Tests of Comparison Bug?

[Steve]

I have noticed that if I try to perform a "Test of Comparison" with Weibull++ and override the autocalculated limits, I can get different answers depending which set I have chosen as the "first" set and the "second" set. That is, if I reverse the same two sets of data, I get different results. More specifically, Data Set "A" looks more relialble than Data Set "B" in one case, then, with the data sets reversed Data Set "B" looks more reliable than Data Set "A". Has anyone else seen this? :confused:

[David]

Hi Steve,

In general, changing the settings of the integration (e.g. upper limit) can change the results since this alters the area over which the integration is taking place. In regards to the switching of results, if you could email the data set to support@reliasoft.com we will be happy to take a closer look.

Thanks.

[Pantelis]

To add to the above. Changing the limits changes the problem... so unless you really know what the function looks like and what these limits should be they should be left on autocalculate ...