Some of the goodness of fit plots fail for directed networks.

[jrisi256]

So for undirected graphs, plotting gof works fine. This example is taken directly from the help pages.

data(florentine)
gest <- ergm(flomarriage ~ edges + kstar(2))
gofflo <- gof(gest)
plot(gofflo)

However for directed graphs, not all of the plots get plotted.

data(emon)
gest <- ergm(emon[[1]] ~ edges + ostar(2))
gofemon <- gof(gest)
plot(gofemon)

Specifically the geodesic distances (distance) and shared partner distributions (espartners) do not graph as I get the error: Error in out[, na.omit(i), drop = FALSE] : subscript out of bounds.

From R Session Info:

R version 4.0.0 (2020-04-24)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 22.04.1 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.20.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8       
 [4] LC_COLLATE=en_US.UTF-8     LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C              
[10] LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices datasets  utils     methods   base     

other attached packages:
[1] ergm_4.2.3         network_1.18.0     ina_0.0.12         RColorBrewer_1.1-3

loaded via a namespace (and not attached):
 [1] lpSolveAPI_5.5.2.0-17.8 rstudioapi_0.14         knitr_1.40              magrittr_2.0.3         
 [5] MASS_7.3-53.1           lattice_0.20-41         rlang_1.0.6             fastmap_1.1.0          
 [9] fansi_1.0.3             tools_4.0.0             parallel_4.0.0          grid_4.0.0             
[13] rle_0.9.2               xfun_0.33               utf8_1.2.2              cli_3.4.1              
[17] coda_0.19-4             tibble_3.1.8            lifecycle_1.0.3         Matrix_1.3-2           
[21] purrr_0.3.5             vctrs_0.4.2             trust_0.1-8             robustbase_0.95-0      
[25] memoise_2.0.1           glue_1.6.2              cachem_1.0.6            statnet.common_4.7.0   
[29] DEoptimR_1.0-11         compiler_4.0.0          pillar_1.8.1            renv_0.15.5            
[33] pkgconfig_2.0.3  

[mbojan]

Thank you @jrisi256 . Looks like a bug in CRAN and master versions of the package.

[mbojan]

Interestingly when computed individually distance works while espartners fails (and in both cases model is included even though it should not, but that's something separate):

gof(gest, GOF = ~ espartners - model) |>
  plot()
## Error in out[, na.omit(i), drop = FALSE] : subscript out of bounds

gof(gest, GOF = ~ distance - model) |>
  plot()

@jrisi256 until we fix it you can produce the gofplot for distances as in the above.

[mbojan]

Given

gof(gest, GOF = ~ espartners - model) |>
  plot()
## Error in out[, na.omit(i), drop = FALSE] : subscript out of bounds

It seems that i computed at

https://github.com/statnet/ergm/blob/033dd693abd7c93d38daa938eea369fa645b83bb/R/gof.ergm.R#L502-L512

is too long (here 1:14) for the actual dimension of out being subscripted (here it is 100 x 13). The logic behind computing i is obscure to me. Is this your design @krivit ?

[krivit]

Is this your design @krivit ?

"Design" isn't the right word. gof() used to be a mess with a lot of repetition, which I eventually turned into a mess with less repetition.

[krivit]

This problem turns out to have been affecting not so much directed networks as any networks dense enough to have maximum observed ESP close to the size.

[jrisi256]

Do you know when this will be pushed to CRAN?

[krivit]

Do you know when this will be pushed to CRAN?

Already has been. It's up to CRAN now.