1. Have a script of what the person interacting with the participant should do. The script should include
1. A greeting
2. The procedures (what the researcher will say and do while interacting with the participant). Be sure you have addressed
A. Order effects: A big threat to internal validity and one that is often dealt with by randomizing.
B. Participant and subject bias: A big threat to construct validity—and a threat that is often dealt with by using “blind” (masking techniques).
3. A debriefing (which should include a “thank you”). For more information on debriefing, see http://www.markwebtest.netfirms.com/Appendix/App_Conducting/Practical Tips for Conducting an Ethical and Valid Study.pdf
Note: If you have trouble accessing that site, go to http://kroxy.net and then cut-and-paste the URL into the search box.
2. Have a data sheet (see sample on attached sheet) that will allow you to record
1. What condition the participant was in.
2. The participant’s score on the dependent measure.
3. Any demographic information that might be relevant (e.g., gender, member of this class)
4. Any other notes about the trial that might be relevant (e.g., the participant guessed the hypothesis, the researcher deviated from procedure, a cup spilled).
3. Have all your materials—and enough copies of those materials. If you are going to have participants make a rating, you should have a rating scale.
4. Practice using your script and coding your data. Be sure everyone in your group can conduct the study the same way.
5. Assign tasks to each member of your group. This may be particularly important if you wish to use blind (masking) techniques.
6. If you are varying a treatment or the order of exposing participants to a treatment, know which participants are getting which treatment in which order. That is, know what treatments—and in what order—the first participant will get, the second participant will get, and so forth. Note that, for construct validity reasons, you may not want the person who interacts with the participant to know the condition. However, it may be that, to avoid confusion, you let the person interacting with the participant know the condition. (If making the experimenter blind confuses everyone so much that you don't know which drink the participant took, you don't have any reliable data from that participant.)
8. Pretend to collect data from each other (for now, this may require some pantomiming and other acting skills. But don't use that data.
9. If you have time, fill out an abbreviated IRB form (available from http://www.markwebtest.netfirms.com/teachRDE/C02/shortIRBform.doc ) for your study. ( Note: If you have trouble accessing that site, go to http://kroxy.net and then cut-and-paste the URL above into the gray box).
Notes:
1. If you have questions about how to conduct a study, you may find the following link helpful:
Note: If you have trouble accessing that site, go to http://kroxy.net and then cut-and-paste the URL into the search box.
2. To randomly assign participants each participant to one of several sequences (e.g., to either condition 1, then condition 2 or condition 1, then condition 2), follow the instructions on page 620 of your book. Alternatively, you could
go to http://faculty.vassar.edu/lowry/rand01.html
enter the number of conditions you have (e.g., 2) for n
press the randomize button
give the first participant the conditions in the sequence given by the randomizer (e.g., “1, 2” or “2,1”)
repeat for as many participants as you expect to test.
Another option is to use the tutorial at http://www.randomizer.org/tutorial.htm
Finally, you could have a list randomizer do the work for you. To see how that would work, imagine that you had 2 conditions (we will call them A and B) and you planned to run 20 participants in each condition. In that case, you could enter 20 As and 20 Bs into the list randomizer at http://www.random.org/lists/
To analyze your data, go to the "Frequency Data" section of http://faculty.vassar.edu/lowry/VassarStats.html Then, go to the "Binomial probabilities" section. For your first analysis, enter the number of participants you have for n, the number who picked the favored drink (e.g., Dr. Pepper) for k, and enter .5 for p.