Skip to contents

Sample size for a two-arm continuous endpoint

Source: R/power.R
sampleSizeContinuous.Rd

Two-sample comparison of means. Delegates the balanced case's core to the noncentral-t (matching stats::power.t.test) and adds unequal allocation, cluster design effect, and dropout inflation.

Usage

sampleSizeContinuous(
  delta,
  sd,
  power = 0.8,
  sig_level = 0.05,
  ratio = 1,
  sides = 2L,
  dropout = 0,
  cluster_size = NULL,
  icc = NULL
)

Arguments

delta

Target difference in means (treatment - control).

sd

Common (residual) standard deviation of the endpoint.

power

Target power (default 0.8).

sig_level

Two-sided (or one-sided, see sides) alpha (default 0.05).

ratio

Allocation ratio n2/n1 (treatment/control; default 1).

sides

1 or 2 (default 2).

dropout

Expected proportion lost to follow-up; inflates enrolment (default 0).

cluster_size

Average subjects per cluster for a cluster-randomized design (NULL = individually randomized).

icc

Intra-cluster correlation (required with cluster_size).

Value

A trial_power object (n1, n2, n_total, n_enrolled, design effect, clusters, ...).

See also

sampleSizeANCOVA(), sampleSizeBinary(), powerContinuous()

Examples

sampleSizeContinuous(delta = 5, sd = 10)          # d = 0.5 -> 64/arm
#> <trial_power> two-sample continuous
#>   effect: delta=5, sd=10, d=0.5
#>   power 0.80 | alpha 0.05 (2-sided) | allocation 1:1
#>   n per arm: control 64, treatment 64  (total 128)
sampleSizeContinuous(delta = 5, sd = 10, dropout = 0.2)
#> <trial_power> two-sample continuous
#>   effect: delta=5, sd=10, d=0.5
#>   power 0.80 | alpha 0.05 (2-sided) | allocation 1:1
#>   n per arm: control 64, treatment 64  (total 128)
#>   enrol 160 to allow 20% dropout