Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

Chi square on JASP 0.9

edited June 29 in JASP & BayesFactor

I've been trying out the latest JASP release (0.9.0.1) and I'm a bit confused regarding Chi Square. I used to perform it via the Frequencies>Contingency tables menu, but I've noticed that there is also a Chi Square option on the Frequencies>Multinomial tests menu. Using the same data, I get different results with both methods.

Why is that? Can someone explain what might be going on with this? Many thanks!

Comments

  • Hi ttome,

    Suppose we have a 2x2 contingency table. This yields 4 cells; if you format your data set so that it can be analyzed with the multinomial chi square and you indicate the expected frequencies, you will get the identical result. If you want a demonstration you can send me the data.

    Cheers,
    E.J.

  • edited July 2

    EJ, many thanks for your reply. I'm guessing this is actually my mistake, for I was using a 2x3 contingency table. I can use "regular" chi square on it, but not the multinomial schi squareb (JASP doesn't even allow me to define the two data fields I used on the contingency table-based chi square).

  • You can use the multinomial chi-square but it requires some extra work. You have to indicate the expected and observed frequencies for each of the six cells -- importantly, in the multinomial chi-square each of the six cells needs a different name. So basically, if you have a contingency table it is easiest to just use the contingency table functionality (but if you really want to you can use the chi-square under the multinomial)
    E.J.

  • Many thanks, I will be looking into it in a few days!

Sign In or Register to comment.