The Ethics of Research Data Northeast Ethics Education Partnership Researchers spend much of their time collecting data. Data are used to confirm or reject hypotheses, to identify new areas of investigation, to guide
the development of new investigative techniques, and more. (8) Data Has Changed Technology has driven major changes in data management, analysis, and sharing -- and thus new ethical and legal considerations(1):
Data Has Changed Old model: lab notebooks, field notebook Old model: single researcher or small group (lab) Data Has Changed New: multiple forms of recording data, e.g.
computers, digital images, photos, gels, notes, printouts, etc. New: many researchers, groups, collaboration between different labs, disciplines, institutions, sectors (private vs. public) (1) Data Life Cycle Data Management Practices
Data management practices are becoming increasingly complex and should be addressed before any data are collected by taking into consideration four important issues (8): ownership collection storage sharing Data Ownership
Research produces data. Common sense might suggest that the person who conducts the research should own the productthe data. In fact, conditions imposed by funders, research institutions, and data sources may dictate otherwise (8). Data Ownership Funders. Funders provide support for research
for different reasons. Government is interested in improving the general health and welfare of society. Private companies are interested in profits, along with benefits to society. Philanthropic organizations are interested in advancing particular causes. Data Ownership These different interests translate into different
ownership claims. Typically: Government gives research institutions the right to use data collected with public funds as an incentive to put research to use for the public good Private companies seek to retain the right to the commercial use of data. Philanthropic organizations retain or give away ownership rights depending on their interests. Data Ownership
Since the claims of funders can and do vary considerably, researchers must be aware of their obligations to them before they begin collecting data. With government funding, it is important to distinguish between grants and contracts. Under grants, researchers must carry out the research as planned and submit reports, but control of the data remains with the institution that received the funds.
Contract Requirements Contracts require the researcher to deliver a product or service, which is then usually owned and controlled by the government. If your research is supported with government funds, make sure you know whether you are working under a grant or a contract. Data Ownership
Research institutions (universities). Support for research is typically awarded to research institutions (universities), not to individual researchers. As the recipients of research funds, research institutions have responsibilities for budgets, regulatory compliance, contractual obligations, and data management (8). Data Ownership
To assure that they are able to meet these responsibilities, research institutions claim ownership rights over data collected with funds given to the institution. This means that researchers cannot automatically assume that they can take their data with them if they move to another institution. The research institution that received the funds may have rights and obligations to retain control over the data.
Data Ownership Data sources (8). Increasingly research subjects and other entities that are the source of data are seeking some control over data derived from them. Countries with unique resources, such as tropical rain forests, individuals with rare medical conditions, and entities with unique databases, have at one time or another claimed ownership of research results based
on their data. Data Ownership Research subjects, place-based communities, Native sovereign nations, and indigenous peoples and entities may no longer be willing to provide data or be the source of data without some ownership stake in the end results. Tribal review committees are more common in Native nations and govern the use of research
data through various protocols. Data Ownership Well before any data are collected, ownership issues and the responsibilities that come with them need to be carefully worked out. Before undertaking any work, make sure you can answer the following questions: - Who owns the data I am collecting? - What rights do I have to publish the data?
- Does collecting these data impose any obligations on me? Data Collection There is no one best way to collect data. Different types of research call for different collection techniques. Four important considerations apply to all data collection to ensure the overall integrity of both the process and the information collected (8).
These considerations are: - Appropriate Methods Attention to Detail Authorized Recording Data Collection (8) Appropriate methods. Reliable data are vitally
dependent on reliable methods. Some funders require Quality Assurance/Quality Control (QAQC) as part of the grant process. Attention to detail. Quality research requires attention to detail. Experiments must be set up properly and the results accurately recorded, interpreted, and published. Data Collection Authorized. Many types of data collection
need to be authorized before they can proceed. Typically permission is needed to use: human and animal subjects in research hazardous materials and biological agents information contained in some libraries, databases, and archives Types of Authorized Data information posted on some Web sites
published photographs and other published information; and other copyrighted or patented processes or materials. Data Collection Recording The final step in data collection is the physical process of recording the data in some type of notebook (hard copy), computer file (electronic copy), or other
permanent record of the work done. The physical formats for recording data vary considerably, from measurements or observations to photographs or interview tapes. Data Collection Hard-copy evidence should be entered into a numbered, bound notebook so that there is no question later about the date the experiment was run, the order in which the data were collected, or
the results achieved. Do not use loose-leaf notebooks or simply collect pages of evidence in a file. Do not change records in a bound notebook without noting the date and reasons for the change. Electronic evidence should be validated in some way to assure that it was actually recorded on a particular date and not changed at some later date
If you collect your data electronically, you must be able to demonstrate that they are valid and have not been changed. Research Records Research process: protocols, Standard Operating Procedures (SOPs), applications for regulatory approval Research outcomes/products: reports, monographs
Research project management: contracts, invoices, staff records, funding applications and budgetary information Research Data: questionnaires, notes, photographs, samples, databases, recordings Research Protection Once collected, data must be properly protected. They may be needed later:
- to confirm research findings, - to establish priority, or - to be reanalyzed by other researchers (8). Over time, data, as the currency of research, become an investment in research. Data Protection Act If research involves live subjects, its subject to the 8 principles of the Data Protection Act of 1998 (3): 1. Personal data must be fairly and lawfully processed;
2. Personal data must be processed for limited purposes; 3. Personal data must be adequate, relevant and not excessive; 4. Personal data must be accurate; 5. Personal data must not be kept longer than necessary; 6. Personal data must be processed in accordance with the data subject's rights; 7. Personal data must be secure;
8. Personal data must not be transferred to countries without adequate protection Data Protection Act Informed Consent For processing to be deemed fair, subjects must be aware of: What will be done with the data? Who will hold the data? Who will have access to or receive copies of the data ? (if the data is to be shared with third parties,
this should be made explicitly clear)(3). Data Protection Act Informed Consent Additional requirements for these following sensitive data racial or ethnic origin political opinions religious beliefs trade union membership physical or mental health/condition
sexual life criminal offences or record (3) Ethical Issues in Data Management Prospective data collection
Explicit consent Awareness of data type Secure data storage Consider archiving and re-use of data Retrospective/existing data analysis Research using existing organizational data Individual knowledge of data usage Protecting Data
Participants should be informed on the nature of study and which data are going to be used: Be aware of data protection legislation Only collect what is necessary (4) Protecting Data Participants should be informed on the nature of study and which data are going to be used: More secure storage for more sensitive/identifiable data (encryption vs
password protection) Be aware of security standards for online/cloud data collection and storage Protecting Data (4) Participants should be informed on the nature of study and which data are going to be used: What happens when data travels? What happens if a journal requests your data?
Data Archiving Archiving data for secondary analysis (4,1) Seen as ethical practice: Ability to validate and refine or refute published findings using publically available data May ensure greater transparency of data Reduces burden of repeat data collection on participants Data should be retained for enough time to allow for others to answer questions that arise
from the research (5 -7 years minimum) Issues (4) Masking of potentially identifiable data though: Data swapping, Microaggregation, Adding random observations Distinction between formal managed archives and researchers holding data for personal use. Issues (4)
Ethical issues in personal archives (who manages them) Data quality: Each set of data an investigator obtains is a slice of reality that may or may not be valid or reliable Data Storage The responsible handling of data begins with proper storage and protection from accidental damage, loss, or theft: Lab notebooks should be stored in a safe place.
Computer files should be backed up and the backup data saved in a secure place that is physically removed from the original data. Samples should be appropriately saved so that they will not degrade over time. Care should be taken to reduce the risk of fire, flood, and other catastrophic events. Properly store and protect your data. They are valuable (8). Data Storage
Confidentiality. Some data are collected with the understanding that only authorized individuals will use them for specific purposes. In such cases, care needs to be taken to assure that privacy agreements are honored. This is particularly true of data that contain personal information that can be linked to specific individuals. It is also true of confidential information about protected processes and materials. If a company shares confidential data about a process with a researcher prior to seeking a patent on that process, the researcher must take care to make
sure the data are kept confidential. Data Storage Period of retention. Data should be retained for a reasonable period of time to allow other researchers to check results or to use the data for other purposes. There is, however, no common definition of a reasonable period of time. NIH generally requires that data be retained for 3 years following the submission of the final financial report. Some government programs require retention
for up to 7 years. A few universities have adopted dataretention policies that set specific time periods in the same range, that is, between 3 and 7 years. Aside from these specific guidelines, however, there is no comprehensive rule for data retention or, when called for, data destruction. Data Analysis Data analysis usually involves cleaning up the data and exclusion of bad results, such as
statistical outliers and results due to human or technical error (e.g. dropped the test tube, cant understand the interviewees speech, etc.). Sometimes data analysis involves dealing with missing data, i.e. data the was not recorded or edited properly (1). Data Analysis
If some data points from a particular measurement are missing, it can affect the statistical significance of the whole analysis. How do you deal with this? Image Manipulation Fields such as cell biology and remote sensing depend on digital images. Computer programs such as Photoshop make it easy to manipulate digital images (1).
Manipulation of images of gels Improper Statistics Statistics software makes it easy to misapply statistical methods. Emphasis on P-value (statistical significance) of 0.05 or less can lead people to manipulate statistical tests to lower the P-value to 0.05 level.
Example: A researcher might choose nonparametric regression over linear regression to get a better P-value (without having a good reason to) Data Mining Data Mining is an analytic process designed to explore data) in search of consistent patterns and/or systematic relationships between variables (7)
Data Mining Issues: Invisible data mining which builds data mining functions as an invisible process in the system so that users may not even sense that data mining has been performed Privacy-preserving data mining that aims to performing effective data mining without disclosure of private or sensitive information to outsiders.
Data Sharing Opennessthe sharing of ideas, data, methods, and resultsis a key ethical principle of research. Openness is essential for collaboration, criticism, confirmation, and replication of results (1). Data Sharing Funding organizations, such the NIH and NSF,
require researchers to disseminate their results and share data (following publication). In the US, private citizens can gain access related to government-sponsored research through the Freedom of Information Act. Data Sharing Protect private information about human subjects. Prior to publication, data pertaining to research involving human subjects should
be stripped of personal identifiers. Protect confidential business information (trade secrets), if your work is industrysponsored. Protect classified information, if your work is sponsored by the military (1). Data Sharing vs. Privacy Protection There is tension between global data sharing and privacy protection in genomics research (5):
Difficult to guarantee anonymity for participants Hard to provide information to satisfy informed consent requirements Data Sharing vs. Privacy Protection Need to improve research governance systems so they accommodate individual privacy, while still being effective at the global level:
E-governance, development of new IT interfaces Enable statistical analysis without compromising privacy Move to participant-centric system Research Data Management Planning (6) Case Study (1) Post-doctoral fellow Margot OToole accuses assistant professor
Thereza Imanishi-Kari of falsifying/fabricating data in a paper published in Cell in 1986, coauthored by Nobel prize- winner David Baltimore. Investigations by MIT, Tufts, Office of Research Integrity, and a Congressional Committee. Federal agents seized the records. A DHHS appeals panel found that Imanishi-Kari was not guilty in 1996. She admitted only to poor record keeping. Baltimore described the episode as a witch
Hunt. He resigned his position as President of Rockefeller University. MORAL: KEEP GOOD RESEARCH RECORDS! Case Study (8) Dr. Marion W. long ago learned that good data management practices are essential to responsible research. She therefore carefully supervises the work of her assistants and students, checking notebooks, backing up computer files, and from time to time verifying results for herself. As she is wrapping up work on one project before starting
another, the technology transfer officer at her university calls. A graduate student who previously worked in her laboratory has moved to another university and filed a patent for work that may have been done in Dr. W.s laboratory on her research funds? If this is the case, the graduate student may not be able to lay claim to the patent. Case Study (8) Cont. What records will Dr. W. need to prove that the work was done in her laboratory?
Who owns and controls the data collected in her laboratory? Do computer records pose any unique problems in this case? Hypothetical Case from NIH (1) Dr. Young is attending a conference in England where he meets Dr. Zenith. One night after a long session, Dr. Young and Dr. Zenith are socializing in a
pub. After a few beers, Dr. Young tells Dr. Zenith what he has found out about the mop gene and its role in cardiac myopathy. A month later a friend informs Dr. Young that Dr. Zenith has just submitted a paper about the mop gene (and the results sound almost identical to what Dr.Young has found). Dr. Young is quite angry about the situation, especially because he now has to rush to submit a paper (initially he wanted to submit a more complete paper).
Hypothetical Case from NIH (1) Was Dr. Zenith obligated to tell Dr. Young he was working on the same gene that evening in the pub? Should Dr. Zenith have told Dr. Young he was submitting a paper about the mop gene? If so, at what point? How should Dr. Young approach Dr. Zenith to
discuss the situation? Hypothetical Case 2 from NIH (1) Dr. Williams is a Principal Investigator who has a large laboratory at NC State University. The laboratory includes about 15 junior researchers, post- doctoral fellows, and graduate students. Twelve members of his group have been working on a project related to the relationship between hormones and obesity. They have isolated a key hormone in mice that is necessary to maintain normal
weight. They publish a paper on this new finding, with Dr. Williams as the senior author. Two months after the paper has been published, Dr. Williams receives an inquiry from a researcher at a large university who has had difficulty replicating some of the groups work. The researcher requests to see the original data used to support a figure presented in the paper. Dr. Williams asks members of his team for the original data related to the figure and they report that the experiments that
generated that data were conducted by Dr. VF, a post-doctoral fellow who recently left the laboratory to return to his native country. When Dr. VF left the institute, he was told to leave the original data at the institute and to take copies. A search of the laboratory for the original data has been less than satisfactory. The group discovers that there are several problems with the data, including the lack of a bound notebook and the availability of some postit sticky notes written in Dr. VF's native language. They also have trouble retrieving data that were stored on his computer, which has been infected by a virus. How should Dr. Williams deal with this issue?
Acknowledgement This slide show was compiled and developed by Austin Arrington, PhD candidate at SUNY-ESF, Syracuse, NY with review by NEEP faculty, funded by the National Science Foundation, Ethics Education in Science and Engineering (EESE), GEO-1032754. Bibliography
(1) Resnik, David. Who Owns Your Data?: Ethical Issues in Data Management. North Carolina State University. http ://www.ncsu.edu/grad/preparing-future leaders/docs/ Ethical_Issues_in_Data_Management_Handout.pdf (2)University of Glasgow. Guidance on Managing Research Records. http ://www.gla.ac.uk/services/dpfoioffice/guidanceonresearch / (3) University of Leicester. Research Data. http ://www2.le.ac.uk/services/research-data/create-data/dpethics (4) Guerin, Suzanne. Ethical Issues in Data Management,
University College Dublin -- School of Psychology. https ://www.ucd.ie/t4cms/Data_protection_ethics_library.pdf Bibliography (5) Kaye, Jane. "The tension between data sharing and the protection of privacy in genomics research." In Ethics, Law and Governance of Biobanking, pp. 101-120. Springer Netherlands, 2015. (6) The University of Western Australia. Research Data Management Toolkit. http ://guides.is.uwa.edu.au/content.php?pid=319161&sid=
2612215. 2015. (7) Han, Jiawei, and Jing Gao. "Research challenges for data mining in science and engineering." Next Generation of Data Mining (2009): 1-18. (8) Steineck, N. Responsible Conduct of Research, Report from the Office of Research Integrity. Data Management, Chapter 6, from Part Three, p. 87-102
Sector model- created in 1939 by Homer Hoyt, it is essentially a modification of the concentric zone model. Rather than rings however, this model uses sectors that lead out from the CBD generally in a radial pattern, with transportation and...
Statistical Capacity Building Project Database Technical Assistance in Statistics Team Development Economic Data Group World Bank 1. Develop material to make the case for statistics to a range of audience.
High use routes should be short and simple. Closed floor plans separate rooms Open floor plans have less walls, save on cost of building materials, and home feels more spacious Work Triangle Is the route from the center of the...
What the forum means to the people: "Being an active person on the Patient & Public Forum for Medicines during the last 4 years has been an amazing journey. Constantly learning, being involved in local medicines campaigns and medicines safety...
LANta is currently self-insured for health coverage. Total FY11 costs, including stop-loss premium, were $3M. From 2007 to 2011, costs increased 2% per year. LANta engages a consultant to review the estimated costs of self-insurance versus premium based plan. Employee...
Of the 8 cranial bones that fit together to form the braincase, the sphenoid bone is the "keystone". Like the keystone of a roman arch, the sphenoid is the "center brick" that balances the outward thrust of . the other...
It just creates a destination within your document Working with Web Site Structures A storyboard is a diagram of a Web site's structure, showing all the pages in the site and indicating how they are linked together It is important...
Anemia/Erythrocyte Disorders Laboratory Procedures Anemia Defined: A condition of reduced oxygen carrying capacity of erythrocytes Erythrocyte disorders May be associated with: Decreased production of RBC's Increased destruction of RBC's Inappropriate loss of RBC's Diagnosis A systemic, diagnostic approach to anemia...
Ready to download the document? Go ahead and hit continue!