Presentation on theme: "Part 6 Meta-topics: Ethics, sampling, and consent Psychological Science on the Internet: Designing Web-Based Experiments From the Ground Up R. Chris Fraley."— Presentation transcript:
Part 6 Meta-topics: Ethics, sampling, and consent Psychological Science on the Internet: Designing Web-Based Experiments From the Ground Up R. Chris Fraley | AMASS 2006
Ethics and informed consent According to federal guidelines ( ), most research conducted on-line should be deemed “exempt” by IRBs (IRBs, not you, must make that determination.) Why? Most tests and surveys are exempt unless “information obtained is recorded in such a manner that human subjects can be identified, directly or through identifiers linked to the subjects” (sect b 2) If personal information is not obtained, there is no way to trace a person’s responses back to him or her.
Ethics and informed consent Moreover, consent can be waived if the research involves no more than minimal risk to subjects, where minimal risk is defined such that “the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests” (section i). People can withdraw literally at any time—with no social norms influencing their participation.
Sampling concerns Are Internet participants representative of the population? –No. Are traditional college samples less representative than Internet samples? –Yes.
WebJPSP Male43%23% SES 1.1% poor 15.4% working class 5.5% lower middle class 46% middle class 29.2% lower middle class 2.8% upper middle class SES only reported for 5- 10% of articles; 85% university students (only 27% of general population have college degrees) AgeM = 27.6M = 23 CountrybroadTypically based on one culture Ethnicity77% white80% white Comparison of Web-based samples and JPSP-based samples From Gosling, Vazire, Srivastava, & John (2004)
Careless data An assumption that web participants won’t take the research seriously: –Provide false answers –Careless –Sabotage This assumption begs the question of why people would want to participate in research in the first place.
Careless data People participate because they want to gain some self-insight, they want their opinions to be known, their friends suggested the site, or they just want to have some fun and do something interesting. Web folks are fickle; they are not going to waste their time as a means to wasting your own.
Careless data Assuming people are intrinsically interested in participating, there are problems that emerge: –Press submit button twice, leading to duplicate entries –Try the study multiple times, to see how their feedback is affected Discussion of solutions to these problems. –“done it before” buttons with default to “yes” –Random code that gets submitted with response; allows duplicate entries to be sorted and deleted
Making your site known Create a stable (always present) home page for your lab or research site. Register the URL for that site at search engines, such as yahoo: APS, SPSP, and other organizations sometimes list URLs to on-line psychological research
Seven guidelines for maximizing participation rates Make the study interesting. –If the main research is not interesting, add something interesting to the design. Make it short. –No more than 10 to 15 minutes. Minimize legalese –People surfing the web know they can leave at any time. Don’t bog them down in text that they won’t want to read.
Seven guidelines for maximizing participation rates Promise feedback –On your introductory page, explain clearly (and concisely) what people will be doing and what they will receive in response Deliver good feedback –If your feedback is informative, people are likely to pass the link on to friends and feel that they did not waste their time Do not collect personal information –Do not make feedback or participation contingent upon registration. Do not use pop-up ads on your site. –If it seems commercial, people may not trust it as much.