notes by jeff borland internet research torture numbers, and they'll confess to anything....
TRANSCRIPT
Notes by Jeff Borland
Internet Research
Torture numbers, and they'll confess to anything.
-Gregg Easterbrook
Notes by Jeff Borland
Internet Statistics Pros• Pro
– Up to date– Easy to access– Enormous amount
• Con– Reliability?
• Anyone can put stuff up• Data is difficult to attain• People may spin the stats • Can be outdated
Notes by Jeff Borland
Who’s data is it?
If you don’t know the source, check it out.
If you cant verify it, don’t use it.
Notes by Jeff Borland
Who’s data is it?• Get data from:
– Sources you have heard of:• New York Times, other papers• Newsweek, other magazines• Census or other governmental departments• Encyclopedias
– Wikipedia is good, but remember its user edited
• Nationally known research sites like PEW, or ROPER.– If you haven’t heard of it:
• Find what the agenda of the source is– What is their motivation? Do they have a bias?
• Find how they collected their data• Find what others are saying about the source, if its reputable.
Notes by Jeff Borland
Is it relevant?• Find the date of
the data, some statistics, like political polls need to be very up to date to be relevant.
• For other data, being very recent, doesn’t matter as much.
Notes by Jeff Borland
Is there a spin?• Does the site
have some bias in the information?
A survey of experts from the American Society of Criminology, the Academy of Criminal Justice Sciences, and the Law and Society Association showed that the overwhelming majority did not believe that the death penalty is a proven deterrent to homicide. Over 80% believe the existing research fails to support a deterrence justification for the death penalty. Similarly, over 75% of those polled do not believe that increasing the number of executions, or decreasing the time spent on death row before execution, would produce a general deterrent effect.
-from http://www.deathpenaltyinfo.org
Notes by Jeff Borland
How did they acquire the data?• Are there varying estimates?• Would the data be hard to get accurately?