Dataset Information

From Regression Analysis to Deep Learning: Development of Improved Proxy Measures of State-Level Household Gun Ownership.

ABSTRACT: In the absence of direct measurements of state-level household gun ownership (GO), the quality and accuracy of proxy measures for this variable are essential for firearm-related research and policy development. In this work, we develop two highly accurate proxy measures of GO using traditional regression analysis and deep learning, the former accounting for non-linearities in the covariates (portion of suicides committed with a firearm [FS/S] and hunting license rates) and their statistical interactions. We subject the proxies to extensive model diagnostics and validation. Both our regression-based and deep-learning proxy measures provide highly accurate models of GO with training R² of 96% and 98%, respectively, along with other desirable qualities-stark improvements over the prevalent FS/S proxy (R² = 0.68). Model diagnostics reveal this widely used FS/S proxy is highly biased and inadequate; we recommend that it no longer be used to represent state-level household gun ownership in firearm-related studies.

SUBMITTER: Gomez DB

PROVIDER: S-EPMC7733878 | biostudies-literature | 2020 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

From Regression Analysis to Deep Learning: Development of Improved Proxy Measures of State-Level Household Gun Ownership.

Gomez David Benjamin DB Xu Zhaoyi Z Saleh Joseph Homer JH

Patterns (New York, N.Y.) 20201127 9

In the absence of direct measurements of state-level household gun ownership (GO), the quality and accuracy of proxy measures for this variable are essential for firearm-related research and policy development. In this work, we develop two highly accurate proxy measures of GO using traditional regression analysis and deep learning, the former accounting for non-linearities in the covariates (portion of suicides committed with a firearm [FS/S] and hunting license rates) and their statistical inte ...[more]

PMID: 33336203

Similar Datasets

Project description:Americans can be divided into two groups: those who own guns and those who do not. Although people who own guns and people who do not own guns are often separated along social, cultural, and political lines, it is unclear whether these divisions might extend to population differences in emotional experience. In this paper, we use national cross-sectional data from the 2014 Chapman University Survey on American Fears (n?=?1385) to test whether gun owners are more or less afraid than people who do not own guns. We build on previous work by testing two hypotheses with a broad range of fear-related outcomes, including specific phobias and fears associated with being victimized. The symptom perspective argues that gun ownership is a behavioral expression of fear, that gun owners need guns to protect themselves because they are irrational cowards. Although binary logistic regression models provided minimal support for this idea, there was some evidence to suggest that the odds of gun ownership are higher for people who report being afraid of being victimized by a random/mass shooting. The palliative perspective claims that gun ownership mitigates fear, that owning a powerful weapon is somehow soothing to individuals and their families. Ordinary least squares and negative binomial regression models suggest that people who own guns tend to report lower levels of phobias and victimization fears than people who do not own guns. This general pattern is observed across multiple indicators of fear (e.g., of animals, heights, zombies, and muggings), multiple outcome specifications (continuous and count), and with adjustments for age, gender, race/ethnicity, education, household income, marital status, the presence of children, religious identity, religiosity, religious attendance, political orientation, region of residence, and urban residence. Additional longitudinal research is needed to confirm our findings with a wider range of covariates and fear-related outcomes.

Project description:BackgroundMyanmar lies in the Greater Mekong Subregion where there is artemisinin-resistant Plasmodium falciparum malaria. As the artemisinin compound is the pillar of effective antimalarial therapies, containing the spread of artemisinin resistance is a national and global priority. The use of insecticide-treated bed nets/long-lasting insecticidal nets (ITNs/LLINs) is the key intervention for ensuring the reduction of malaria transmission and the spread of resistant strains, and for eventually eliminating malaria. This study aimed at assessing household ownership of, access to, and utilization of bed nets in areas of Myanmar with and without artemisinin resistance containment measures.MethodsSecondary data from a nationwide community-based malaria survey conducted by the National Malaria Control Program in 2014 were analyzed. Based on evidence of artemisinin resistance, Myanmar was divided into tiers 1, 2, and 3: townships in tiers 1 and 2 were aggregated as the Myanmar Artemisinin Resistance Containment (MARC) areas and were compared with tier 3 townships, which were defined as non-MARC areas. The chi-square test was used to compare groups, and the level of significance was set at P ≤ 0.05.ResultsOf the 6328 households assessed, 97.2% in both MARC and non-MARC areas had at least one bed net (any type), but only 63% of households had ITNs/LLINs. Only 44% of households in MARC areas and 24% in non-MARC areas had adequate numbers of ITNs/LLINs (one ITN/LLIN per two persons, P < 0.001). Nearly 44% of household members had access to ITNs/LLINs. Regarding the utilization of ITNs/LLINs, 45% of household members used them in MARC areas and 36% used them in non-MARC areas (P < 0.001, desired target = 100%). Utilization of ITNs/LLINs among children aged below five years and pregnant women (high malaria risk groups) was low, at 44% and 42%, respectively.ConclusionsThis study highlights the nationwide shortfalls in the ownership of, access to, and utilization of ITNs/LLINs in Myanmar, which is of particular concern in terms of containing the spread of artemisinin resistance. It highlights the need for priority attention to be paid and mobilization of resources in order to improve bed net coverage and utilization through bed net distribution and/or social marketing, information dissemination, and awareness-raising.

Dataset Information

From Regression Analysis to Deep Learning: Development of Improved Proxy Measures of State-Level Household Gun Ownership.

Publications

From Regression Analysis to Deep Learning: Development of Improved Proxy Measures of State-Level Household Gun Ownership.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure