Protecting Science and Human Subjects in Online Research
Protecting Science and Human Subjects in Online Research Jos A. Bauermeister, MPH, PhD John G Searle Assistant Professor Health Behavior & Health Education University of Michigan School of Public Health Outline
Online methods Transitioning into Online Data Collection When? Why?
For What? Issues to Tackle Scientific Integrity
Human Subjects Regulation Other Considerations HIV/AIDS Prevention Online 90% of Americans use the Internet
Broadened discourse of sexuality HIV/AIDS e-prevention needed
Online networks are a part of every day life Pequegnat, W., et al. (2007). Conducting Internet-based HIV/STD Prevention Survey Research: Considerations in Design and Evaluation, AIDS & Behavior, 11(4), 505-521. Online Data Collection Benefits:
Incredible reach of internet 90% of YA use internet 85% of YAs have broadband at home Speed and ease of survey administration
Challenges: Response rates Generalizability Digital Divide The Family Tree Interviewer-administered surveys Self-administered surveys
Paper and pencil surveys Mail/postal surveys Telephone interviews Disk-by-mail (DBM)
Replacement technology Limitations of mail or telephone surveys Individual burden Data collection costs Systematic data collection Adaptive Complex skip patterns Couper, M.P. (2008). Designing Effective Web Surveys. New York: Cambridge University Press.
Why? (cont) Supplemental technology Mixed-method data collection Visual/interactive elements
Additional tool to acquire data Social validity Couper, M.P. (2008). Designing Effective Web Surveys. New York: Cambridge University Press. Collecting Sensitive Information Sensitive Information
Questions are contingent on the study goal. Online questions Avoid social desirability regarding sensitive questions
Improve Ensure Tailor recall data consistency
content to participants Remove irrelevant content Program change into survey Approaches
Avoid interviewer-based social desirability regarding sensitive questions Pregnancy Substance Use Sexual
Behavior Mental Health *** Symptoms vs. Diagnosis
Approaches Improve recall Use of prior data entries
Use of event calendars Approaches Ensure data consistency
Approaches Tailor content to participants Remove irrelevant content Use formal vs street language
Approaches Tailor content to participants Language Preference The following questions refer to your sexual behavior during the past 30 days. Our focus will be exclusively on anal, vaginal, and oral sex. Therefore, do not include in your answers references to partners with whom you did not engage in anal, vaginal, and oral
sex. Would you like to see these questions in formal language or street language? Formal language 60(45.5%) Street language 72 (54.5%) Online Data Collection
Considerations re: Human Subjects Participant Understanding of Study Participant Burden
Data Security Collection of (non-consented) third party data Standard recommendations
Piecemeal consent by sections Certificate of Confidentiality
In-house survey administration and data management 128-bit SSL Encryption
Use of well-validated psychometric measures In-house survey administration Collect through a third-party?
Buy licensed software? Data Repository What strategies are in place to offset a breach? Where are the data?
Who accesses the data? How are the data files (en)coded?
When do you download/clean the data? DATA QUALITY Why does data quality matter?
Increasing adoption of web-based data collection in sex research. Important to make sure data collected through this modality are valid and reliable, and that conclusions are accurate. Duplication Falsification
Decrease research costs. Bauermeister, J.A., Pingel, E., Zimmerman, M.A., Couper, M., Carballo-Diguez, A., & Strecher, V.J. (2012). Data quality in web-based HIV/AIDS research: Handling Invalid and Suspicious Data. Field Methods, 24(3), 272-291. How common are these issues?
Invalid entries occur commonly in web-based research. Konstan and colleagues (2005) found that 11% of entries in their MSM sample (N=1,150) were duplicate entries from participants. Bowen
and colleagues (2008) found that approximately 1/3 of the 1,900 total submissions among MSM were multiple entries. Bauermeister and colleagues (2012) found that 16% of entries in their YA sample (N = 3,448) were falsified. Bauermeister
and colleagues (2014) found that 15% of entries in their YMSM sample (N = 2,329) were falsified. Preventing falsified data Automated Procedures Eligibility
screeners Restriction of one submission per IP address Reverse IP lookup Creating statistical algorithms Manual Procedures Cross-checking
Flagging entries from similar IPs cases Geocoding spatial data (if available) Dear participant, We appreciate your interest and willingness to complete our survey on young men who
have sex with mens (YMSM) online dating behavior. Unfortunately, we have noticed irregularities during data collection and have had to stop our study. Specifically, a few individuals have chosen to provide false data and/or create multiple entries so that they may receive one or more $15 iTunes incentives. We cannot underscore how disappointing this has been for us. As public health practitioners, we strive to collect quality and robust data through research that will inform HIV/AIDS and sex education programs for young men. False data diminishes our ability and actually harms the population that we seek to help through science and social services. We hope that similar events will not occur in future efforts. It is only through the
honesty, integrity, and willingness of participants that we can help to contribute to the eradication of HIV and other sexually transmitted diseases from our communities. Recommendations Pre and post hoc decisions regarding how to handle suspicious data are warranted.
We encourage researchers using web-surveys to: Mention whether data exclusion criteria are set. Explicitly state whether the presence of invalid data will be examined and how it will be handled.
Quality criteria for web-survey research may be an important covariate in meta-analyses. Questions?
Tessellations M.C. Escher (1898-1972) Jim McNeill (1967- ) Hand with Reflecting Sphere 1935 Drawing Hands, 1948 Relativity, 1953 Waterfall, 1961 House of Stairs, 1951 Reptiles, 1943 Sky and Water, 1938 Liberation Monkeymen Mirror Circle Limit III, 1959 Butterfly Tessellations M.C....
* A One-Way Analysis of Variance (ANOVA) was conducted on the results of the three scenarios, in order to determine which scenario resulted in the smallest number of entities in the Process queue. After opening Arena's Output Analyzer, the user...
Significance and Introduction Genes and proteins are often associated with multiple names Apo3, DR3, TRAMP, LARD, and lymphocyte associated receptor of death Authors often use different synonyms Information extraction benefits from identifying those synonyms Synonym knowledge sources are not complete...
Lab Procedures. Never touch lab equipment without a teacher's permission.. Always. wear your lab safety goggles and lab apron when in the lab. Pull back . long. hair. Wear . closed-toe. shoes during a lab. Do . not
Topological insulators. Magnetic Monopole. Gauge Transformation. Vector potential cannot be defined globally. Matter field. wave-function on each semi-sphere is single valued. ... QUANTUM LIQUID CRYSTAL PHASES IN STRONGLY CORRELATED FERMIONIC SYSTEMS
The modern, spa and beauty centres offer luxurious and indulging spa treatments for body and soul, which make everybody feel at home. The spacous beaches, sea water, together with the mineral water and peloids transform Varna into a a centre...