Data Entry: Verifying Data Quality

1 view

Skip to first unread message

Ajay Shenoy

unread,

Jul 30, 2013, 8:05:42 AM7/30/13

to income-dynam...@googlegroups.com

Hey, folks,

For those of you on digitization: as you start to get the data from spreadsheets into Stata, please

1. Keep do files and spreadsheets on hand - we want to be able to have the original spreasheets to check for errors at the source

2. Save the spreadsheets as .csv files - Stata is much less likely to screw up reading a csv file than an xls(x) file

3. Once the data are in stata, run some scatter plots of each variable and look for crazy outliers. In particular, if you plot births in 1969 against births in 1968 and find something really far from the 45 degree line it probably means ABBYY missed a decimal point but failed to express its uncertainty about it. I know you've all come across this issue many times, so ex post scatter plots are an easy way to spot them.