The recent proliferation of cell phones has generated much speculation regarding their potential effect on surveys and particularly RDD surveys. To collect information on cell phone usage and cell phone users, the 2004 JPSM Practicum of the University of Maryland conducted a survey that included a substantial cell phone sample. One open-ended question was included to investigate response differences between those interviewed on cell phones and those interviewed on landlines ("What do you like about having a cell phone?"). This was asked of all respondents in both samples who reported having a cell phone. All responses were recorded verbatim and a standard probe for additional information was delivered. We plan to analyze this open-ended question by coding the content of responses and doing counts of the number of characters, words, and phrases. We wish to investigate whether there are any differences in open-ended responses by telephone device used. Cell phone respondents may be concerned about the minutes used (and related cost) for the interview, which may result in brevity of response for cell phone interviews as compared to landline interviews. For the cell phone sample, the data set includes their location at the time of interview, which allows examination of contextual effects, that is, effects related to being interviewed while in the car, grocery store, etc. We will also investigate possible interviewer effects by telephone device by conducting a hierarchical analysis with the interviewer at one level and respondent at the second level. To avoid confounding the analysis (i.e., individual-level analysis from a household-level survey), only household level variables will be used. For the hierarchical modeling, we expect this not to be an issue because we have interviewer characteristics at the first level and respondent household characteristics at the second level, with probably no interactions across levels