At AAPOR some years ago, Chang and Krosnick reported a national survey experiment in which the same questionnaire was administered via RDD telephone interviewing and by Internet survey firms. Those investigators found that the Internet respondents provided self-reports containing less random and systematic measurement error than did the telephone respondents. We have conducted a new experiment in which a single questionnaire was administered by eight different, major national survey firms to national samples. One survey was done by telephone by SRBI, and the others were done by Internet survey firms. The questionnaire included an array of embedded experiments designed to assess the extent of survey satisficing, including measures of non-differentiation, response order effects, don't know filter effects, acquiescence response bias, and more. Comparisons across the firms will be reported showing that the extent of survey satisficing varied depending on the mode of data collection. Additional analyses will show that these effects of mode and firm are especially prominent among subsets of the populations likely to be most sensitive to such data collection factors.