This space-delimited data file is an extract of the data from a large national telephone survey of a random sample of 465 residents of the United States at least 18 years of age.The columns in the data file are in the order left to right as they appear in the table below from top to bottom.The top of the data file also contains the variable names in the order the data appear from left to right.

Variable WTSC IDEOL DNEIGH DFAM DWORK DMEN DWOMEN PARTYID SEX Description Score on Hayes, Glynn, and Shanahan's (2005) Willingness to Self-Censor scale. Scores can range between 1 and 5. High scores reecting a greater willingness to self-censorthe tendency to censor one's opinion expression around audiences believed to disagree with that opinion. Political ideology. This is the arithmetic average response to two questions: \"The terms liberal and conservative may mean different things to different people, depending on the kind of issue one is considering. In terms of =economic issues, would you say you are..." =social issues and people's behavior, would you say you are...\" Response options were \"very liberal" (coded 1), \"liberal" (coded 2), \"somewhat liberal" (coded 3), \"moderate" (coded 4), \"somewhat conservative" (coded 5), \"conservative" (coded 6), \"very conservative\" (coded 7). Higher scores on IDEOL reect a more conservative ideology. A score of 4 reects an ideologically moderate or neutral person. Five of the variables are responses to a question about political discussion: \"I would like to lcnow which groups you talk to about politics and how often. On a scale from one to ten, where 'one' means NEVER and 'ten' means VERY FREQUENTLY, how often do you talk to. . .\": . .. neighbors . .. family . .. people at work .. men . . women Political party self-identication. Respondents were asked whether they consider themselves a Democrat (1), a Republican (2), an Independent (3), or \"Other\" (4). Respondent's sex (0 = female, 1 = male)