Abstract
Beef, pork and broiler producers face complexities in prioritizing sustainability actions. We use Q methodology to explore diverging patterns in how US producers approach sustainability priorities, even among same-species and same-scale operations, with emphasis on animal well-being and environmental stewardship. Three distinct worldviews emerged: one prioritizing animal welfare, another emphasizing holistic environmental stewardship, and a third focused on balancing business viability with other concerns. Our findings can inform strategies for aligning producer actions and consumer expectations for sustainable meat production.
Subject terms: Environmental impact, Sociology, Environmental studies, Business and management, Business and management
A sample of US producers’ worldviews on sustainable meat production fall into three categories, which can guide decision-makers towards actions with broad support and highlight those requiring more targeted efforts.
Main
Decisions about meat production practices and priorities are central within a complex landscape of risks, benefits and trade-offs spanning social, economic and environmental dimensions of sustainability1,2. Such decisions are made by millions of farmers globally, representing myriad perspectives3,4. Producer perspectives are often examined in the context of farm characteristics (for example, size, farming system), but less is known about producers’ underlying viewpoints or worldviews, how these influence decision-making, and whether they align with farm characteristics5–8. In this study, we present a novel application of Q methodology to explore how US beef, pork and broiler producers prioritize competing sustainability considerations. We employ a rank-ordering exercise and principal component analysis to examine consensus and divergence around common worldviews, paired with interviews that contextualize the results9,10. This approach was selected because it enables systematic exploration of subjectivity, focusing on the range of perspectives that exist, and the forced-choice nature of the ranking exercise is ideally suited to the competing priorities and complexity of sustainability decision-making. We emphasize animal well-being and environmental impact because they receive significant attention among both producers and consumers11–13 and thus hold potential as areas in which improved understanding could enable better-aligned signals across supply chains to enhance sector sustainability. We focus on the United States because of the country’s high levels of meat consumption and production and diverse scales of operation14,15.
Results
Producers clustered into three distinct worldviews, or discourses, based on the way they sorted sustainability priorities (Table 1 and Fig. 1).
Table 1.
Factor characteristics
| Discourse | ||||
|---|---|---|---|---|
| Animal husbandry, first and foremost (AH) | Environmental stewardship, a holistic approach (ES) | Business viability, a balancing act (BV) | ||
| Loading Q sorts (count (%))a | 15 (43%) | 13 (37%) | 7 (20%) | |
| Explained variance (%) | 21 | 17 | 9 | |
| Eigenvalue | 7.9 | 6.5 | 4.2 | |
| Loading Q sorts by species (count (%))b | Beef | 6 (86%) | 1 (14%) | 0 (0%) |
| Pork | 4 (40%) | 1 (10%) | 5 (50%) | |
| Broilers | 1 (25%) | 2 (50%) | 1 (25%) | |
| Multiple | 4 (29%) | 9 (64%) | 1 (7%) | |
| Loading Q sorts by scale (count (%))c | Very small | 3 (33%) | 6 (67%) | 0 (0%) |
| Small | 8 (47%) | 7 (41%) | 2 (12%) | |
| Medium | 3 (43%) | 4 (57%) | 0 (0%) | |
| Large | 2 (67%) | 0 (0%) | 1 (33%) | |
| Very large | 0 (0%) | 0 (0%) | 4 (100%) | |
aDistribution of participant Q sorts loading to each factor (n = 35, with 3 of the 38 having been confounded)
bBeef, pork and broiler rows include only single-species operations. Any operations raising multiple species are included in the ‘multiple’ row.
cCounts add up to more than 35 because some participants raised multiple species, sometimes at different scales. For example, a participant raising beef at a small scale and pork at a medium scale will be counted in both the small- and medium-scale rows. See Supplementary Table 2 for scale definitions.
Fig. 1. Sustainability priorities by discourse.
a, Sustainability statement factor scores. Statements are listed in the order in which they were displayed to participants. Factor scores for each statement are listed by discourse (AH, animal husbandry, first and foremost; ES, environmental stewardship, a holistic approach; BV, business viability, a balancing act). Factor scores indicate placement along the Q sort grid; for example, 0 indicates placement in the middle, 6 indicates placement in the far-right column (highest priority), and −6 indicates placement in the far-left column (lowest priority). Higher intensity of shading corresponds to more positive factor scores. b, Priority statements grouped by the discourse in which they received the highest factor score. c, Priority statements grouped by the discourse in which they received the lowest factor score. Each box in b,c represents a statement, with numbers corresponding to statement numbers. Statements in overlapping areas received the same factor score in one or more discourse. In a–c, sustainability domains with which the statements most closely align are indicated by colours/icons (pink/pig, animal well-being, distinct from other social considerations; beige/people, social considerations other than animal well-being; green/leaf, environmental; yellow/calculator, economic).
The ‘animal husbandry, first and foremost’ discourse (AH) was characterized by prioritizing the humane treatment and well-being of animals above other considerations. The operations of participants belonging to this group ranged in size, and most raised a single species (Table 1). Members of this discourse ranked the majority of statements about animal handling and well-being higher than did members of other discourses (Fig. 1b), with such statements comprising the seven highest priorities of AH (Fig. 1a). Compared with other discourses, AH placed less emphasis on statements about environmental, economic and broader societal considerations (Fig. 1c), and noted their unwillingness to compromise principles related to animal well-being, often tying their decision-making to personal values. However, members of this group also viewed the prioritization of animal well-being as making practical business sense, describing happy, healthy animals as essential to or synergistic with other goals such as a high-quality product, strong customer relationships, profitability and resource stewardship (Extended Data Table 1).
Extended Data Table 1.
Common sentiments by discourse, with illustrative quotes, organized into overarching categories
The ‘environmental stewardship, a holistic approach’ discourse (ES) was characterized by prioritizing environmental stewardship, viewing the farm or ranch as an integrated ecosystem achieving multiple goals simultaneously. Unlike with AH, most participants belonging to the ES discourse raised multiple species and operated at smaller scales (Table 1). Members of this discourse tended to rank statements related to environmental sustainability as a higher priority than did members of other groups (Fig. 1b). However, environmental sustainability does not fully define this group’s focus because their next-most highly ranked priorities included those related to animal health, producing nutritious food and building trust with consumers (Fig. 1a). Compared with other discourses, ES placed less emphasis on statements related to efficiency and technology (Fig. 1c), instead noting the longer finishing times and more natural conditions of the pasture-based systems that predominated among this group. Animals were described as integral components of healthy agroecosystems, promoting resource use efficiency, circularity and soil health, and contributing fertility to crop-based systems. While expressing concern about the affordability of their products, ES members stressed that remaining financially viable was essential to enabling their holistic stewardship approach (Extended Data Table 1).
The ‘business viability, a balancing act’ discourse (BV) was characterized by seeking to balance a mix of priorities related to production efficiency, resource stewardship and animal health with the overarching goal of operating a viable business producing nutritious food. The majority of BV members were larger-scale pork operations, but this discourse also included small-scale broiler and multispecies operations (Table 1). The group’s most highly ranked statement, ‘providing nutritious food to consumers’ (S7), sets it apart, while other highly ranked statements were shared with other discourses (for example, natural resource stewardship and consumer trust priorities shared with ES, prioritization of animal health shared with AH and ES). However, members of the BV discourse tended to rank statements with direct bearing on business operations and economic viability more highly than did members of the other discourses (Fig. 1), making this a defining feature of the group. Statements ranked lowest overall by this group (S21 and S24) related to giving animals more space, and deprioritization was explained as primarily for animal health and comfort (for example, protecting pigs from harsh winters) and because additional space in confinement systems can lead to injury. However, two members deprioritized these statements because they did not operate confinement-based systems and thus considered them non-issues. A common BV viewpoint was that business profitability is a prerequisite for all other objectives, and that activities such as trust-building, regulatory compliance and stewardship are all existential necessities (Extended Data Table 1). The BV discourse also emphasized efficiency as a core principle, ranking several environmental- and labour-related efficiency statements more highly than other discourses (Fig. 1b), noting the importance of efficiency to achieving environmental and economic sustainability and underscoring the importance of animal health to efficient production (Extended Data Table 1).
In addition to revealing divergent views, the Q sort also revealed areas of alignment among the discourses. Three statements, each ranked moderately by participants, were identified as statistical points of consensus (Extended Data Fig. 1): ‘minimizing water pollution’ (S23) was not considered a high priority because other practices, such as managing soil health or overall resource stewardship, accounted for it. ‘Managing manure effectively and efficiently’ (S20) was not a high priority because participants felt they had already addressed it in their operations. ‘Increasing consumer understanding of farming operations’ (S33) was rated a moderate priority across groups but for different reasons: AH and ES discussed wanting to showcase their animal husbandry practices and high-quality sustainable products, respectively, and explain the associated high costs, while BV emphasized raising awareness of their role in producing affordable, nutritionally dense food.
Extended Data Fig. 1. Sustainability priority statement z-scores by discourse.
Points are coloured by discourse (AH = animal husbandry, first and foremost; ES = environmental stewardship, a holistic approach; BV = business viability, a balancing act). A higher z-score indicates a higher priority. Clustering of points indicates consensus across discourses about the relative priority of that statement. Greater distance between points indicates greater disagreement between discourses. Filled symbols indicate distinguishing statements (for which the z-score for one discourse was significantly different from the z-scores for the other two discourses). Statements are ordered from the most distinguishing (top) to most consensus (bottom).
‘Maintaining traditions in farming practices’ (S10) stands out as the most consistently low priority across discourses (Fig. 1a and Extended Data Fig. 1). Although some participants noted their respect for older generations, traditional knowledge and rural cultural values, there was agreement that maintaining traditions impeded progress and sustainability. Intriguingly, no similar consensus emerged around a high priority. The ES and BV discourses aligned in their prioritization of several statements related to consumer relations and stewardship (S7, S8, S34; Extended Data Fig. 1), but these priorities were not shared with AH. Priorities shared between AH and either ES or BV tended to receive somewhat lower overall rankings, such as ‘providing optimal rations to animals’ (S26, shared with BV) and ‘building or expanding local, direct-to-consumer markets’ (S14, shared with ES). Statements that were statistically distinguishing between all three discourses included several related to animal husbandry (S4, S18, S21, S28), environmental stewardship (S3, S27) and the workforce (S25). Statements that were distinguishing for only one discourse were equally revealing, such as ‘improving production and profitability by using new technologies’ (S37), which strongly distinguished BV from AH and ES, or ‘providing affordable products to consumers’ (S6), which strongly distinguishes AH from ES and BV (Extended Data Fig. 1).
Across discourses, producers’ explanations of what drove their prioritizations fell into several categories. Operational viability was discussed by members of all discourses as always top-of-mind and something that frequently drove trade-offs (for example, between economic and environmental priorities, see Extended Data Table 1). Personal drivers strongly influenced the priority-setting of most participants, except for the four very large-scale producers belonging to the BV discourse, and customer-related drivers were also mentioned by many and drove additional trade-offs. For example, the prioritization of animal well-being, which was sometimes ascribed to personal values and sometimes to customer preferences, was often noted as taking precedence over profit. Regulatory drivers were described as less influential except for those operating at very large scales, and were typically viewed as constraints leading to further trade-offs and compromises among competing priorities.
Discussion
Farmers and ranchers face a complex task as they work to improve the sustainability of meat production. Agricultural sustainability encompasses everything from environmental stewardship to economic prosperity to rural and societal quality of life16, and in animal agriculture additional considerations such as animal welfare and the affordability of high-quality animal-sourced food must be balanced. It is readily apparent that few solutions exist that will simultaneously maximize all possible sustainability priorities: most have trade-offs, and producers must navigate these. The forced-choice nature of Q methodology is ideally suited to examining prioritization and perspectives in this setting. Previous studies have used Q methodology to examine discrete elements of sustainability (for example, environmental stewardship or animal welfare) and have found an illuminating diversity of worldviews on these subjects8,17,18, yet to our knowledge none have combined the many competing facets of animal agriculture sustainability in a single study with a highly diverse sample of producers.
Studies of farmer perspectives frequently reveal distinctions between those who are primarily environmentally versus economically oriented17,19,20. However, the emergence of a discourse oriented strongly towards animal health, welfare and husbandry speaks to the importance of these unique sustainability considerations for many animal agriculture producers, and stands in contrast to worldviews identified by previous studies not primarily focused on animal agriculture16,19. It is also noteworthy that in this study the BV discourse was characterized by prioritizing not only economic but also the majority of social sustainability priorities not related to animals (Fig. 1b). The motivations for prioritization choices articulated by participants align with other research on this topic21–23. Willock et al.24 categorized non-financial influences that may affect decision-making among producers as a mix of attitudes (for example, stress, risk aversion, off-farm work), goals (for example, quality of life, management goals) and behaviours (for example, profit maximization, diversification)—influences that were also found in the present study.
While some loose patterns emerged related to operation type, this may have been due to the nature of the questions (Supplementary Discussion 2), and it is clear from the diverse make-up of the three discourses identified here that worldview is dictated by far more than operational characteristics. Efforts aimed at incentivizing sustainable practices often segment producers by scale, species and other descriptive characteristics, potentially obscuring important variations by worldview25,26. As used here, Q methodology can illuminate such variations within homogeneous-appearing segments and identify non-obvious points of convergence and divergence. By carefully analysing priorities across groups, policymakers and other stakeholders can identify areas in which broad support is likely and where more targeted or collaborative efforts are needed to advance sustainability goals. For example, in developing an outreach campaign on animal welfare, one could look across statements related to animal well-being to assess which strategies might be met with full or partial consensus and which areas are low, medium and high priorities for different groups (Supplementary Discussion 1). This approach highlights framings that might be compelling across groups and strategies that might be successful with some groups but not others.
Findings can also be used to examine a specific subgroup of issues within a sustainability domain. For example, those wishing to assess the feasibility of action on topics related to resource use and energy efficiency might observe that these statements (S3, S5, S15, S22, S23, S31) were all ranked as low priorities except for S5 (‘increasing production in resource-efficient ways’). This suggests that focusing on increasing resource-efficient production might be more compelling than other metrics not directly tied to productivity. One might also observe that many of these statements (such as S22 on renewable energy) were ranked especially low by AH, indicating that attempts to incentivize action in these areas may be less likely to succeed among members of this group, particularly when such actions could force a trade-off with other priorities.
The absence of any high-priority consensus statements is striking, although perhaps not surprising because the discourses are defined by their differing patterns of prioritizing competing considerations. This suggests that silver bullet sustainability solutions—those that will be equally compelling to all producers—are generally unlikely. However, although not a statistical point of consensus, ‘ensuring the health of animals’ (S32) stands out as having the highest average factor score across discourses. This points to prioritization of animal health as an important area of common ground across producers of differing worldviews, and suggests that measures that improve animal health may be particularly likely to see widespread support.
We find that animal agriculture producers have diverging patterns in the ways they approach the complex landscape of sustainability priorities, even among same-species and same-scale operations. The three discourses identified here reveal distinct overarching mindsets: animal-focused, environment-focused or business-focused. Understanding the nature of these distinct worldviews can help policymakers and stakeholders determine which sustainability priorities are likely to have broad support and which may require more targeted efforts. These findings also advance our understanding of motivations unique to animal agriculture producers. Similar research into how others in the value chain, including consumers, rank sustainability priorities and the degree to which they align with producer discourses and motivations is needed to create meat production systems that synergize across forces of supply and demand to advance sustainability.
Methods
This study was reviewed by the University of Washington Human Subjects Division and determined to qualify for exempt status.
Q methodology combines elements of both quantitative and qualitative research to explore human subjectivity around a set of issues or topics9. Using a rank-ordering exercise and principal components analysis paired with qualitative analysis of participant interviews, the method allows within-group and between-group exploration of consensus and divergence around common discourses (that is, worldviews or perspectives)10. Q methodology involves seven core steps: developing a concourse (a collection of statements representing the scope of dialogue and perceptions around a topic), refining the concourse into a Q set (a collection of statements to be sorted), recruiting a P set (participants), conducting Q sorts (in which participants sort the Q set), conducting post-Q sort interviews, performing quantitative analysis and performing qualitative interpretation27. Each of these as applied in the present study is described in detail in the Supplementary Information.
In brief, we developed a concourse based on key informant interviews with 14 stakeholders across beef, pork and broiler industries. We systematically winnowed and refined excerpts from this concourse into a final Q set of 38 statements encompassing a breadth of sustainability priorities. We conducted Q sorting exercises and interviews with 42 US producers of beef, pork and/or broilers across geographies and scales, and retained 38 of those interviews for analysis. Three factors, representing three distinct worldviews, were extracted from the Q sort dataset based on principal component analysis. We then conducted a qualitative interpretation of factors and decision-making drivers based on the factors’ highest- and lowest-ranked statements and undertook qualitative coding and analysis of interview transcripts.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Supplementary information
Supplementary Discussion, Methods, References, Figs. 1 and 2, and Tables 1 and 2.
Source data
Raw Q sort data.
Statement z-scores and statistical source data.
Acknowledgements
We thank the producers who provided their insights through Q sorts and interviews, to H. McKinley who assisted with transcript processing, and to the project’s stakeholder advisory committee. This study was supported by Agriculture and Food Research Initiative award number 2022-68006-37269 (S.M.C., J.J.O.) from the USDA National Institute for Food and Agriculture.
Extended data
Author contributions
J.J.O., S.M.C. and M.L.S. designed the study with input from J.S. and N.J. N.J., M.F., J.R.-G. and E.A. collected and analysed the data with assistance from A.I. S.M.C. prepared the paper with contributions from N.J., M.F., A.I., J.J.O. and M.L.S. All authors reviewed and approved the submitted paper.
Peer review
Peer review information
Nature Food thanks Lorraine Balaine, Nadine Lehrer, Elizabeth Ransom and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
Data availability
Raw data underlying quantitative analyses have been made available in the associated source data files. The participant key and interview data that support qualitative findings are not openly available due to reasons of sensitivity and are available from the corresponding authors upon reasonable request. Requests must include a description of intended use, agreement not to distribute or make data public, and a plan for protecting participant confidentiality. Data requests will be responded to within one month of their receipt. Data are located in controlled-access data storage maintained by the University of Washington. Source data are provided with this paper.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Sarah M. Collier, Email: scollier@uw.edu
Marie L. Spiker, Email: mspiker@uw.edu
Jennifer J. Otten, Email: jotten@uw.edu
Extended data
is available for this paper at 10.1038/s43016-026-01300-9.
Supplementary information
The online version contains supplementary material available at 10.1038/s43016-026-01300-9.
References
- 1.Breure, T. et al. A systematic review of the methodology of trade-off analysis in agriculture. Nat. Food5, 211–220 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Herrero, M., Thornton, P. K., Gerber, P. & Reid, R. S. Livestock, livelihoods and the environment: understanding the trade-offs. Curr. Opin. Environ. Sustain.1, 111–120 (2009). [Google Scholar]
- 3.Roche, L. M. et al. Sustaining working rangelands: insights from rancher decision making. Rangel. Ecol. Manag.68, 383–389 (2015). [Google Scholar]
- 4.Baur, P. in Social Innovation and Sustainability Transition (eds Desa, G. & Jia, X.) 241–260 (Springer, 2022); 10.1007/978-3-031-18560-1_17
- 5.McKendree, M. G. S., Tonsor, G. T. & Wolf, C. A. Animal welfare perceptions of the US public and cow-calf producers. J. Agric. Appl. Econ.50, 544–578 (2018). [Google Scholar]
- 6.Kauppinen, T., Vainio, A., Valros, A., Rita, H. & Vesala, K. Improving animal welfare: qualitative and quantitative methodology in the study of farmers’ attitudes. Anim. Welf.19, 523–536 (2010). [Google Scholar]
- 7.Pereira, M. A., Fairweather, J. R., Woodford, K. B. & Nuthall, P. L. Assessing the diversity of values and goals amongst Brazilian commercial-scale progressive beef farmers using Q-methodology. Agric. Syst.144, 1–8 (2016). [Google Scholar]
- 8.Walder, P. & Kantelhardt, J. The environmental behaviour of farmers—capturing the diversity of perspectives with a Q methodological approach. Ecol. Econ.143, 55–63 (2018). [Google Scholar]
- 9.Stephenson, W. The Study of Behavior: Q-technique and its Methodology (University of Chicago, 1953).
- 10.McKeown, B. B. & Thomas, D. Q Methodology—Quantitative Applications in the Social Sciences (Sage, 1988).
- 11.Schmiess, J. S. & Lusk, J. L. Trade-off between animal welfare and environmental impacts of beef production: an analysis of presentation effects on consumer choice. J. Agric. Resour. Econ.47, 278–299 (2022). [Google Scholar]
- 12.Johnston, J. L., Fanzo, J. C. & Cogill, B. Understanding sustainable diets: a descriptive analysis of the determinants and processes that influence diets and their impact on health, food security, and environmental sustainability. Adv. Nutr.5, 418–429 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Perry, B. D. & Grace, D. C. How growing complexity of consumer choices and drivers of consumption behaviour affect demand for animal source foods. EcoHealth12, 703–712 (2015). [DOI] [PubMed] [Google Scholar]
- 14.FAOSTAT. FAO Statistical Databases (FAO, 2024); http://www.fao.org/faostat/en/
- 15.2022 Census of Agriculture: Summary and State Data, Vol. 1, Geographic Area Series, Part 51 (USDA, 2024); https://www.nass.usda.gov/Publications/AgCensus/2022/Full_Report/Volume_1,_Chapter_1_US/usv1.pdf
- 16.Food, Agriculture, Conservation, and Trade Act of 1990 (USC, 1990); https://nationalaglawcenter.org/wp-content/uploads/assets/farmbills/1990-1.pdf
- 17.Davies, B. B. & Hodge, I. D. Exploring environmental perspectives in lowland agriculture: a Q methodology study in East Anglia, UK. Ecol. Econ.61, 323–333 (2007). [Google Scholar]
- 18.Maurer, L., Schenkenfelder, J. & Winckler, C. Resource, collaborator, or individual cow? Applying Q methodology to investigate Austrian farmers’ viewpoints on motivational aspects of improving animal welfare. Front. Vet. Sci.7, 607925 (2021). [DOI] [PMC free article] [PubMed]
- 19.Sutherland, L. A., Barnes, A., McCrum, G., Blackstock, K. & Toma, L. Towards a cross-sectoral analysis of land use decision-making in Scotland. Landsc. Urban Plan.100, 1–10 (2011). [Google Scholar]
- 20.Brodt, S., Klonsky, K. & Tourte, L. Farmer goals and management styles: implications for advancing biologically based agriculture. Agric. Syst.89, 90–105 (2006). [Google Scholar]
- 21.Balzani, A. & Hanlon, A. Factors that influence farmers’ views on farm animal welfare: a semi-systematic review and thematic analysis. Animals10, 1524 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hayden, M. Sensemaking and the influencing factors on farmer decision-making. J. Rural Stud.84, 31–44 (2021). [Google Scholar]
- 23.Brown, C. et al. Simplistic understandings of farmer motivations could undermine the environmental potential of the common agricultural policy. Land Use Policy101, 105136 (2021). [Google Scholar]
- 24.Willock, J. et al. Farmers’ attitudes, objectives, behaviors, and personality traits: the Edinburgh study of decision making on farms. J. Vocat. Behav.54, 5–36 (1999). [Google Scholar]
- 25.Huber, R. et al. Farm typologies for understanding farm systems and improving agricultural policy. Agric. Syst.213, 103800 (2024). [Google Scholar]
- 26.Piñeiro, V. et al. A scoping review on incentives for adoption of sustainable agricultural practices and their outcomes. Nat. Sustain.3, 809–820 (2020). [Google Scholar]
- 27.Brown, S. R. A primer on Q methodology. Operant Subjectivity16, 91–138 (1993). [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Discussion, Methods, References, Figs. 1 and 2, and Tables 1 and 2.
Raw Q sort data.
Statement z-scores and statistical source data.
Data Availability Statement
Raw data underlying quantitative analyses have been made available in the associated source data files. The participant key and interview data that support qualitative findings are not openly available due to reasons of sensitivity and are available from the corresponding authors upon reasonable request. Requests must include a description of intended use, agreement not to distribute or make data public, and a plan for protecting participant confidentiality. Data requests will be responded to within one month of their receipt. Data are located in controlled-access data storage maintained by the University of Washington. Source data are provided with this paper.



