I'm getting a warning message (Warning: no valid starting value for precision parameter found, using 1 instead) while fitting a beta regression model using betareg::betareg. What does it mean? Is that something of concern? My response variable is positive, continuous with values between 0 & 1. Thanks for any assistance.

data_assessment_num$humidity_catog <- 
 levels = c("RH1(0, 65)", 
 "RH2(65.01, 80)", 
 "RH3(80.01, 90)", 
 "RH4(90.01, 96)"))

mod01 <- betareg(severity ~ humidity_catog, data = data_assessment_num)

Warning: no valid starting value for precision parameter found, using 1 instead
  • $\begingroup$ there's no betareg function that comes with R. If you're using a package, you should (a) read the help of the function you call from whatever package it's in, and (b) when you ask about it, make it clear to us what package you're using -- it's common for several packages to use identical function names. $\endgroup$
    – Glen_b
    Commented Oct 25, 2022 at 23:00
  • $\begingroup$ Thanks for the comment. The function is from betareg R package. I saw the package documentation but couldn't found anything on this. $\endgroup$
    – Ahsk
    Commented Oct 26, 2022 at 10:46
  • $\begingroup$ Edited to indicate that. $\endgroup$
    – Glen_b
    Commented Oct 27, 2022 at 3:25

1 Answer 1


By default, betareg() runs an auxiliary regression (see Ferrari & Cribari-Neto 2004) to determine the starting value of the precision parameter phi. If this is not strictly positive, then phi = 1 is used as the starting value and the warning is issued.

This may not be critical at all. But it may also indicate a poor fit, possibly just for certain factor levels. So I would recommend to have a close look at your fitted model to check whether everything seems to work ok.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.