| T O P I C R E V I E W |
| wbturnbull |
Posted - 10/24/2002 : 05:21:49 AM
I have two datasets with equal numbers of data points, but one covers twice the range of the x-axis as the other. I expected that the dataset covering a wider spread of x values would give me better fitting, but instead it returns a higher value of chi^2 than for the other dataset. Can I directly compare the two chi^2 values in this case, or do I have to resort to an alternative statistical test to determine which set of parameters is more likely to be correct?
|
| 1 L A T E S T R E P L I E S (Newest First) |
| easwar |
Posted - 10/24/2002 : 10:45:43 AM
Hi,
I think your two datasets, one with a smaller range of x and one with a larger range of x, are essentially sampling two different populations of data. So it may be incorrect to compare the two chi-squares to see which set of parameters is better. This is also the reason why you do not see a better chi-sq value for the data with the larger x range.
If you are drawing samples from the same population, and if you draw two different samples, one with more points than the other, and perform fitting with the same model to both, the one with more number of data points should return a better fit/chi-sq value.
Hope this helps.
Easwar
OriginLab. |
|
|
