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.

Comparison with G*Power

library(anovapowersim)

The balanced simulation and calculation-only functions accept gpower = TRUE when results should follow G*Power’s noncentrality convention. For within-subject designs, this corresponds to selecting G*Power’s as in Cohen (1988) option.

A typical 2 x 2 design

In a 2 x 2 mixed design, the default and G*Power conventions usually give very similar results. Here they differ by one participant per group:

default_2x2 <- power_n_calc(
  between = c(group = 2),
  within = c(time = 2),
  term = "group:time",
  target_pes = 0.08,
  power = 0.90
)

gpower_2x2 <- power_n_calc(
  between = c(group = 2),
  within = c(time = 2),
  term = "group:time",
  target_pes = 0.08,
  power = 0.90,
  gpower = TRUE
)

c(default = default_2x2$n_needed, gpower = gpower_2x2$n_needed)
#> default  gpower 
#>      63      62

Set gpower = TRUE in power_n(), power_curve(), power_achieved(), or power_sensitivity() in the same way when running simulations.

More than three within-subject levels

The conventions can diverge more noticeably for smaller samples and for terms with more degrees of freedom, such as within-subject factors with more than three levels. With four within-subject levels, the same inputs produce:

default_4 <- power_n_calc(
  between = c(group = 2),
  within = c(time = 4),
  term = "group:time",
  target_pes = 0.08,
  power = 0.90
)

gpower_4 <- power_n_calc(
  between = c(group = 2),
  within = c(time = 4),
  term = "group:time",
  target_pes = 0.08,
  power = 0.90,
  gpower = TRUE
)

c(default = default_4$n_needed, gpower = gpower_4$n_needed)
#> default  gpower 
#>      29      83

In this case, use the package default, gpower = FALSE, so target_pes continues to match the partial eta squared you supplied. Use gpower = TRUE only when reproducing G*Power’s as in Cohen (1988) result is specifically required.

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.