# Publications

For a full list of publications, see my Google Scholar page. Here, I summarize key publications across three research areas: (1) the empirics and ethics of algorithmic prioritization, (2) ethics and decision-making at the nexus of healthcare and K-12 education, and (3) social psychology and educational interventions.

## Empirics and ethics of algorithmic prioritization

In this cluster of papers, I examine (1) the role of supervised machine learning in improving equity in civil rights enforcement, (2) the role that new tools from machine learning can play in helping researchers better connect theory to evidence in quantitative sociology, and (3) public views of and new methods for using large-scale genotyping data in policy-relevant contexts.

1. Lundberg, Johnson, and Stewart. What is your estimand? Defining the target quantity connects statistical evidence to theory. American Sociological Review, 2021.
• Abstract: We make only one point in this article. Every quantitative study must be able to answer the question: what is your estimand? The estimand is the target quantity—the purpose of the statistical analysis. Much attention is already placed on how to do estimation; a similar degree of care should be given to defining the thing we are estimating. We advocate that authors state the central quantity of each analysis—the theoretical estimand—in precise terms that exist outside of any statistical model. In our framework, researchers do three things: (1) set a theoretical estimand, clearly connecting this quantity to theory; (2) link to an empirical estimand, which is informative about the theoretical estimand under some identification assumptions; and (3) learn from data. Adding precise estimands to research practice expands the space of theoretical questions, clarifies how evidence can speak to those questions, and unlocks new tools for estimation. By grounding all three steps in a precise statement of the target quantity, our framework connects statistical evidence to theory.

• Journal publication

• Replication code

1. Johnson and Rostain. Tool for surveillance or spotlight on inequality? Big data and the law. Annual Review of Law and Social Sciences, 2020.
• Abstract: The rise of big data and machine learning is a polarizing force among those studying inequality and the law. Big data and tools like predictive modeling may amplify inequalities in the law, subjecting vulnerable individuals to enhanced surveillance. But these data and tools may also serve an opposite function, shining a spotlight on inequality and subjecting powerful institutions to enhanced oversight. We begin with a typology of the role of big data in inequality and the law. The typology asks questions—Which type of individual or institutional actor holds the data? What problem is the actor trying to use the data to solve?—that help situate the use of big data within existing scholarship on law and inequality. We then highlight the dual uses of big data and computational methods—data for surveillance and data as a spotlight—in three areas of law: rental housing, child welfare, and opioid prescribing. Our review highlights asymmetries where the lack of data infrastructure to measure basic facts about inequality within the law has impeded the spotlight function.

• Journal publication

1. Ye, Johnson, et al. Using machine learning to help vulnerable tenants in New York city. ACM COMPASS ‘19, 2019
• Abstract: To keep housing affordable, the City of New York has implemented rent-stabilization policies to restrict the rate at which the rent of certain units can be increased every year. However, some landlords of these rent-stabilized units try to illegally force their tenants out in order to circumvent rent-stabilization laws and greatly increase the rent they can charge. To identify and help tenants who are vulnerable to such landlord harassment, the New York City Public Engagement Unit (NYC PEU) conducts targeted outreach to tenants to inform them of their rights and to assist them with serious housing challenges. In this paper, we1 collaborated with NYC PEU to develop machine learning models to better prioritize outreach and help to vulnerable tenants. Our best-performing model can potentially help TSU find 59% more buildings where tenants face landlord harassment than the current outreach method using the same resources. The results also highlight the factors that help predict the risk of experiencing tenant harassment, and provide a data-driven and comprehensive approach to improve the city’s policy of proactive outreach to vulnerable tenants.

• Journal publication

• Replication code

1. Johnson, Sotoudeh (equal first authorship), and Conley. Polygenic Scores for Plasticity: A New Tool for Studying Gene-Environment Interplay. Forthcoming at Demography, 2022
• Abstract: Outcomes of interest to demographers—fertility; health; education—are the product of both an individual’s genetic makeup and his or her social environment. Yet Gene $\times$ Environment research (GxE) currently deploys a limited toolkit on the genetic side to study gene-environment interplay: polygenic scores (PGS, or what we call mPGS) that reflect the influence of genetics on levels of an outcome. The purpose of the present paper is to develop a genetic summary measure better suited for GxE research. We develop what we call variance polygenic scores (vPGS), or polygenic scores that reflect genetic contributions to plasticity in outcomes. The first part of the analysis uses the UK Biobank (N $\sim$ 326,000 in the training set) and the Health and Retirement Study (HRS) to compare four approaches for constructing polygenic scores for plasticity. The results show that widely-used methods for discovering which genetic variants affect outcome variability fail to serve as distinctive new tools for GxE. Then, using the polygenic scores that do capture distinctive genetic contributions to plasticity, we analyze heterogeneous effects of a UK education reform on health and educational attainment. The results show the properties of a new tool useful for population scientists studying the interplay of nature and nurture and for population-based studies that are releasing polygenic scores to applied researchers.

• Biorxiv

1. Zhang, Johnson, Novembre, Freeland, and Conley. Public attitudes toward genetic risk scoring in medicine and beyond. Social Science and Medicine, 2021.
• Abstract: Advances in genomics research have led to the development of polygenic risk scores, which numerically summarize genetic predispositions for a wide array of human outcomes. Initially developed to characterize disease risk, polygenic risk scores can now be calculated for many non-disease traits and social outcomes, with the potential to be used not only in health care but also other institutional domains. In this study, we draw on a nationally-representative survey of U.S. adults to examine three sets of lay attitudes toward the deployment of genetic risk scores in a variety of medical and non-medical domains: 1. abstract belief about whether people should be judged on the basis of genetic predispositions; 2. concrete attitudes about whether various institutions should be permitted to use genetic information; and 3. personal willingness to provide genetic information to various institutions. Results demonstrate two striking differences across these three sets of attitudes. First, despite almost universal agreement that people should not be judged based on genetics, there is support, albeit varied, for institutions being permitted to use genetic information, with support highest for disease outcomes and in reproductive decision-making. We further find significant variation in personal willingness to provide such information, with a majority of respondents expressing willingness to provide information to health care providers and relative finder services, but less than a quarter expressing willingness to do so for an array of other institutions and services. Second, while there are no demographic differences in respondents’ abstract beliefs about judging based on genetics, demographic differences emerge in permissibility ratings and personal willingness. Our results should inform debates about the deployment of polygenic scores in domains within and beyond medicine.

• Journal website

• Replication code

## Ethics and decision-making at the nexus of healthcare and K-12 education

In this cluster of papers, I examine institutional linkages between two sectors that help people living with disabilities: (1) the healthcare sector, which takes responsibility for medical aspects of disability and (2) K-12 school districts, which take responsibility for disability as they impact learning-related outcomes. I study the role of parents and patient advocacy groups in both defining which institutions take responsibility for disability and decision-making within those institutions. Other papers examine the role of disease stigma in shaping public views of people with disabilities. These are a select sample and the full list is found on my CV and google scholar.

1. Johnson, Danis, and Hafner-Eaton. US state variation in autism insurance mandates: Balancing access and fairness. Autism, 2014.
• Abstract: This article examines how nations split decision-making about health services between federal and sub-federal levels, creating variation between states or provinces. When is this variation ethically acceptable? We identify three sources of ethical acceptability—procedural fairness, value pluralism, and substantive fairness—and examine these sources with respect to a case study: the fact that only 30 out of 51 US states or territories passed mandates requiring private insurers to offer extensive coverage of autism behavioral therapies, creating variation for privately insured children living in different US states. Is this variation ethically acceptable? To address this question, we need to analyze whether mandates go to more or less needy states and whether the mandates reflect value pluralism between states regarding government’s role in health care. Using time-series logistic regressions and data from National Survey of Children with Special Health Care Needs, Individual with Disabilities Education Act, legislature political composition, and American Board of Pediatrics workforce data, we find that the states in which mandates are passed are less needy than states in which mandates have not been passed, what we call a cumulative advantage outcome that increases between-state disparities rather than a compensatory outcome that decreases between-state disparities. Concluding, we discuss the implications of our analysis for broader discussions of variation in health services provision.

• Journal website

1. Johnson, Barrett, and Sisti. The ethical boundaries of patient and advocate influence on DSM-5. Harvard Review of Psychiatry, 2013.

• Journal website

1. Sisti and Johnson. Revision and representation: The controversial case of the DSM-5. Public Affairs Quarterly, 2015.
1. Johnson, Harkins, Cary, and Karlawish. The relative contributions of disease label and disease prognosis to Alzheimer’s stigma: A vignette-based experiment. Social Science and Medicine, 2015.
• Abstract: The classification of Alzheimer’s disease is undergoing a significant transformation. Researchers have created the category of “preclinical Alzheimer’s,” characterized by biomarker pathology rather than observable symptoms. Diagnosis and treatment at this stage could allow preventing Alzheimer’s cognitive decline. While many commentators have worried that persons given a preclinical Alzheimer’s label will be subject to stigma, little research exists to inform whether the stigma attached to the label of clinical Alzheimer’s will extend to a preclinical disorder that has the label of “Alzheimer’s” but lacks the symptoms or expected prognosis of the clinical form. The present study sought to correct this gap by examining the foundations of stigma directed at Alzheimer’s. It asked: do people form stigmatizing reactions to the label “Alzheimer’s disease” itself or to the condition’s observable impairments? How does the condition’s prognosis modify these reactions? Data were collected through a web-based experiment with N = 789 adult members of the U.S. general population (median age = 49, interquartile range, 32–60, range = 18–90). Participants were randomized through a 3 × 3 design to read one of 9 vignettes depicting signs and symptoms of mild stage dementia that varied the disease label (“Alzheimer’s” vs. “traumatic brain injury” vs. no label) and prognosis (improve vs. static vs. worsen symptoms). Four stigma outcomes were assessed: discrimination, negative cognitive attributions, negative emotions, and social distance. The study found that the Alzheimer’s disease label was generally not associated with more stigmatizing reactions. In contrast, expecting the symptoms to get worse, regardless of which disease label those symptoms received, resulted in higher levels of perceived structural discrimination, higher pity, and greater social distance. These findings suggest that stigma surrounding pre-clinical Alzheimer’s categories will depend highly on the expected prognosis attached to the label. They also highlight the need for models of Alzheimer’s-directed stigma that incorporate attributions about the condition’s mutability.

• Journal website

## Social psychology and educational interventions

In this cluster of papers from early in my research career, I contributed to projects on how to use tools from social psychology to reduce inequalities in K-12 educational outcomes.

1. Yeager, Johnson, Spitzer, Trzesniewski, Powers, Dweck. The far-reaching effects of believing people can change: Implicit theories of personality shape stress, health, and achievement during adolescence. Journal of Personality and Social Psychology, 2014.
• Abstract: The belief that personality is fixed (an entity theory of personality) can give rise to negative reactions to social adversities. Three studies showed that when social adversity is common—at the transition to high school—an entity theory can affect overall stress, health, and achievement. Study 1 showed that an entity theory of personality, measured during the 1st month of 9th grade, predicted more negative immediate reactions to social adversity and, at the end of the year, greater stress, poorer health, and lower grades in school. Studies 2 and 3, both experiments, tested a brief intervention that taught a malleable (incremental) theory of personality—the belief that people can change. The incremental theory group showed less negative reactions to an immediate experience of social adversity and, 8 months later, reported lower overall stress and physical illness. They also achieved better academic performance over the year. Discussion centers on the power of targeted psychological interventions to effect far-reaching and long-term change by shifting interpretations of recurring adversities during developmental transitions.

• Journal website