 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: Fraud methods in Stata
On 26 Sep 2008, at 16:20, Williams, Rachael wrote:
If anyone has any ideas of other ways to detect possible fraud I would
love to hear from you too!
Having been peripherally involved in a case of academic fraud, there
are two interesting tendencies that people show when inventing data:
1. They do not understand random variability. Typically, invented data
has a smaller dispersion than real data, and the distribution tends to
be uniform rather than normal. People often confuse 'random' with
'evenly distributed'. In particular, differences between measurements
supposedly taken on different occasions has a very narrow distribution.
2. The previous observation influences the value if the person is
inventing data at the keyboard. They try to make each observation
different, causing a correlation between the values for adjacent
observations.
In addition, watch out for data organised in a way that makes it hard
to transcribe from the written source but easy to enter on a computer.
In this case, instead of all the pretest items from a questionnaire
being entered, followed by the post-test items, each item was entered
with the pretest and post-test values in adjacent columns. Clearly,
this would be hopelessly inefficient if you really had the
questionnaires, but easy if you were inventing the data to show
pretest-posttest differences.
Ronan Conroy
=================================
[email protected]
Royal College of Surgeons in Ireland
Epidemiology Department,
Beaux Lane House, Dublin 2, Ireland
+353 (0)1 402 2431
+353 (0)87 799 97 95
+353 (0)1 402 2764 (Fax - remember them?)
http://www.flickr.com/photos/ronanconroy/sets/72157601895416740/
P Before printing, think about the environment
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/