Meet {RaPhlot}

A R package that makes it easier to track the evolution of public opinion in cumulative American National Election Studies for both individual questions and their connections.

ANES

American National Election Studies (ANES) is possibly the most authentic survey used in studies of American political behavior. Conducted every four years (and sometimes every two years) since 1948, ANES provides a consistent and enduring window for political scientists to observe both the cross-sectional landscape and time-serial evolution of public opinion. Though its name suggests electoral behavior as the focus of the survey, ANES actually include a diverse range of questions that captures a respondent’s social, cultural, and psychological orientations. The relative stability of questions asked in ANES, many of which span across decades, is especially helpful for those interested in how the political mind of the American public is shaped by periodical and generational forces.

Consistent and high-quality as it is, there are also some conditions of ANES that holds back a smooth experience of exmaining long-term public opinion using this dataset. First, variables in ANES data are named with a somehow random combination of letters (usually “V”) and numbers, which suggests little information about its content at the absence of codebooks or variables labels¹. This means that every time we input the original version of ANES data, we need to rename the variables of interest with the user guide’s help, and take some time to revive our memory of what each variable literally asks the respondents.

Second, though variables pertaining to essential political behavior like partisanship or ideology are almost asked in every survey from 1948 to 2020, a large number of variables are either relatively new or had only been asked during several releases. The official website of ANES offers continuity guides that report the timing span of a variable, but there is no standalone support in statistical softwares to report such continuity with just one click. This adds some potential frustration into our explorations of ANES since descriptive or regression analyses in R would report error or return NA values when no observation is detected.

Say Hi to {RaPhlot}

The {RaPhlot} package is created to address these issues with additional support for analyzing how attitudinal indicators evolve in terms of their own trend and their relationship to other attitudes.

scatter.smooth(Nile)

Footnotes

1. Unlike the data inspector in Stata that shows greater details of variables, the data viewer in RStudio usually detach a variable from its original label, making the problem of abstract variable naming in ANES even more salient.