Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
Time period assessment
Message posted by Jack (via 213.122.92.56) on October 20, 2001 at 2:02 PM (ET)
In trying to find trends in global conflicts, what is the optimum period for study? And is 10 years enough? That is statistically speaking.
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: Time period assessment
Message posted by Tomi (via 154.32.142.159) on October 22, 2001 at 3:40 PM (ET)
It all depends on how often you are taking measurements (granularity of data). You really need at least 30 data items to do a decent time series analysis. So ten years would only be OK if you were counting the number of conflicts per quarter or (better) per month.It is possible with 10 data items to assess whether there is an overall increase, but any tests will not be very powerful.
Re: Time period assessment
Message posted by Jack (via 62.7.79.254) on October 28, 2001 at 6:55 PM (ET)
Thanks for the response. Most authors sample yearly (i.e. how many conflicts per year per type of conflict) Is that enough?
Re: Time period assessment
Message posted by Tomi (via 154.32.143.146) on October 28, 2001 at 11:57 PM (ET)
Yearly will do, but in that case you could do with more than 10 years worth of data.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.