Information for Institutional Review Boards (IRBs) Regarding Practices at PsychExperiments

Back to About PE

The experiments that collect data at PsychExperiments are of two types: laboratory experiments and research experiments.  IRB concerns for the two types of experiments differ slightly.

The laboratory experiments constitute an experiment package that is the on-line equivalent of experiment packages used in psychology departments everywhere in the country to provide students with research training.  Most of the experiments are drawn from what might be termed the “experimental canon” within psychology. Experiments on basic processes in perception and memory typify the laboratory experiments.  From a student’s point of view, the experiments look as if they are being run from a local exe file; however, there is the advantage that students can conduct their data collection any hour of the day from any web-connected computer that has the Authorware Web Player installed.

        We do not seek IRB approval for laboratory experiments.  The reason is that these experiments are used for educational purposes and most are simply replications with slight variations of classic studies in experimental psychology.  Nonetheless, it is conceivable that there are research questions of interest that could be addressed using aggregate data from the laboratory data archive.  Anyone interested in using data from the archive for research purposes should notify their local IRB of their intent to do so and obtain permission to use the archived data for research purposes.  

        Research experiments are experiments that have been submitted by professional researchers wishing to collect proprietary data.  We are pleased to host these experiments because we have convenient technology for doing so and because we believe in the concept of web-based collaboratories in the social sciences.  Anyone is free to submit an experiment to be hosted at PsychExperiments but the experiment must first be approved for web delivery by the researcher’s home institution.  The IRB approval is linked to the experiment in pdf file format for anyone to view with the Acrobat Viewer.  The researcher’s e-mail address is available for anyone who wants to contact the researcher with concerns over the purpose or content of an experiment. 
« Return to Top»

       Web-delivery of experiments precludes obtaining signed consent forms from participants. Nonetheless, one can make participation contingent on the participant’s consent. 

A scrolling text box that contains the information needed for participants to provide informed consent precedes each experiment at PsychExperiments.  The text in the box is identical to what would be used locally on a signed informed consent agreement, except for the editing required to reflect the fact that the researcher is not available in person to answer questions and a signature cannot be obtained.  Specifically, the form covers the topics of research description, risks and benefits, confidentiality procedures, and right to withdraw.  The information in the text box is contributed by whoever develops the experiment. To see a typical consent form, click here .  For laboratory experiments, an informed consent form is included for pedagogical purposes primarily.  Nonetheless, because the data in the PsychExperiments database may acquire archival value over time (see Paragraph 2 under Laboratory Experiments above), providing a consent form that makes note of this seems warranted. Participants must agree to the consent statement before any data can be collected. 
« Return to Top»

Procedures for Storing Data and Assuring Confidentiality

       IRBs are charged with assuring the ethical handling of data, particularly as concerns participant confidentiality.  The following information on data transmission, storage, and retrieval will outline the procedures used at PsychExperiments. 

Data Transmission and Storage . At the end of each experiment, participants are presented with the option of sending or not sending their data to the PsychExperiments database.  If they choose “Don’t Send,” the only record retained from the experiment session is an entry in a database table that records the information below:

UserIndex is a number assigned by the database to link records across tables, UserID is a randomly chosen ID assigned by the Authorware program to provide confidentiality, UserName is whatever name participants supply (the default entry is Interested Person), ExpName is the experiment that was accessed, ExDate and ExpTime are date and time on the server not the client computer.  The database table with this information is only available to system administrators.  Thus only system administrators have access to the names submitted by participants.  When a researcher or instructor with a legitimate right to participant names inquires if a particular person or student has submitted data , we provide that information via e-mail but we  do not provide the ID number that would allow matching a dataset with a name.

When participants choose the “Send Data” option at the conclusion of an experiment, data are sent to two separate tables.  The first is for experiment data (i.e., measures of the dependent variable for each of the experimental conditions); the second is for any participant information submitted along with the experiment data. The experiment data tables are specific to experiments; therefore, there are as many tables as experiments.  The data, which typically consist of trial information and measures of the dependent variable (e.g., correct-incorrect choices, reaction times, survey ratings, etc.), are illustrated with sample entries from the experiment data table for Dichotic Listening.  The entries are for Trial 1 data recorded for two participants identified by their UserIndex and UserID.

The second table to which data are sent is a personal information table that contains the UserIndex and UserID along with any participant information that was collected as part of the experiment (e.g., gender, age, race, handedness).  Whereas there are as many data tables as experiments, there is only one personal information table. The table below shows the structure of this table. The actual table has more columns than are shown because there is a column for any subject variable used by any experiment

 The fact that are no user names in either of the two tables created by using the “Send Data” option is important to participant confidentiality because the data that are publicly available at PsychExperiments are generated from these two tables. To obtain data from the tables, one uses a download link at the website which connects to an asp page on the server. The asp page creates a query that selects data according to experiment name, participant affiliation, and date and, then, writes a comma separated (.csv) text file that contains the selected data using one text line per data record. The text file can be downloaded along with an Excel macro that processes the data in ways appropriate to the experimental design.

The data that are sent to the three tables described above are sent unencrypted which means that it is feasible the transmissions could be intercepted and read. Although we cannot guarantee that data from individual participants won't be intercepted, the actual questions and stimulus materials are not included in the data transmission, making interpretation difficult if not impossible.
« Return to Top»

       An important IRB concern is whether research participants are coerced into providing data.  A form of de facto coercion is the use of experimental participation to acquire benefits such as extra points toward grades, experimental credit, or entry in a lottery.  The procedures at PsychExperiments assure that participants can receive the benefits of participation without actually submitting data. 

Whatever benefits are offered by researchers to their research participants are not dependent on a participant choosing the “Send Data” option.  Even if a participant elects “Don’t Send,” the program used to deliver the experiment will proceed to the benefits screen in cases where this option was elected by the researcher or the instructor.  Typically benefits are secured by an e-mail that is sent automatically to the research owner (in the case of research experiments) or the classroom instructor (in the case of lab experiments).  By default the e-mail contains the participant’s name, experiment name, time, and date.  The UserID is not included to prevent cross checking of Ids with actual data entries.

       Another form of coercion arises from the social context of typical experiments.  Once a participant has agreed to participate in a study, withdrawing becomes difficult because of the social demands that are especially salient in high status laboratories where high status professionals are involved in data collection.  Because data at PsychExperiments are collected in a non-social context, social demands that might inhibit withdrawal in a laboratory are not present. Participants can comfortably withdraw at any point by closing the experiment session.  Also, they can withdraw by choosing the “Don’t Send” option when asked whether they want their data to be entered in the experiment archive.  As noted above, choosing “Don’t Send” does not preclude receipt of whatever benefits are available to participants.  So long as a person completes all the experiment activities, the benefits are available.  They are not available if a person exits before completing the experiment activities.
« Return to Top»

      Among the benefits of experimental participation is learning about the purpose and results of experiments.  Feedback regarding laboratory experiments is provided in-class and via instructional material at the site under the link “Instructors Only.”  For research experiments, feedback is the ethical obligation of the research owner.  To make it possible to provide feedback, research participants are asked if they would like to receive feedback. Those who say “Yes” generate an automatic e-mail to the research owner in which they insert their personal e-mail or postal address. In addition, general information regarding the purpose of the experiment is given in text screens at the conclusion of research experiments.
« Return to Top»

Last revised:June 03, 2003 11:46:45 AM
Copyright© 1995 The University of Mississippi. All rights reserved.
Questions about this page?PsychExps