1	Ethics & Values of Data: Coping with Complexity & Uncertainty Joan E. Sieber jsieber@bay.csuhayward.edu
2	To Do Practical Ethics: Thoughtfully select the values to be promoted. Minimize or balance conflicts among values. Consider how context can change priorities, nuances and values themselves.
3	Why Ethics of Data is Complex: Data archives are for future use. Anticipate the future nature, problems and methods of science. Assemble data archives likely to be useful in the future. Anticipate possible ways of combining diverse kinds of data in informative ways.
4	The Course of Science is... Complex and unpredictable. Man-made, but its course is somewhat beyond our control. Possible to conceptualize and cope with via chaos theory.
5	Important Data Satisfy: Micro-Ethics of Science: Build new knowledge Validity Transparency and appropriateness of methodology Adequate documentation Are shared with other scientists Macro-Ethics Have broader social implications and uses. Foster important social values, policies. Address important current concerns: e.g., education, health, environment, building science infrastructure.
6	NSF’s “Broader Impacts” Criteria: The U.S. Congress and the National Science Foundation Require Funded Projects to Seriously Address Macro-Ethical Issues
7	Find 5 pages of examples of “Broader Impacts” at: Http://www.nsf.gov/pubs/2002/nsfo22/bicexamples.pdf
8	Normative Ethical Theory DEONTOLOGY - follow the most inviolable rule or value. RULE UTILITARIANISM - follow the rule most likely to lead to the most good for the most people. ACT UTILITARIANISM - do what seems, in the particular case, most likely to lead to the most good for the most people.
9	Values: Many, changing, complex, uncertain Pre-September 11th - emphasis on openness Post-September 11th - consideration of national security: e.g., should gene sequences of human pathogens be freely available?
10	Technology Licensing at Universities 20 years ago: Seen as stimulus to innovation. Cash cow.
11	Technology Licensing at Universities Today: Inhibiting to productive young scientists. $ to Lawyers.
12	Data Related Values... Evolve rapidly with new discoveries, priorities, technologies. Vary with context -- academe, industry, government laboratories. Change with changing reward structures in science.
13	Data … ? Raw, cleaned, digitized? Qualitative, descriptive? Cell lines? Samples of rock, sediment, ice cores, DNA, bacteria? Fossils? Carbon dating? Financial records of the research administration?
14	Kinds of Archives Public data in public archives. Data of individual researchers shared informally via “invisible college.” Data of individual researchers shared via an organized archive. Privatized data: Produced by private industry. Publicly produced, value added.
15	Some Major Data-Related Values to Consider in Ethical Decision Making
16	Publicize. But How? Hard copy or electronic? Early and incomplete, or Later after elaboration? The file drawer problem ----> What about null results? The role of peer review, especially with null results and electronic publication.
17	Sharing Useful Data Avoiding data graveyards Serving methods of data integration: -- Meta-analysis -- EITM Assembling panel and longitudinal data in useful formats. Otherwise deciding what’s useful.
18	Responsibilities for Sharing Who releases data for sharing? How soon after publication? Who operates archive, answers users’ questions, updates the archive? How much sharing-related service is expected from the individual researcher? The public archive? The private archive?
19	Costs of Sharing Who pays for sharing services? What pricing formulas are used? Trade-offs between funding -- new research -- quality documentation -- quality services -- quality state-of-art technology
20	Protections Proportional to Main Goal Academe - Main Goal: Education - Mostly Open Government - Main Goal: Public Service - Mostly Open Business - Main Goal: Production & Profit - Mostly Protected Intellectual Property
21	Protections Proportional to Immediacy of Application
22	Astronomy No immediate applicability. Total openness of technology and data seems appropriate.
23	Natural Sciences Less business applicability. Educational value. Understanding of environment
24	NSF’s Broader Impact Criteria Advance discovery while promoting teaching / learning at all levels (K-post doc). Broaden participation in science. Enhance scientific infrastructure Disseminate broadly via many media. Benefit society; educate non-scientists, partner with all kinds of institutions.
25	How to “do ethics” with this kaleidoscopic array of values and contexts? Use of Concepts from Chaos Theory
26	Use complexity, change and uncertainty as a context for creative problem solving.
27	Experiment with small local changes, which can have big impacts. Avoid sweeping changes.
28	Chaotic structures self organize. See what scientists want in order to be productive.
29	Complex structures can have rich simple subtleties. Beware of stereotyped ideas.
30	Show-case and enjoy the elegant self-organization you find in scientific institutions.
31	Time may be regarded as a process. Use it that way to think about new data practices.
32	Reductionistic notions of dissection and control are only one approach. Accept chaos; be creative.