Enhancing big data in the social sciences with crowdsourcing: Data augmentation practices, techniques, and opportunities

Relationships, race/ethnicity, gender, age, and living kidney donation evaluation willingness

Racial/ethnic and gender disparities in living donor kidney transplantation are large and persistent but incompletely explained. One previously unexplored potential contributor to these disparities is differential willingness to donate to recipients in specific relationships such as children, parents, and friends. We collected and analyzed data from an online sample featuring an experimental vignette in which respondents were asked to rate their willingness to donate to a randomly chosen member of their family or social network. Results show very large differences in respondents' willingness to donate to recipients with different relationships to them, favoring children, spouses/partners, siblings, and parents, and disfavoring friends, aunts/uncles, and coworkers. Evidence suggesting an interactive effect between relationship, respondent race/ethnicity, respondent or recipient gender, was limited to a few cases. At the p < 0.05 level, the parent-recipient gender interaction was statistically significant, favoring mothers over fathers, as was other/multiracial respondents' greater willingness to donate to friends compared to Whites. Additionally, other interactions were significant at the p < 0.10 level, such as Hispanics' and women's higher willingness to donate to parents compared to Whites and men respectively, women's lower willingness to donate to friends compared to men, and Blacks' greater willingness to donate to coworkers than Whites. We also examined differences by age and found that older respondents were less willing to donate to recipients other than their parents. Together these results suggest that differential willingness to donate by relationship group may be a moderately important factor in understanding racial/ethnic and gender disparities in living donor kidney transplantation.

Response Consistency of Crowdsourced Web-Based Surveys on Type 1 Diabetes

Although Amazon Mechanical Turk facilitates the quick surveying of a large sample from various demographic and socioeconomic backgrounds, it may not be an optimal platform for obtaining reliable diabetes-related information from the online type 1 diabetes population.

Explaining Firm Performance During the COVID-19 With fsQCA: The Role of Supply Network Complexity, Inventory Turns, and Geographic Dispersion

The COVID-19 pandemic has significantly affected firm performance. As a result, many studies have examined the significance of supply network complexity. Our paper uses the fuzzy set qualitative comparative analysis (fsQCA) method to investigate the causal relationships among the supply network complexity, geographic dispersion, inventory turns, and firm performance. Using a sample of 263 Chinese listed firms, we find that no single factor is necessary to achieve high firm performance during COVID-19 and reveal four paths to produce high performance: operational level driven, supply base complexity driven, and customer base complexity driven with supplier distance, and supply network complexity absence. Furthermore, our findings suggest that supply-based complexity-driven and customer-based complexity-driven can improve firm performance, but not all supply network complexity dimensions can improve firm performance. Hence, firms need to choose the suitable path based on their specific situations.

Thirty-three myths and misconceptions about population data: from data capture and processing to linkage

Databases covering all individuals of a population are increasingly used for research and decision-making. The massive size of such databases is often mistaken as a guarantee for valid inferences. However, population data have characteristics that make them challenging to use. Various assumptions on population coverage and data quality are commonly made, including how such data were captured and what types of processing have been applied to them. Furthermore, the full potential of population data can often only be unlocked when such data are linked to other databases. Record linkage often implies subtle technical problems, which are easily missed. We discuss a diverse range of myths and misconceptions relevant for anybody capturing, processing, linking, or analysing population data. Remarkably, many of these myths and misconceptions are due to the social nature of data collections and are therefore missed by purely technical accounts of data processing. Many are also not well documented in scientific publications. We conclude with a set of recommendations for using population data.

The Acoustic Dissection of Cough: Diving Into Machine Listening-based COVID-19 Analysis and Detection

OBJECTIVES

The coronavirus disease 2019 (COVID-19) has caused a crisis worldwide. Amounts of efforts have been made to prevent and control COVID-19's transmission, from early screenings to vaccinations and treatments. Recently, due to the spring up of many automatic disease recognition applications based on machine listening techniques, it would be fast and cheap to detect COVID-19 from recordings of cough, a key symptom of COVID-19. To date, knowledge of the acoustic characteristics of COVID-19 cough sounds is limited but would be essential for structuring effective and robust machine learning models. The present study aims to explore acoustic features for distinguishing COVID-19 positive individuals from COVID-19 negative ones based on their cough sounds.

METHODS

By applying conventional inferential statistics, we analyze the acoustic correlates of COVID-19 cough sounds based on the ComParE feature set, i.e., a standardized set of 6,373 acoustic higher-level features. Furthermore, we train automatic COVID-19 detection models with machine learning methods and explore the latent features by evaluating the contribution of all features to the COVID-19 status predictions.

RESULTS

The experimental results demonstrate that a set of acoustic parameters of cough sounds, e.g., statistical functionals of the root mean square energy and Mel-frequency cepstral coefficients, bear essential acoustic information in terms of effect sizes for the differentiation between COVID-19 positive and COVID-19 negative cough samples. Our general automatic COVID-19 detection model performs significantly above chance level, i.e., at an unweighted average recall (UAR) of 0.632, on a data set consisting of 1,411 cough samples (COVID-19 positive/negative: 210/1,201).

CONCLUSIONS

Based on the acoustic correlates analysis on the ComParE feature set and the feature analysis in the effective COVID-19 detection approach, we find that several acoustic features that show higher effects in conventional group difference testing are also higher weighted in the machine learning models.

Using K-Means Cluster Analysis and Decision Trees to Highlight Significant Factors Leading to Homelessness

Homelessness has been a persistent social concern in the United States. A combination of political and economic events since the 1960s has driven increases in poverty that, by 1991, had surpassed 1928 depression era levels in some accounts. This paper explores how the emerging field of behavioral economics can use machine learning and data science methods to explore preventative responses to homelessness. In this study, machine learning data mining strategies, specifically K-means cluster analysis and later, decision trees, were used to understand how environmental factors and resultant behaviors can contribute to the experience of homelessness. Prevention of the first homeless event is especially important as studies show that if a person has experienced homelessness once, they are 2.6 times more likely to have another homeless episode. Study findings demonstrate that when someone is at risk for not being able to pay utility bills at the same time as they experience challenges with two or more of the other social determinants of health, the individual is statistically significantly more likely to have their first homeless event. Additionally, for men over 50 who are not in the workforce, have a health hardship, and experience two or more other social determinants of health hardships at the same time, the individual has a high statistically significant probability of experiencing homelessness for the first time.

A Web-Based Digital Contact Tracing Strategy Addresses Stigma Concerns Among Individuals Evaluated for COVID-19

Background: Conventional contact tracing approaches have not kept pace with the scale of the coronavirus disease 2019 (COVID-19) pandemic and the highly anticipated smartphone applications for digital contact tracing efforts are plagued by low adoption rates attributed to privacy concerns; therefore, innovation is needed in this public health capability. Methods: This study involved a cross-sectional, nonrepresentative, online survey in the United States of individuals tested for COVID-19. Testing survey items measured the performance of conventional contact tracing programs, quantified the stigma related to the notification of COVID-19 close contacts, and assessed the acceptability of a website service for digital contact tracing. Results: A sample of 668 (19.9%) individuals met the inclusion criteria and consented to participation. Among the 95 participants with COVID-19, results were received after a median of 2 days, 63.2% interacted with a contact tracing program a median of 2 days after receiving test results, 62.1% had close contacts, and 37.1% of participants with COVID-19 and close contacts did not disclose their results to all close contacts. Among all participants, 17% had downloaded a mobile application and 40.3% reported interest in a website service. One hundred and nine participants perceived stigma with the disclosure of COVID-19 test results; of these, 58.7% reported that a website service for close contact notification would decrease this stigma. Discussion: Conventional contact tracing programs did not comprehensively contact individuals who tested positive for COVID-19 nor did so within a meaningful time frame. Digital contact tracing innovations may address these shortcomings; however, the low penetration of mobile application services in the United States indicates that a suite of digital contact tracing tools, including website services, are warranted for a more exhaustive coverage of the population. Conclusions: Public health officials should develop a complementary toolkit of digital contact tracing strategies to enable effective pandemic containment strategies.

Mobile Sensing Apps and Self-management of Mental Health During the COVID-19 Pandemic: Web-Based Survey

BACKGROUND

During the COVID-19 pandemic, people had to adapt their daily life routines to the currently implemented public health measures, which is likely to have resulted in a lack of in-person social interactions, physical activity, or sleep. Such changes can have a significant impact on mental health. Mobile sensing apps can passively record the daily life routines of people, thus making them aware of maladaptive behavioral adjustments to the pandemic.

OBJECTIVE

This study aimed to explore the views of people on mobile sensing apps that passively record behaviors and their potential to increase awareness and helpfulness for self-managing mental health during the pandemic.

METHODS

We conducted an anonymous web-based survey including people with and those without mental disorders, asking them to rate the helpfulness of mobile sensing apps for the self-management of mental health during the COVID-19 pandemic. The survey was conducted in May 2020.

RESULTS

The majority of participants, particularly those with a mental disorder (n=106/148, 72%), perceived mobile sensing apps as very or extremely helpful for managing their mental health by becoming aware of maladaptive behaviors. The perceived helpfulness of mobile sensing apps was also higher among people who experienced a stronger health impact of the COVID-19 pandemic (β=.24; 95% CI 0.16-0.33; P<.001), had a better understanding of technology (β=.17; 95% CI 0.08-0.25; P<.001), and had a higher education (β=.1; 95% CI 0.02-0.19; P=.02).

CONCLUSIONS

Our findings highlight the potential of mobile sensing apps to assist in mental health care during the pandemic.