How to solve "the leading minor of order # is not positive definite"?

[maxjohnson1]

I am trying to run the following regression with ivprobit:

reg1 <- ivprobit(YVAR ~ CONTROL1 |XVAR|INSTRUMENT, data=finaldata)

However, I keep getting the message "Error in chol.default(mat) : the leading minor of order 3 is not positive definite". I saw in some blogs that this could be due to the model being overspecified, but I am just including one control variable at the moment so am not sure how this can be possible.

Does anybody know how to solve this issue? Any help would be greatly appreciated.

Thank you very much!

[technocrat]

@maxjohnson1

leading minor of order 3 is not positive definite

See Ben Bolker's answer for another reason this might arise. Without finaldata, it's not possible to test whether this is the case here. See the FAQ: How to do a minimal reproducible example reprex for beginners.

[maxjohnson1]

Hi, thank you very much for referring me to this post, but I am not sure it would apply to my case as I am not creating any sequences for my regression?

my dataframe "finaldata" looks as follows:

     YVAR CONTROL1                  XVAR    INSTRUMENT
1      NA      6                    0            0
2       0      5                    NA           NA
3       0      5                    0            0
4       1      6                    NA           NA
5      NA      4                    0            0
6       0      6                    52           23
7       1      6                    0            1
8       0      4                    31           3
9       0      4                    43           98
10      1      3                    0            0

The variables XVAR and YVAR have a lot of 0 values and NA but over half are positive values.

Any suggestions would be really appreciated!

Thank you very much