Scientists' data sharing practices and perceptions were studied through a survey of 1329 researchers. The study found that while most scientists are satisfied with their current data collection and short-term storage practices, they are not satisfied with long-term data preservation. Many organizations lack support for data management in both the short and long term. Scientists are more likely to share data if certain conditions are met, such as formal citation and sharing reprints. There are significant differences in data management practices based on primary funding agency, subject discipline, age, work focus, and world region. Data sharing is important for scientific progress, but many researchers do not share their data due to reasons such as insufficient time, lack of funding, and concerns about privacy or future publishing opportunities. The study also found that data sharing practices vary by discipline, with some fields being more open than others. Researchers who choose to withhold data often have specific reasons, such as the desire to retain exclusive rights or concerns about data misuse. Institutional policies and practices significantly influence data sharing. Many researchers reported that their organizations lack formal policies for data management and preservation. The study suggests that new data management mandates from agencies like the NSF and global attention to data sharing could lead to changes. Large-scale programs like DataONE aim to improve data sharing by providing resources and infrastructure for data management and open access. The study also highlights the importance of metadata standards and the need for better tools and education to improve data management practices. Researchers cited many reasons for not sharing data electronically, including insufficient time and lack of funding. However, some researchers believe that data sharing is important and would be willing to share data if certain conditions are met. Overall, the study shows that while data sharing is important for scientific progress, there are significant barriers to effective data sharing, including institutional policies, lack of funding, and concerns about data privacy and misuse. The study suggests that new data management mandates and global attention to data sharing could help overcome these barriers and improve data sharing practices.Scientists' data sharing practices and perceptions were studied through a survey of 1329 researchers. The study found that while most scientists are satisfied with their current data collection and short-term storage practices, they are not satisfied with long-term data preservation. Many organizations lack support for data management in both the short and long term. Scientists are more likely to share data if certain conditions are met, such as formal citation and sharing reprints. There are significant differences in data management practices based on primary funding agency, subject discipline, age, work focus, and world region. Data sharing is important for scientific progress, but many researchers do not share their data due to reasons such as insufficient time, lack of funding, and concerns about privacy or future publishing opportunities. The study also found that data sharing practices vary by discipline, with some fields being more open than others. Researchers who choose to withhold data often have specific reasons, such as the desire to retain exclusive rights or concerns about data misuse. Institutional policies and practices significantly influence data sharing. Many researchers reported that their organizations lack formal policies for data management and preservation. The study suggests that new data management mandates from agencies like the NSF and global attention to data sharing could lead to changes. Large-scale programs like DataONE aim to improve data sharing by providing resources and infrastructure for data management and open access. The study also highlights the importance of metadata standards and the need for better tools and education to improve data management practices. Researchers cited many reasons for not sharing data electronically, including insufficient time and lack of funding. However, some researchers believe that data sharing is important and would be willing to share data if certain conditions are met. Overall, the study shows that while data sharing is important for scientific progress, there are significant barriers to effective data sharing, including institutional policies, lack of funding, and concerns about data privacy and misuse. The study suggests that new data management mandates and global attention to data sharing could help overcome these barriers and improve data sharing practices.