Dear Brendan,
That's curious, because the use argument to hetcor() works fine for me (see
below). Is it possible that you tried to use this argument without
specifying a data frame as the first argument to hetcor? If so, please see
?hetcor. If not, it would help to have an example.
I hope this helps,
John
------------ snip -------------
> set.seed(12345) # adapting the example in ?hetcor
> R <- matrix(0, 4, 4)
> R[upper.tri(R)] <- runif(6)
> diag(R) <- 1
> R <- cov2cor(t(R) %*% R)
> round(R, 4) # population correlations
[,1] [,2] [,3] [,4]
[1,] 1.0000 0.5848 0.5718 0.6233
[2,] 0.5848 1.0000 0.7374 0.6249
[3,] 0.5718 0.7374 1.0000 0.5923
[4,] 0.6233 0.6249 0.5923 1.0000> data <- rmvnorm(1000, rep(0, 4), R)
> round(cor(data), 4) # sample correlations
[,1] [,2] [,3] [,4]
[1,] 1.0000 0.5933 0.5659 0.6088
[2,] 0.5933 1.0000 0.7334 0.6230
[3,] 0.5659 0.7334 1.0000 0.5802
[4,] 0.6088 0.6230 0.5802 1.0000> x1 <- data[,1]
> x2 <- data[,2]
> y1 <- cut(data[,3], c(-Inf, .75, Inf))
> y2 <- cut(data[,4], c(-Inf, -1, .5, 1.5, Inf))
> data <- data.frame(x1, x2, y1, y2)
> data[1,1] <- NA
> hetcor(data)
Two-Step Estimates
Correlations/Type of Correlation:
x1 x2 y1 y2
x1 1 Pearson Polyserial Polyserial
x2 0.5932 1 Polyserial Polyserial
y1 0.5952 0.7409 1 Polychoric
y2 0.624 0.6316 0.5708 1
Standard Errors:
x1 x2 y1
x1
x2 0.02053
y1 0.03092 0.02296
y2 0.02027 0.01995 0.0374
n = 999
P-values for Tests of Bivariate Normality:
x1 x2 y1
x1
x2 0.4782
y1 0.4023 0.8871
y2 0.1166 0.5077 0.05526
> hetcor(data, use = "complete.obs")
Two-Step Estimates
Correlations/Type of Correlation:
x1 x2 y1 y2
x1 1 Pearson Polyserial Polyserial
x2 0.5932 1 Polyserial Polyserial
y1 0.5952 0.7409 1 Polychoric
y2 0.624 0.6316 0.5708 1
Standard Errors:
x1 x2 y1
x1
x2 0.02053
y1 0.03092 0.02296
y2 0.02027 0.01995 0.0374
n = 999
P-values for Tests of Bivariate Normality:
x1 x2 y1
x1
x2 0.4782
y1 0.4023 0.8871
y2 0.1166 0.5077 0.05526
> hetcor(data, use = "pairwise.complete.obs")
Two-Step Estimates
Correlations/Type of Correlation:
x1 x2 y1 y2
x1 1 Pearson Polyserial Polyserial
x2 0.5932 1 Polyserial Polyserial
y1 0.5952 0.7409 1 Polychoric
y2 0.624 0.6317 0.5711 1
Standard Errors/Numbers of Observations:
x1 x2 y1 y2
x1 999 999 999 999
x2 0.02053 1000 1000 1000
y1 0.03092 0.02295 1000 1000
y2 0.02027 0.01994 0.03738 1000
P-values for Tests of Bivariate Normality:
x1 x2 y1
x1
x2 0.4952
y1 0.4023 0.878
y2 0.1166 0.5255 0.05615
--------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox
--------------------------------
> -----Original Message-----
> From: r-help-bounces at stat.math.ethz.ch
> [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of BRENDAN KLICK
> Sent: Friday, December 01, 2006 1:42 PM
> To: r-help at stat.math.ethz.ch
> Subject: [R] error in hetcor function (polycor package)?
>
> I have been using the hetcor function in the polycor package.
> When I don't specify the use option everything runs
> smoothly. However, when I specify use either as
> "pairwise.complete.obs" or "complete.obs" I get this
error
>
> Error in optim(rho, f, control = control, hessian = TRUE, method >
"BFGS") :
> non-finite value supplied by optim
>
> Is this an error in the hetcor function or am I missing something.
> Thanks for your help.
>
> Brendan Klick
> Johns Hopkins University School of Medicine bklick1 at jhmi.edu
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.