I'm working with birdstrike data, and I would like to run statistics to verify if the frequency of birdstrikes does follow the number of aircraft movements around specific airports.
I have the data of birdstrikes and movements for a certain period. I can have two sets of data with ten years each. Could I use the aircraft movement on the total of birdstrikes to get the expected number of birdstrikes?
Exemple:
lets say: set 1 = 900 mov and set 2= 1100 mov
900+1100=2000
900/2000=0.45 | 1100/2000=0.55
set 1 = 45% and set 2= 55%
total of birdstrikes= 200+300=500
Expected:
Set 1=500*0.45=225
Set 2=500*0.55=275
set 1 | set 2
Obs: 200 | 300
Exp: 225 | 275
Is this possible inside the Chi-Square, Goodness of fit? Or is it recommended another kind of test for this kind of analysis?
hope this helps, and I hope I have understood it correctly.
I think chi^2 is your friend. One element of the chi^2 sum is the Pearson residual, which reads:
p_i = (obs_i - model_i)/sqrt(model_i)
The model_i in your case is just what you expect, i.e. model_i = f*mov_i
So you can from your model compute the expected values, and finally compute the chi^2 value as the sum over the squared Pearson residuals, and give the result to a chi^2 test which then returns whether your model fits the observations good or not good.
Kind regards --Peter
- Badges
- Science topic
- Similar topics
- Books