Difficulty Obtaining Bayesfactor for Bayesian Correlation (correlationBF)
Jakob-Kasper opened this issue · comments
I am trying to compute a Bayesian correlation between two variables (val_pvv
, ther_pvv
) within my dataset (df_clean
), using the correlationBF
function:
cor_pvv_ther_val <- correlationBF(df_clean$val_pvv, df_clean$ther_pvv)
This results in the following warning:
Warning message: In genhypergeo_series_pos(U = c((n - 1)/2, (n - 1)/2), L = ((n + : Series not converged.
The describe_posterior
function provides a summary of the posterior distribution, but not the Bayes Factor (BF):
Summary of Posterior Distribution
Parameter | Median | 95% CI | pd | ROPE | % in ROPE | BF | Prior
-----------------------------------------------------------------------------------------
rho | 0.92 | [0.90, 0.93] | 100% | [-0.05, 0.05] | 0% | | Beta (1 +- 1)
As can also be seen here:
> extractBF(cor_pvv_ther_val)
bf error time code
Alt., r=0.333 NA 0 Fri Apr 5 11:10:22 2024 504709623a0
Interestingly, JASP allows me to obtain the BF = 8.251*10^191.
Given the high median, I suspect the issue with the BF calculation might be related to the extremely high correlation. To give a reproducible example: the same happens when I simulate two variables that are very highly correlated (correlationBF(c(1,2,3,4,5,6,7,8,9), c(0.7, 2.3, 2.5, 3.8, 5.2, 5.8, 7.5, 8.4, 8.8))
). I have the same issue when modifying the rscale
argument (e.g., to wide
or ultrawide
).
I would be very happy about some further information on the warning and about some suggestions on how to solve the issue.
Thank you very much in advance!
This seems to be the solution to the issue: https://stats.stackexchange.com/questions/644203/difficulty-obtaining-bayesfactor-in-baysian-correlation-in-r-correlationbf-from/644954