The public opinion poll stories that appeared in six major United States newspapers during the 1972 Presidential campaigns provide the basis for this social science analysis. Every polling story reported in the sample newspapers during the one month time frame was checked against the eight criteria of the Opinion Research Corporation. These criteria pose the following questions: Who was interviewed? How many people were interviewed? How were the people interviewed selected? How were the interviews conducted? When was the interviewing done? How were the questions worded? Who conducted the poll? and What was the confidence interval (range of possible error)? The analysis rates each newspaper separately on a scale of eight, with one point possible for each of the criteria. Results suggest that none of the papers did a genuinely scientific survey and none apparently attempted to do one. The analysis indicates a need in American journalism to scientifically report poll results and, at the same time, deflate poll importance by stressing clearly the limitations involved in obtaining accurate, reliable poll results. (MAI)