| Comment Number: | 524083-00010 |
| Received: | 8/10/2006 5:32:00 PM |
| Organization: | CLASP |
| Commenter: | Christine Egan |
| State: | DC |
| Agency: | Federal Trade Commission |
| Rule: | Appliance Labeling Research |
| No Attachments |
Comments:
Dear Hampton, I must apologize that my travel schedule in July made responding to the FTC's latest Federal Register notice out of reach. Nonetheless, I have now had some time to review the materials you provided and wanted to provide you feedback. I understand this is not official or on the public record but I thought it might be useful to you and it will definitely make me feel better to contribute. First, I was very impressed overall with the elements of the research design included in the notice. The inclusion of a pre-test is very important and I was thrilled to see it in included. Also, the fact that FTC plans to test comprehension experimentally is EXTREMELY important as all the research shows that asking people which one is easy to understand (without experimental validation) is highly flawed and thus not a basis for sound policy decisions. One item I thought I would raise for your consideration is that showing consumers a comparison of 4 labels seems like it might be overwhelming. It is a lot for consumers to digest and try to choose between. The most I have ever tested at once was three for this reason. Given the computer-based platform of the survey, it is also a lot to try and fit on a computer screen at once. My suggestion is that FTC monitor this pretty closely in the pre-test. Longer than expected interview times might be one sign that this is fatiguing interviewees. With 8 to 10 treatment groups, the sample sizes for each (300) could potentially be small--perhaps too small for making comparisons among each one that are statistically significant and/or nationally representative? If you combine the Energy Star and Non-Energy Star treatment groups by label type (e.g., current labels with and without the Energy Star versus Star labels with and without the energy Star) is the sample size of 600 enough to make statistically significant comparisons among label types? And would FTC be comfortable with this approach? FTC should be able to have its firm run some stats to test this before hand. My comments are just meant to consider how the results would be able to be analyzed up-front before implementation. It is always frustrating to find out your results are limited in ways you didn't want after the fact. If your research firm says valid comparisons among the groups off 300 is possible, than its probably not an issue. On the dishwasher labels, the words electric and natural gas probably need to be highlighted. Otherwise the presence of those two operating costs is likely to be confusing. In fact, FTC might want to add a question asking folks why they think there are two numbers and what they mean. I am afraid some people might think the two #s are a comparison between themselves. Does the internet-based platform raise the overall affluence of the sample and potentially increases likely comprehension? Is Harris' system of weighting to make up for lack of representation in low-income homes considered within the range of best practice? I hope these comments are useful. Again, I was really impressed with the overall approach. It seems like you guys are on a good track for getting some sound and unbiased answers. Best regards, Christine Christine Egan Executive Director, CLASP