pwrss (version 1.3.1)

power.lp.test: Statistical Power for the Lambda-Prime Distribution

Description

Determines the power, the non-centrality parameter, or the degrees of freedom for the lambda-prime distribution with (optional) Type 1 and Type 2 error plots.

Usage

power.lp.test(
  power = NULL,
  ncp = NULL,
  req.sign = "+",
  null.ncp = 0,
  df = NULL,
  alpha = 0.05,
  alternative = c("two.sided", "one.sided", "two.one.sided"),
  plot = TRUE,
  verbose = 1,
  utf = FALSE
)

Value

power

statistical power \((1 - \beta)\).

ncp

non-centrality parameter under alternative.

null.ncp

non-centrality parameter under null.

df

degrees of freedom.

alpha

type 1 error rate (user-specified).

alternative

the direction or type of the hypothesis test.

t.alpha

critical value(s).

beta

type 2 error rate.

type.s

type S error rate (only for two-tailed test).

type.m

type M error rate (only for two-tailed test).

Arguments

power

statistical power \((1 - \beta)\); either power, ncp or df needs to be NULL (and is then estimated).

ncp

non-centrality parameter for the alternative; either power, ncp or df needs to be NULL (and is then estimated).

req.sign

whether ncp is expected to be greater '+1', less than '-1', or within '0' the null.ncp bounds; only relevant if ncp is to be estimated.

null.ncp

non-centrality parameter for the null. When alternative = "two.one.sided", the function expects two values in the form c(lower, upper). If a single value is provided, it is interpreted as the absolute bound and automatically expanded to c(-value, +value).

df

degrees of freedom; either power, ncp or df needs to be NULL (and is then estimated).

alpha

type 1 error rate, defined as the probability of incorrectly rejecting a true null hypothesis, denoted as \(\alpha\).

alternative

character; the direction or type of the hypothesis test: "two.sided", "one.sided", or "two.one.sided". "two.one.sided" is used for equivalence and minimal effect testing.

plot

logical; FALSE switches off Type 1 and Type 2 error plot. TRUE by default.

verbose

1 by default (returns test, hypotheses, and results), if 2 a more detailed output is given (plus key parameters and definitions), if 0 no output is printed on the console.

utf

logical; whether the output should show Unicode characters (if encoding allows for it). FALSE by default.

Examples

# two-sided
# power defined as the probability of observing test statistics greater
# than the positive critical value OR less than the negative critical value
power.lp.test(ncp = 1.960, df = 100, alpha = 0.05,
              alternative = "two.sided", plot = FALSE)
power.lp.test(power = 0.800, df = 100, alpha = 0.05,
              alternative = "two.sided", plot = FALSE)

# the two examples below estimate the df's based upon the first example
# (revealing a power of 0.498; df = 94.11) and the second example (revealing
# a ncp of 2.825; df = 101.06)
power.lp.test(ncp = 1.960, power = 0.498, alpha = 0.05,
              alternative = "two.sided", plot = FALSE)
power.lp.test(ncp = 2.825, power = 0.800, alpha = 0.05,
              alternative = "two.sided", plot = FALSE)

# one-sided
# power is defined as the probability of observing a test statistic greater
# than the critical value
power.lp.test(ncp = 1.960, df = 100, alpha = 0.05, alternative = "one.sided")
power.lp.test(power = 0.800, df = 100, alpha = 0.05, alternative = "one.sided")
# the two examples below estimate the df's based upon the first example
# (revealing a power of 0.6207; df = 100.323) and the second example (revealing
# a ncp of 2.506; df = 99.12)
power.lp.test(ncp = 1.960, power = 0.6207, alpha = 0.05,
              alternative = "one.sided", plot = FALSE)
power.lp.test(ncp = 2.506, power = 0.8000, alpha = 0.05,
              alternative = "one.sided", plot = FALSE)

# equivalence
# power is defined as the probability of observing a test statistic greater
# than the upper critical value (for the lower bound) AND less than the
# lower critical value (for the upper bound)
power.lp.test(ncp = 0, null.ncp = c(-3, 3), df = 100, alpha = 0.05,
              alternative = "two.one.sided", plot = FALSE)
power.lp.test(power = 0.80, req.sign = "0", null.ncp = c(-3, 3),
              df = 100, alpha = 0.05, alternative = "two.one.sided", plot = FALSE)
# adjust the power based upon what is returned from the example above in
# order to get a valid estimate of the df's (100.321; power = 0.8 -> 58.911)
power.lp.test(ncp = 0, power = 0.8103, req.sign = "0", null.ncp = c(-3, 3),
              alpha = 0.05, alternative = "two.one.sided", plot = FALSE)

# minimal effect testing
# power is defined as the probability of observing a test statistic greater
# than the upper critical value (for the upper bound) OR less than the lower
# critical value (for the lower bound).
power.lp.test(ncp = 2, null.ncp = c(-1, 1), df = 100, alpha = 0.05,
              alternative = "two.one.sided", plot = FALSE)
power.lp.test(power = 0.80, req.sign = "+", null.ncp = c(-1, 1),
              df = 100, alpha = 0.05, alternative = "two.one.sided")
# the first example (ncp = 2) reveals insufficient power (0.169), hence
# use the ncp returned from the example above for estimating the df's
power.lp.test(ncp = 3.844, power = 0.8, req.sign = "+", null.ncp = c(-3, 3),
              alpha = 0.05, alternative = "two.one.sided", plot = FALSE)