I would like to report some Bayesian Independent Samples t-Test results in a table, but I don't know how (cannot seem to find any info on the Web).
Table 1 shows how I did for the Frequentist Independent Samples t-Tests:
Table 1. Prevalence, Odds Ratio (OR), and Chi-Square test of association (X2) of DV at two cut-offs broken down by the two modalities of IV.
Code: Select all
--------------------------------------------------------------------
DV
------------------------------------------------
IV N at Cut-off1 (%) N at Cut-off2 (%)
--------------------------------------------------------------------
Modality1 26 (13.0) 49 (17.4)
Modality2 33 (13.8) 55 (14.1)
OR (95% CI) 0.969 (0.362-2.592) 1.720 (0.591–2.516)
X2 (df=1, N=350) 6.415, p < .01 7.811, p < .05
--------------------------------------------------------------------
Now, in another table, instead of the Chi-square, I'd like to put some BF01 (Bayesian Factor where I posit that H0 is true). Here are my questions:
- Can I simply swap the X2 for BF (and keep both the Ns and % for the 2 modalities, and ORs and CIs)?
- Is there something to add (the equivalent of df and N in a chi-square: like the error % and/or the median and 95% CI given in the Prior and posterior plots with additional info)?
- If yes, how should it be presented?
Table 2. Prevalence, Odds Ratio (OR), and Bayesian Independent Samples t-Test (BF) of DV at two cut-offs broken down by the two modalities of IV.
Code: Select all
--------------------------------------------------------------------
DV
------------------------------------------------
IV N at Cut-off1 (%) N at Cut-off2 (%)
--------------------------------------------------------------------
Modality1 26 (13.0) 49 (17.4)
Modality2 33 (13.8) 55 (14.1)
OR (95% CI) 0.969 (0.362-2.592) 1.720 (0.591–2.516)
BF01 (error %) 7.714 (1.718e-5) 3.531 (1.718e-5)
--------------------------------------------------------------------
Does it make sense?
Any help is highly appreciated.
Best regards,
pdeli
P.S.: Numbers in the tables are made up.
P.P.S.: If nobody can answer, is there another place on the web where I could ask this?
++++