The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

Nested loop plots

Alessandro Gasparini

2024-03-03

As of version 0.6.0, rsimsum supports the fully automated creation of nested loop plots (Rücker and Schwarzer, 2014).

library(rsimsum)

A dataset that can be purposefully used to illustrate nested loop plots is bundled and shipped with rsimsum:

data("nlp", package = "rsimsum")

This data set contains the results of a simulation study on survival modelling with 150 distinct data-generating mechanisms:

head(nlp)
#>   dgm  i model           b        se baseline  ss beta esigma pars
#> 1   1  1     1  0.17119413 0.2064344        E 100    0    0.1  0.5
#> 2   1  1     2  0.19822898 0.2048353        E 100    0    0.1  0.5
#> 3   1 50     2 -0.03404229 0.2071766        E 100    0    0.1  0.5
#> 4   1 82     1 -0.09263968 0.2040281        E 100    0    0.1  0.5
#> 5   1 82     2 -0.05095914 0.2026813        E 100    0    0.1  0.5
#> 6   1 33     1 -0.17013365 0.2038076        E 100    0    0.1  0.5

Further information on the data could be find in the help file (?nlp).

We can analyse this simulation study using rsimsum as usual:

s <- rsimsum::simsum(
  data = nlp, estvarname = "b", true = 0, se = "se",
  methodvar = "model", by = c("baseline", "ss", "esigma")
)
#> 'ref' method was not specified, 1 set as the reference
s
#> Summary of a simulation study with a single estimand.
#> True value of the estimand: 0 
#> 
#> Method variable: model 
#>  Unique methods: 1, 2 
#>  Reference method: 1 
#> 
#> By factors: baseline, ss, esigma 
#> 
#> Monte Carlo standard errors were computed.

Finally, a nested loop plot can be automatically produced via the autoplot method, e.g. for bias:

library(ggplot2)
autoplot(s, type = "nlp", stats = "bias")

However:

  1. Nested loop plots are suited for several DGMs but not for several methods;
  2. The decision on how to nest the results is subjective - the top-level of nesting receives most emphasis;
  3. It gives an overall impression, without focusing too much on details;
  4. It is cumbersome to incorporate Monte Carlo errors in the plot.

References

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.