora of scientific data and explosion of published materials often leave it challenging to develop a clear and concise overview of many scientific topics. A number of factors may contribute to our misunderstanding. It is the focus of this article to describe primary reasons for failure to establish a clear, factual and functional understanding regarding scientific areas of inquiry.
  • Thumbnail Image
    Item
    An Interactive Approach to Bias Mitigation in Machine Learning
    (IEEE, 2021-10) Wang, Hao; Mukhopadhyay, Snehasis; Xiao, Yunyu; Fang, Shiaofen; Computer and Information Science, School of Science
    Underrepresentation and misrepresentation of protected groups in the training data is a significant source of bias for Machine Learning (ML) algorithms, resulting in decreased confidence and trustworthiness of the generated ML models. Such bias can be mitigated by incorporating both objective as well as subjective (through human users) measures of bias, and compensating for them by means of a suitable selection algorithm over subgroups of training data. In this paper, we propose a methodology of integrating bias detection and mitigation strategies through interactive visualization of machine learning models in selected protected spaces. In this approach, a (partially generated) ML model performance is visualized and evaluated by a human user or a community of human users in terms of potential presence of bias using both objective and subjective criteria. Guided by such human feedback, the ML algorithm can implement a variety of remedial sampling strategies to mitigate the bias using an iterative human-in-the-loop approach. We also provide experimental results with a benchmark ML dataset to demonstrate that such an interactive ML approach holds considerable promise in detecting and mitigating bias in ML models.
  • Thumbnail Image
    Item
    Pre-Clinical Medical Students' Attitudes Towards Psychiatry
    (2022-05) Opperman, Michael; Smith, Alyssa; McCann, Joseph; Chastain, Jonathan; Schiller, Brennan; Thomas, Alexander; Jivens, Morgan; Schargorodsky, David; Scofield, David; Grant, Larrilyn; Sweazey, Robert; Richardson, Jenelle; Plawecki, Martin
  • Thumbnail Image
    Item
    Women in trades: Barriers and challenges
    (Office of the Vice Chancellor for Research, 2016-04-08) Walker, Marquita
    This research addresses the implicit bias women seeking to enter and remain in the male-dominated building trades experience as a result of their gender. Implicit bias within masculine dominated workplaces has a deleterious effect on the hiring and retention of women in trades, so it is important for policy makers, employers, stockholders, and union officials to address these deficiencies through a strategy for decreasing masculine dominance in the workplace. As skill shortages and a weakened labor supply loom for the construction industry, it is important to seriously consider why women’s participation in the construction industry remains below legal and necessary limits. Hiring and retaining more women in the building trades would fill the predicted future construction vacancies. Situated in political economy theory, this study surveyed 29 women in union-sponsored apprenticeship programs. Analysis of data collected from survey instruments and personal interviews reveal gatekeeping barriers and covert discriminatory practices to women seeking to enter the building trades. Recommendations for addressing these barriers include enforcing government policies which mandate more women in the trades, changing the masculine culture of union/employer construction workplaces through the promotion of mentoring components in apprenticeship programs which provide to women one-on-one support, and making concerted efforts within the firm toward implementation of more gender neutral, family-friendly, and work-life balance policies.