Gallup data quality depends on original file format integrity

Evidence

Notes: Individual-level data from Gallup polls (accessed via ipoll), 1944-2004. South is defined throughout as the eleven states of the former Confederacy. Throughout the paper, we restrict ourselves to those ipoll surveys which were originally submitted as ascii files (as opposed to originally binary and converted into ascii by Roper). Roper warns that unreadable characters may have arisen in the conversion process and indeed every binary file we tried to read (roughly a dozen) had characters for some of our key variables that we could not decipher. We thus avoid binary files to ensure data quality. For this figure, we use every non-binary file that includes state and party-id from 1944-1980. From 1980-2004 Gallup surveys become more frequent and we choose just one per quarter to limit the burden of reading-in raw data files. — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 1022

Categorization

• XREF: connects to: Survey data-file encoding artifacts can corrupt key variables. — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 1022

Broader Connections

• Survey data-file encoding artifacts can corrupt key variables. #expands (0.78) — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 1022