Amazon Mechanical Turk Platform Research Articles

Strategic philanthropy of athletes after transgression: examining the congruence effect of pre-transgression and transgression issues

PurposeThis study comprehensively explains how sports consumers evaluate athletes’ post-transgression philanthropic activities. We specifically focus on the congruence effect between the pre-transgression philanthropic endeavors and the transgression issue, shedding light on the effectiveness of these strategies in reinstating the positive image of athletes and associated entities.Design/methodology/approachThis study employed a rigorous research design, replicating two studies using sports-related (n = 409) and non-sports-related (n = 404) transgression cases. Data were collected by the Amazon Mechanical Turk platform. A series of experimental studies aimed to investigate the congruence mechanism underlying athletes’ post-transgression philanthropic efforts.FindingsWhen the post-transgression philanthropic initiative is related to the transgression, sport consumers are less likely to view it skeptically and are more inclined to positively evaluate the brand attitude and purchase intention, especially when the transgression is unrelated to the pre-transgression philanthropic efforts.Research limitations/implicationsAs is the case with most research, this study has a limitation. This study used a fictitious athlete name to prevent any prior biases or preconceived notions about the athletes and to avoid any unforeseen influences of personal attitudes toward the athlete. However, designing this study around a fictitious athlete may pose construct validity issues because it may not reflect real-life interactions with the athletes. To increase the validity of findings, future research should aim to replicate the current findings using the names of actual athletes.Originality/valueThis unique approach provides valuable insights and equips sports marketers and brand managers with effective strategies to restore the positive image of athletes and associated entities after a transgression has been made public, empowering them to make informed decisions in challenging situations.

It depends: Platform dependence, schedule control, and satisfaction among gig workers on Amazon's Mechanical Turk platform

AbstractUnderstanding the risks and rewards of gig work requires examining how gig workers themselves feel about their work and how work satisfaction may vary even among workers on the same gig platform. Platform dependence theory posits that financially disadvantaged workers who depend on gig work for essential income will be less satisfied, in part, because they are less able to work their preferred schedules. These predictions, however, have not been tested quantitatively. Using survey data from workers who do gig microtasks on Amazon's Mechanical Turk platform (MTurk), we examine how platform dependence, schedule mismatches (between preferred and actual work schedules), and work stress are related to work satisfaction. We find that all three factors are associated with lower satisfaction. Furthermore, we find support for a hypothesized mediating mechanism: platform dependence lowers satisfaction, in part, because it is associated with more mismatches between workers' preferred and actual schedules. We conclude by discussing reforms that could improve satisfaction for all microtask workers.

The relationship between existential isolation and suicidal ideation among formerly incarcerated individuals

Background and purpose Formerly incarcerated people die by suicide at a rate 6.75 times higher than the general population. In this study we ask whether existential isolation, the feeling of being alone in one’s subjective experiences, offers new insights into this staggering statistic. We specifically ask whether existential isolation predicts suicidal ideation among formerly incarcerated individuals even after accounting for their levels of interpersonal isolation, emotional support, and instrumental support. Method We examined cross-sectional survey data from American participants (n = 946) recruited over the Amazon Mechanical Turk (MTurk) platform. Results Consistent with prediction, we found that people with an incarceration history reported more existential isolation, but not more interpersonal isolation, than those without an incarceration history. A hierarchical regression analysis revealed that existential isolation explained additional variance in suicidal ideation beyond what was explained by interpersonal isolation, instrumental support, and emotional support. Fisher’s r-to-z transformations demonstrated that the magnitude of the relationship between existential isolation and suicidal ideation was stronger among those with an incarceration history than among those without. Finally, existential isolation mediated the relationship between incarceration and suicidal thoughts, even when holding constant other predictors of suicidal ideation. Conclusion Reentry programs may target existential isolation among formerly incarcerated individuals to improve emotional wellbeing.

Evaluating the Impact of Uncertainty Visualization on Model Reliance.

Machine learning models have gained traction as decision support tools for tasks that require processing copious amounts of data. However, to achieve the primary benefits of automating this part of decision-making, people must be able to trust the machine learning model's outputs. In order to enhance people's trust and promote appropriate reliance on the model, visualization techniques such as interactive model steering, performance analysis, model comparison, and uncertainty visualization have been proposed. In this study, we tested the effects of two uncertainty visualization techniques in a college admissions forecasting task, under two task difficulty levels, using Amazon's Mechanical Turk platform. Results show that (1) people's reliance on the model depends on the task difficulty and level of machine uncertainty and (2) ordinal forms of expressing model uncertainty are more likely to calibrate model usage behavior. These outcomes emphasize that reliance on decision support tools can depend on the cognitive accessibility of the visualization technique and perceptions of model performance and task difficulty.

Solidarity and collective issues in remote crowd work: A mixed methods study of the Amazon Mechanical Turk online forum

AbstractThe article deals with collective issues and emerging forms of solidarity in remote crowd work. Using the example of the Amazon Mechanical Turk platform, we analyse communication in one of the oldest online labour forums (OLF) with a mixed‐methods‐approach (web scraping, topic modelling and qualitative content analysis) over a 10‐year period. We identify six broader themes of collective relevance whose importance varies over time. The results indicate that elements of solidarity emerge as crowd workers communicate about deprivations, develop collective orientations, and invoke injustice frames. However, there are no indications of more profound collective activities, which may be due to different types of crowd workers, some of whom have conflicting orientations. We develop assumptions about restrictions of OLF in facilitating collective action and tasks for organisers. The research contributes to a better understanding of mobilisation issues in remote platform labour.

JSSE: Joint Sequential Semantic Encoder for Zero-Shot Event Recognition

Zero-shot learning (ZSL) is a paradigm in transfer learning that aims to recognize unknown categories by having a mere description of them. The problem of ZSL has been thoroughly studied in the domain of static object recognition, however, ZSL for dynamic events (ZSER) such as activities and gestures has hardly been investigated. In this context, this paper addresses ZSER by relying on semantic attributes of events to transfer the learned knowledge from seen classes to unseen ones. First, we utilized the Amazon Mechanical Turk platform to create the first attribute-based gesture dataset, referred to as ZSGL, comprising the categories present in MSRC and Italian gesture datasets. Overall, our ZSGL dataset consisted of 26 categories, 65 discriminative attributes, and 16 attribute annotations and 400 examples per category. We used trainable recurrent networks and 3D CNNs to learn the spatio-temporal features. Next, we propose a simple yet effective <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">end-to-end</i> approach for ZSER, referred to as Joint Sequential Semantic Encoder (JSSE), to explore temporal patterns, to efficiently represent events in the latent space, and to simultaneously optimize for both the semantic and classification tasks. We evaluate our model on ZSGL and two action datasets (UCF and HMDB), and compared the performance of JSSE against several existing baselines in four experimental conditions: 1. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Within-category</i> , 2. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Across-category</i> , 3. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Closed-set</i> , and 4. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Open-Set</i> . Results show that JSSE considerably outperforms ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$p&lt; 0.05$</tex-math></inline-formula> ) other approaches and performs favorably for both the datasets in all experimental conditions.

Subscales of alexithymia show unique pathways through reappraisal and suppression to anxiety, depression and stress

The goal of this work was to explore associations of constituent factors of alexithymia on mental health and potential mediating effects of emotion regulation strategies, specifically suppression and reappraisal. Data were collected through the crowd-sourcing platform Amazon Mechanical Turk (MTURK). Three hundred seventy-seven individuals completed questionnaires related to distress (Depression Anxiety Stress Scales [DASS]), emotion regulation (Emotion Regulation Questionnaire [ERQ]) and Alexithymia (Bermond-Vorst Alexithymia Questionnaire [BVAQ]). Three mediation models were constructed for depression, anxiety and stress, with BVAQ subscales (verbalizing, identifying, emotionalizing, fantasizing, and analyzing) as predictors and ERQ subscales (suppression and reappraisal) as mediators. Results indicated 37.3 % variance in depression, 25.2 % variance in anxiety, and 35.3 % variance in stress was explained by each model. Direct associations revealed emotionalizing and fantasizing were negatively associated with depression, anxiety, and stress, while verbalizing was additionally associated with depression, identifying was additionally associated with anxiety, and all four BVAQ subscales were associated with stress. BVAQ subscales demonstrated negative associations with reappraisal and positive associations with suppression that mediated anxiety and depression. However, suppression did not mediate relationships between BVAQ subscales with stress. Findings support the importance of examining multiple factors of alexithymia and associations with emotion regulation strategies and distress.

Online crowdsourcing to study trauma and mental health symptoms in military populations: A case for Amazon's Mechanical Turk (MTurk) platform.

Amazon's Mechanical Turk (MTurk) appears to be a reliable resource for studying clinical populations and accessing hard-to-reach populations. Recent research suggests that MTurk may also be a viable option for military recruitment. The goal of the current study was to examine the utility of collecting clinical data on military samples recruited via MTurk. Participants were 535 military veterans (Mage = 37.45; 71.8% men; 69.5% White) who completed measures assessing trauma and mental health. Findings indicate that rates of military traumas and mental health diagnoses were higher than published comparisons; posttraumatic stress disorder (PTSD) and depression symptoms were found to be higher than values found in a nationally representative sample, lower than a treatment-seeking sample, and comparable to a MTurk-recruited military sample. Alcohol misuse was found to be higher than both nationally representative and treatment-seeking samples. Psychometric analyses indicated support for convergent validity of measures, and confirmatory factor analysis results demonstrated that empirically supported factor models of PTSD were replicated in the current sample; the hybrid model demonstrated the best fit. Our findings support the utility of MTurk for collecting clinical data on military samples. Increasing access to and recruitment of military samples is important for advancing the field of military psychology. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Investigating the role of artificial intelligence in the US criminal justice system

PurposeThe purpose of this study is to examine public perceptions and attitudes toward using artificial intelligence (AI) in the US criminal justice system.Design/methodology/approachThe authors took a quantitative approach and administered an online survey using the Amazon Mechanical Turk platform. The instrument was developed by integrating prior literature to create multiple scales for measuring public perceptions and attitudes.FindingsThe findings suggest that despite the various attempts, there are still significant perceptions of sociodemographic bias in the criminal justice system and technology alone cannot alleviate them. However, AI can assist judges in making fairer and more objective decisions by using triangulation – offering additional data points to offset individual biases.Social implicationsOther scholars can build upon the findings and extend the work to shed more light on some problems of growing concern for society – bias and inequality in criminal sentencing. AI can be a valuable tool to assist judges in the decision-making process by offering diverse viewpoints. Furthermore, the authors bridge the gap between the fields of technology and criminal justice and demonstrate how the two can be successfully integrated for the benefit of society.Originality/valueTo the best of the authors’ knowledge, this is among the first studies to examine a complex societal problem like the introduction of technology in a high-stakes environment – the US criminal justice system. Understanding how AI is perceived by society is necessary to develop more transparent and unbiased algorithms for assisting judges in making fair and equitable sentencing decisions. In addition, the authors developed and validated a new scale that can be used to further examine this novel approach to criminal sentencing in the future.

Retrospector: Rapid collaborative reflection to improve collaborative practices

Online platforms for freelancing allow teams performing complex work to be assembled in a matter of minutes and dispersed nearly as quickly. With such short time frames, ad hoc and virtual teams have few opportunities to learn strategies and effective team practices to work with their colleagues. Without such practices, teams are prone to work sub-optimally and lack direction. One key challenge in virtual teams discovering effective team practices is that because the practices ought to involve situated knowledge, it takes time to coalesce, as team members learn about each other over time. This work introduces Retrospector, that ad hoc teams can use to reflect collaboratively and reinforce effective team practices. Our interface accelerates the discovery of practices in situ and then guides them in reinforcing and applying these practices to future tasks. We conducted a between-subjects experiment (N=75) to assess our design with crowdworkers from the Amazon Mechanical Turk platform. This randomized controlled experiment showed that teams using our system for approximately six minutes of collaborative reflection were able to discover effective practices more successfully and had significantly improved team performance and viability. These results indicate that deliberate support for improving team practices can improve outcomes even through very short interaction. We conclude with design implications and opportunities for future work.

Access to medication for opioid use disorder supported by telemedicine and healthcare coverage: A web-based survey during the COVID-19 pandemic

Background and AimsMedications for opioid use disorder (MOUD) are highly effective in improving treatment outcomes and reducing overdose. Concerns about interrupted access to critical MOUD services led to expansion of telemedicine services during the COVID-19 pandemic in the US. The current study tested the hypothesis that telemedicine usage and healthcare coverage would be significantly associated with access to MOUD in the early phase of the COVID-19 pandemic. DesignA cross-sectional online survey was administered to a non-probability sample from June 18-July 19, 2020 using the Amazon Mechanical Turk platform. SettingNortheastern United States during the early phase of the COVID-19 pandemic. At the time of the survey, federal regulators had waived the longstanding requirement for in-office visits for MOUD prescription receipt and provided guidance on increasing third-party payer reimbursement rates for telehealth visits in order to mitigate barriers to care associated with COVID-19 safety guidelines. ParticipantsIndividuals 18 years or older residing in Connecticut, Massachusetts, New Jersey, New York, or Rhode Island were eligible to complete the survey. The analytic sample was participants who reported using opioids not as prescribed by a physician in the past seven days. MeasurementsDemographics, telemedicine usage, and healthcare coverage were assessed as explanatory variables. The primary outcome was whether participants reported ability to access MOUD in the past four weeks. FindingsIn this sample of individuals who used illicit opioids in the past week (N = 191), one in two individuals who utilized telehealth or had healthcare coverage were able to access MOUD, whereas only one in five of their respective counterparts who did not have telehealth access or healthcare coverage were able to access these medications. ConclusionsTelemedicine and healthcare coverage were associated with greater MOUD access early in the COVID-19 pandemic, when barriers to care were high. Such findings speak to the importance of not only extending but also formalizing temporary policy changes instituted during the pandemic to allow MOUD prescribing via telemedicine.

Impact of Beliefs about Local Physician Supply and Self-rated Health on Willingness to See a Nurse Practitioner during the COVID-19 Pandemic: Online Survey and Experiment.

The coronavirus (COVID-19) pandemic overburdened primary care clinicians. For nurse practitioners (NPs) to alleviate the burden, the public must be willing to see a NP over a physician. Those with poor health tended to continue seeking care during the pandemic, suggesting they may be willing to see a NP. Evaluate the public's willingness to see a NP for primary care and how this may be associated with their beliefs about the local supply of physicians and self-rated health. Conduct two studies: 1) a survey to identify correlations; and 2) an experiment to assess how willingness is dependent on information about the local supply of physicians. The survey and experiment were conducted online in April and December 2020, respectively. Participants were U.S. adults recruited from Amazon's Mechanical Turk platform. The key independent variables were self-rated health, which was a dichotomized five-point scale (Excellent, Very Good, Good vs. Fair, Poor), and beliefs about local physician supply. The survey measured beliefs about local physician supply, while the experiment manipulated beliefs via altering information the participants read about the local supply of physicians. Willingness to see a NP was assessed as an overall preference over a physician and as a preference given two clinically significant scenarios in which participants imagined they were experiencing either coughing or a headache (presentation order randomized). Multiple regressions and analysis of variance were used to assess how beliefs as about the local physician supply and self-rated health were associated with overall willingness to see a NP. Bivariate probits simultaneously estimated willingness to see a NP in the two clinically significant scenarios. The survey showed concerns about physician supply was associated with lower willingness to see a NP among respondents with comparatively better health, but a greater willingness among respondents with comparatively worse health. The experiment suggests only the latter is causal. For the two clinically significant scenarios, these patterns appeared for the coughing scenario in the survey and the headache scenario in the experiment. U.S. adults with comparatively worse self-rated health become more willing to see a NP for primary care when they hear information that raises their concerns about the local physician supply. The differences between the survey and experiment results may be useful for interpreting findings from future studies. Findings may aid in managing finite health care resources during public health crises and crafting successful messaging by NP advocacy groups. Efforts to address nursing shortages will also be needed.

Does Mood affect Sexual and Gender Discrimination in Hiring Choices? Evidence from Online Experiments

We explore whether there is a link between mood and hiring decisions. This research examines how positive mood affects the discrimination faced by homosexual and female job candidates compared to heterosexual and male ones. We randomly assign respondents to one of two mood-inducing videos (positive and neutral), and we allow subjects to make a series of hiring choices prior and immediately after watching the mood-inducing video. Our experiment being conducted in the online labor platform Amazon Mechanical Turk, allows us to track the complete hiring process and monitor employers’ behavior within and without our treatment context. Constructing pairs of curriculum vitae, distinguished only by the sexual orientation or the gender of the applicants in each case, leads to the observation that women and gay men faced a significantly lower chance of getting hired. We also find that female employers proposed higher levels of discrimination only in the case of female applicants. Our positive mood manipulation leads to a decrease of discrimination levels. Thus, there is substantial experimental evidence to suggest that discrimination based on sexual orientation and gender also exists in online labor markets. An additional experiment with negative mood manipulation, also, gives evidence for the opposite direction of the effects, contributing to a broader picture of the relationship between mood and discrimination behavior. Contributions to the literature on hiring discrimination, mood research and the online economy are discussed.

A Virtual Reading Center Model Using Crowdsourcing to Grade Photographs for Trachoma: Validation Study

As trachoma is eliminated, skilled field graders become less adept at correctly identifying active disease (trachomatous inflammation-follicular [TF]). Deciding if trachoma has been eliminated from a district or if treatment strategies need to be continued or reinstated is of critical public health importance. Telemedicine solutions require both connectivity, which can be poor in the resource-limited regions of the world in which trachoma occurs, and accurate grading of the images. Our purpose was to develop and validate a cloud-based "virtual reading center" (VRC) model using crowdsourcing for image interpretation. The Amazon Mechanical Turk (AMT) platform was used to recruit lay graders to interpret 2299 gradable images from a prior field trial of a smartphone-based camera system. Each image received 7 grades for US $0.05 per grade in this VRC. The resultant data set was divided into training and test sets to internally validate the VRC. In the training set, crowdsourcing scores were summed, and the optimal raw score cutoff was chosen to optimize kappa agreement and the resulting prevalence of TF. The best method was then applied to the test set, and the sensitivity, specificity, kappa, and TF prevalence were calculated. In this trial, over 16,000 grades were rendered in just over 60 minutes for US $1098 including AMT fees. After choosing an AMT raw score cut point to optimize kappa near the World Health Organization (WHO)-endorsed level of 0.7 (with a simulated 40% prevalence TF), crowdsourcing was 95% sensitive and 87% specific for TF in the training set with a kappa of 0.797. All 196 crowdsourced-positive images received a skilled overread to mimic a tiered reading center and specificity improved to 99%, while sensitivity remained above 78%. Kappa for the entire sample improved from 0.162 to 0.685 with overreads, and the skilled grader burden was reduced by over 80%. This tiered VRC model was then applied to the test set and produced a sensitivity of 99% and a specificity of 76% with a kappa of 0.775 in the entire set. The prevalence estimated by the VRC was 2.70% (95% CI 1.84%-3.80%) compared to the ground truth prevalence of 2.87% (95% CI 1.98%-4.01%). A VRC model using crowdsourcing as a first pass with skilled grading of positive images was able to identify TF rapidly and accurately in a low prevalence setting. The findings from this study support further validation of a VRC and crowdsourcing for image grading and estimation of trachoma prevalence from field-acquired images, although further prospective field testing is required to determine if diagnostic characteristics are acceptable in real-world surveys with a low prevalence of the disease.

The path to good health: Shifting the dialogue and promoting social ecological thinking

The purpose of our research was to assess whether an intervention could produce changes in critical consciousness (CC) relative to participants' understanding of social influences on health and individual health behaviors. The intervention was a 4-min animation, entitled The Path to Good Health , that described how factors in our social environment influence individuals in a variety of ways and thereby our health. We used the same sampling and intervention strategies with two discrete cohorts of participants (Initial study: June 2018, n = 249; Retest study: October 2019, n = 315), who were recruited and incentivized through Amazon's Mechanical Turk platform. We measured direction and extent of change from pre-to post-intervention in four key constructs of CC (Passive Adaptation, Emotional Engagement, Cognitive Awakening, and Intentions to Act) using the 4-Factor Critical Consciousness Scale (4-FCCS), and we investigated differential effects of the intervention related to demographics of participants, including political typology. We also assessed concurrent and predictive validity of the (4-FCCS). Changes of CC subscale scores from pretest to posttest moved in the expected direction in both the Initial and Retest studies; Cohen's d effect sizes were medium to very large in both studies. Overall, findings suggest the video intervention was effective at improving CC among participants drawn from the general population. We demonstrated that it is possible to influence people's cognitive-emotional interpretations in as little as 4 min, regardless of one's political ideology, and that the (4-FCCS) is sufficiently sensitive to measure changes in CC. This study provides preliminary evidence that a brief intervention can broaden cognitive-emotional interpretations from an over-emphasis on personal responsibility for individual health to an appreciation of the impact of social ecological factors on population health.

Algorithms of Suspicion: Authentication and Distrust on the Amazon Mechanical Turk Platform

Algorithms of Suspicion: Authentication and Distrust on the Amazon Mechanical Turk Platform

Toloka platform as a source of online survey participants: an experience of assessing data quality

The article presents the experience of using Yandex Toloka crowdsourcing platform to recruit respondents for an online survey. Analyzing methodological publications on a similar foreign platform Amazon Mechanical Turk we put forward hypotheses about the data quality obtained via Toloka in comparison with the results collected using other convenience sample types –online panels and recruitment of respondents through social networks. Additionally, only based on the Toloka data, we assessed the indicator of respondent’s attentiveness. The main conclusion is that Toloka allows to recruit respondents quickly and at low cost, and the results are comparable in terms of quality to those obtained by other methods. In particular, respondents from Toloka almost always complete the survey, fill out questionnaires faster than other types of respondents, but less often than participants of the online panel have a tendency to “straightline” (i.e., give the same answers in a tabular question), just as often as social media respondents give answers to the open-ended question (but less frequently than online panel participants), although their responses are shorter. Only 36% of the respondents passed the attention check question, attentive participants had a longer questionnaire complete time and were less likely to be straightliners. The increase of reward did not increase the proportion of attentive respondents, but decreased the questionnaire filling out speed, increased the number of answers to the open question, and reduced the proportion of straightliners.

Organic food consumption and contextual factors: An attitude–behavior–context perspective

AbstractThis study is grounded within the theoretical framework of attitude–behavior–context. The study examines how perceived health risk shapes both health and environmental attitudes, leading to organic food consumption. Further, we also assess the role of perceived effectiveness and green trust in moderating the proposed relationships. The data were collected from 461 consumers through Amazon Mechanical Turk platform. Our findings indicate that consumers' perception of health risk plays a significant role in developing positive environmental and health attitudes. Consequently, this leads consumers to adopt organic food. Further, this study confirms that there is an attitude–behavior gap, as both environmental and health attitudes seem to have a weak Cohen f2 effect on organic food consumption. Though perceived effectiveness moderates the relationship of organic food consumption with perceived health risk and attitude, green trust does not bridge the attitude–behavior gap. This study may help organic food marketers, policy‐makers, and retailers bridge the consumers' attitude–behavior gap.

Understanding the Microtask Crowdsourcing Experience for Workers with Disabilities: A Comparative View

Microtask crowdsourcing holds great potential as an employment opportunity with the flexibility and anonymity that individuals with disability may require. Though prior research has explored the accessibility of crowd work, the lived crowd work experiences of the broader community of workers with disability are still largely under-explored, especially when it comes to how their experiences are similar to or different from the experiences of workers without disability. In this work, we aim to obtain a deeper understanding of the microtask crowdsourcing experience for people with disabilities, especially regarding their financial and social experiences of participating in crowd work, along with the benefits and challenges that they encounter through this work. Specifically, we first surveyed 1,200 crowd workers both with and without disability about their experiences using the Amazon Mechanical Turk platform, and the differences we found inspired the design of a follow-up survey to gain greater understanding of the crowd work experience for workers with disability. Our findings reveal that workers with disability receive unique benefits from performing crowd work, such as a greater sense of purpose, but also encounter many challenges, such as completing tasks on time and earning a livable wage, causing them to turn to online communities for assistance. Although many of the challenges they face are not unique to crowd workers with disability, workers with disability may be disproportionately impacted by these challenges. From our findings, we provide implications for crowd platforms, as well as the gig economy as a whole, that seek to promote greater consideration of workers with a diverse range of conditions to create a more valuable work experience for them.

Usability of Hospital Price Estimators for Lumbar Spine MRI.

The aim of this study was to evaluate the usability of online hospital price estimators for a common imaging examination using surrogate patients. Using the Amazon Mechanical Turk platform, the authors recruited adult English-speaking US residents as surrogate patients to find the cash price for a noncontrast lumbar spine MRI examination for a self-pay patient using price estimator tools at four hospitals. All were asked to view a 3-min tutorial video and report their experiences with the task, including the System Usability Scale (SUS) for the estimator, through a paid survey. Participants were queried about demographics, insurance, prior imaging exposure, and assessed health literacy and health insurance literacy using validated measures. Multivariable analysis for correct price identification and price estimator SUS were performed. Of 660 respondents, 476 met eligibility criteria (72.7% <45 years of age, 41.7% female, 86.1% white); 76.9% found all four estimators, and 9.2% were unable to locate any. Only 27.7% found the correct price at all four hospitals, with 67.4% being able to find at least one correct cash price. Average SUS score for the hospitals' estimators ranged between 62.4 and 77.5. The hospital with a similar estimator to that used in the tutorial video had the highest SUS score. Accuracy of price identification improved with later tasks. Higher health insurance literacy was associated with higher identification of at least one correct price (odds ratio, 1.21; 95% confidence interval, 1.02-1.44) and higher SUS score (B= 1.68; 95% confidence interval, 1.07-2.29). Surrogate patients were able to locate hospital price estimators but unable to effectively use them to obtain correct prices. Tutorial videos improved SUS score, but correct price identification improved with practice.