Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: best way to manage data with recoding

From   "Copeland, Laurel A" <>
To   <>
Subject   st: RE: best way to manage data with recoding
Date   Thu, 1 Sep 2005 12:25:48 -0500

I strongly recommend 4+: enter the scores as written, save a copy of the
original dataset somewhere such as a zip file clearly identified as
containing the original data, use Stata to recode the reverse-scored
items INTO NEW VARS, retaining the originals as well and saving the
setup used to carry out the recoding.  Save the new dataset and analyze
that.  Save copies of both the original and the augmented datasets and
the setup on separate media (e.g., another computer, a remote webspace,
a flashdrive).  Two reasons for this sort of compulsiveness: (1) people
make mistakes; we all have lost important data at one time or another
due to a moment's error; backup copies are essential so you can easily
recreate your analytic dataset if needed; and (2) computers / diskettes
/ flashdrives / CDs fail, and it's not "if it fails" it's "when it
fails" that you need a good recovery plan.

Laurel A. Copeland

-----Original Message-----
[] On Behalf Of Christopher
W. Ryan
Sent: Thursday, September 01, 2005 12:22 PM
To: Statalist
Subject: st: best way to manage data with recoding

I am conducting a study that involves some paper-pencil Likert-scale
survey instruments.  Some of the items are reverse-scored to calculate
the final scale score.  I plan to use Epidata to enter the data.

What is the best and most legitimate way to handle the reversed coding:

1.  reverse the scores on the fly at data entry time and enter them the
reversed way?  (seems onerous and prone to error)

2.  enter the scores as written by the subjects, then use Stata to
recode them, and save the now-modified dataset (writing over the
original one), to use for the subsequent analyses?

3.  enter the scores as written by the subjects, and leave the "raw" 
dataset like that, doing a recoding at the start of each analysis
session, thus preserving the original dataset?

4.  make a new dataset, with the relevant items recoded, and save both
the raw and the "working" dataset?

5.  something else?

Christopher W. Ryan, MD
SUNY Upstate Medical University Clinical Campus at Binghamton and Wilson
Family Practice Residency, Johnson City, NY cryanatbinghamtondotedu
GnuPG and PGP public keys available at

"If you want to build a ship, don't drum up the men to gather wood,
divide the work and give orders. Instead, teach them to yearn for the
vast and endless sea."  [Antoine de St. Exupery]
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index