Obviously this book was published before cell phones were in the mainstream, but there were several reasons he gave me to exclude them (in my study) at this point in survey research. (Gonna look for the text of our conversation or a draft of that chapter in my thesis)
(OK found the thesis draft)
Cell phones were not included in this study for three reasons. First, many people who own cell phones have prefixes and area codes different than the one they reside in. For example, someone who purchased a cell phone in Jackson County, Illinois, and then moved to Canton, Ohio could be erroneously included in this study.
Second, many people own both a land line and a cell phone, giving them a higher probability of being contacted than people who just have land lines.
Finally, many people object to participating in studies on their cell phones because they don't want to use their minutes for that purpose (D. Dillman, personal interview, April 17, 2004).
Now, you can do it if you use random digit dialing, by collected a total list of prefixes for land and cell phone lines in the sample area you are drawing from. You just have to make sure that they are sampled in proportion to the frequency of people with the individual phone prefixes. But that still doesn't answer the problems that Dillman called into question during our conversation.
Be ware residents of Florida, New Hampshire, and Pennsylvania - its only a matter of time until a non-specific terror threat hits your states.
P.S. Thanks for the email exchanges Jerome - we are about to take it to the next level locally with Jeff's campaign. We are actually doing some great things nationally on-line, non fundraising related.