Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
kevinburke
on May 16, 2011
|
parent
|
context
|
favorite
| on:
Everyone sucks at interviewing. Everyone.
Work sample tests are the best predictor (p of 0.54) of success on the job.
http://bobsutton.typepad.com/my_weblog/2009/10/selecting-tal...
tomjen3
on May 16, 2011
[–]
Yes isn't p usually supposed to be less than 0.05 for the it to usefull?
Robin_Message
on May 16, 2011
|
parent
|
next
[–]
It (probably) refers to a measure of correlation being 0.54 on a scale of 0=no correlation to 1=perfect linear relationship (
http://en.wikipedia.org/wiki/Pearson_product-moment_correlat...
) rather than a confidence interval, which is what p<0.05 refers to (
http://en.wikipedia.org/wiki/P-value
)
kevinburke
on May 16, 2011
|
parent
|
prev
[–]
Yes - sorry - r of 0.54, not p.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
