Class will meet on:
feb 24 0-12
march 7 1:30-5
- Scope of human subjects research
- Definitions of research
- Definition of human subjects
- "living individual(s) about whom an investigator (whether professional or student) conducting research obtains (1) data through intervention or interaction with the individual, or (2) identifiable information."
- OPRS manual sections
As research becomes more interdisciplinary, people are doing more research without it even occurring to them that it falls into this domain. It's about what you're capturing, not what the intent of your research may be.
Generalizable knowledge is what distinguishes research from marketing research. Even though marketers are publishing stuff, it's not the same type of publishing that IRB legislation deals with.
If the university is doing an internal review, then the assumption of not going to be published as generalizable knowledge for others, then get a waiver/exemption.
It's very easy to combine files and the ease of re-identification has become a research area in itself. (e.g., DNA and geneaology)
Origins of human subjects protection regulations
Research of questionable ethics
- Tuskegee syphilis study
- US prisons and phase I drug trials
- Tearoom trade (1970)
- Milgram's "improved study habits" (simulated shocks; Stanford, 1960s)
- Zimbardo's Stanford prison study
- Nuremberg Code, 1946
- Declaration of Helsinki, 1961
Milgram's research design was extremely rigorous--set up exactly as it should have been.
Basis of current guidelines
Belmont report: basic ethical principles
Autonomy--individual has the information on the risks and can make the decision for him/herself
are the people you're studying capable of making that decision for themselves? (children, patients, prisoners, etc.)
read: Bad Pharma by Ben Goldacre; he also wrote Bad Science; he explains research methods around clinical trials and biomedicine better than anyone; is it fair for every pharma company to conduct their own clinical trials just to get better market share?; withholding and release of partial data; cutting the outliers
the implementation of current guidelines Federal regulations (45 CFR 46)
Criteria for research review and approval
Institutional review bodies
Assurance of compliance
Tora Bikson at RAND (died last year)
Human subjects research process
Research refers to activities aimed at developing or contributing to generalized human knowledge
Generalizable knowledge includes theories, principles, or relationships and the data on which they are based; such knowledge can be corroborated by scientific methods
Research activities usually hace explicit objectives and formal procedures for attaining them; both are usually laid out in a research protocol.
Laying out the research protocol is the meat of the IRB documentation generated after one gets approval
When does research involve human subjects?
intervention, interaction, or identifiable data
What is private identifiable information?
private: about behavior that occurs in contexts in which people would not expect that data are being gathered about them; also info that people have provided with no expectation that the data will be made public (this is what the IRB deals with)
identifiable: data that can be directly linked to an individual research subject or indirectly linked by inference or because the researcher has created a link file
PII is a different legal term: personally identifiable information
Federal category, California category: name, address, phone number, SSN-->regulations about when you have to report a data breach
Mark Rotenberg EPIC and Katie Shilton on differences in terms of privacy and observation and surveillance and capture
Basic elements of informed consent
Purpose of research and source of funding
Research procedures and duration
Reasonably foreseeable risks or discomforts to subjects
Expected benefits to subjects or others
Nature of confidentiality protections
Voluntariness of participation (subjects may refuse or discontinue without negative consequences)
UCLA Investigator's Manual, Ch4
Tora Bikson, RAND
Criteria for approval of research (RAND guidelines)
Risks are identified and minimized (with special attention to vulnerable subjects)
Risks are reasonable in relation to anticipated benefits to subjects
Fully informed consent is sought and documentd
Subject selection is equitable
Data are monitored, if appropriate
Privacy is protected TKTKTK
Principal investigator responsibilities (UCLA guidelines)
Only one member of the research team can be designated as the Principal Investigator (PI). The PI has the following responsibilities described in the Principal Investigator Assurances:
- Utlimate responsibility for the conduct of the study, the ethical performance of the project, the protection of the rights and welfare of human subjects, and strict adherence to any stipulations imposed by the IRB.
- Assurance of compliance with all UCLA policies and procedures, as well as will all applicable federal, state, and local laws regarding the protection of human subjects in research.
Access to the server is much more powerful than access to the file cabinet.
Scaling difference: property records are public record; what happens when you can combine this data with others (i.e., OkCupid)
Privacy principles: OECD 1980-1
Basic principles of national application
Collection limitation principleThere should be limits to the collection of personal data and any such data should be obtained by lawful and fair means and, where appropriate, with the knowledge or consent of the data subject.
Data quality principle
Personal data should be relevant to the purposes for which they are used, and, to the extent necessary for those purposes, should be accurate, complete and kept up to date.
This is the pushback against collecting data from GPS, social media, etc.
Data are assets, but they're liabilities, too.
Purpose specification principle
The purposes for which personal data are collected should be specifiec not later than at the time of data collection and the subsequent use limited to the fulfillment of those purposes or such other as are not incompatible with those purposes and as are specified on each occasion of change of purpose.
Use limitation principle
Personal data should not be disclosed, made available or otherwise used for purposes other than those specificed in accordance with paragraph 9 except:
you need a different kind of informed consent if you're going to make your data open access-->need to anonymize the deeply stripped anonymous data will be shared to get the most valuable application of your time, etc.
Security safeguards principle
There should be a general policy of openness about developments, practices and policies with respect to personal data TKTKTKTK
Individual participation principle
An individual should have the right
a. to obtain from a data controller, or otherwise, confirmation of whether or not the data controller has data relating to him;
b. to have communicated with him, data relating to him
1. within a reasonable time
HIPPA, etc. enforces laws based on these principles, but these are just principles-->need to push for legislation if some violation is a condition of employment, health care, etc.
The more data aggregated in one place, the more valuable it becomes. Greater risk
If you can get ppl to think about these questions prior to designing a rkkg system, a lot of problems are avoided/addressed
data analysis: quantization-->reduce the precision of numbers of values
Want to find clever and innovative ways to ask questions that haven't been done before. Balancing
what can you get that you can use that will get a reasonable approximation of what you want to ask
Topics for journal articles
1/27 Topic 3: Reliability, validity, bias: tradeoffs bt the likelihood that the study can be replicated (reliability) and the truth value of the study (validity)
articles that do a good job in IS that other people are going to enjoy reading; brief;
2/10 Topic 4: Research designs; operationalization: find a good example of a good way to measure a fuzzy concept; taking a concept and deciding exactly how we're going to measure; big picture of when which methods are appropriate
2/24 Topic 5: Experimental and quasi-experimental research: a true experiment you need pre-test, post-test, and control group;control relationships tightly; more idiographic approaches; in IS, you rarely do a full experiment, you do a quasi-experiment; how you determine what should be independent vs. dependent variable (Shadish book is the best for this); determine where bias comes in
3/3 Topic 6: Randomization and sampling: does the sample represent appropriately to address the research question; statistical distributions are relevant
3/7 Topic 7: Survey research: how you design a survey, pick a population, that ppl can understand, will get you an answer, won't bias, will get responses, what are the tradeoffs of conditional questions with Likert scales, etc.; categorical, ordinal, intertktktk scale
3/10 Topic 8: Qualitative field research: catchall for things that aren't quantitative; how you think about more of the idiographic; need to think abt reliability and validity; Lofland
3/17 Topic 9: Data analysis: cross-cut (Babbie and Shadish) what do you do with it once you've got it; Lofland how you keep sifting down and looking at relationships in various ways