BFpower.t.test_two_sample: Sample size determination for two-sample Bayesian t-test

Description

Perform sample size determination or the calculation of compelling and misleading evidence for a two-sample Bayesian t-test.

Usage

BFpower.t.test_two_sample(
  hypothesis = NULL,
  e = NULL,
  interval = NULL,
  D = NULL,
  target = NULL,
  alpha = NULL,
  model = NULL,
  location = NULL,
  scale = NULL,
  dff = NULL,
  model_d = NULL,
  location_d = NULL,
  scale_d = NULL,
  dff_d = NULL,
  de_an_prior = NULL,
  N1 = NULL,
  N2 = NULL,
  r = NULL,
  mode_bf = NULL,
  direct = NULL
)

Value

A data frame with the following columns:

p(BF10 > D | H1): Probability of obtaining compelling evidence in favor of the alternative hypothesis when the alternative is true.
p(BF01 > D | H1): Probability of obtaining misleading evidence in favor of the null hypothesis when the alternative is true.
p(BF01 > D | H0): Probability of obtaining compelling evidence in favor of the null hypothesis when the null is true.
p(BF10 > D | H0): Probability of obtaining misleading evidence in favor of the alternative hypothesis when the null is true.
Required N1: The required sample size for group 1 or the sample size input by the user.
Required N2: The required sample size for group 1 or the sample size input by the user.

If sample size determination fails, the function returns NULL.

Arguments

hypothesis: The hypothesis being tested (e.g., two-sided "!=", right-sided ">", left-sided "<").
e: The bounds for the interval Bayes factor (used when interval = 0).
interval: Integer (1 or 0). If 1, Bayes factor with a point null against a composite alternative hypothesis; otherwise Bayes factor with interval null and alternative hypotheses.
D: The bound of compelling evidence.
target: The targeted true positive rate (if direct = "h1") or true negative rate (if direct = "h0").
alpha: The targeted false positive rate (if direct = "h1") or false negative rate (if direct = "h0").
model: Statistical model of the analysis prior under the alternative hypothesis: Normal distribution ("Normal"), Normal moment ("NLP"), or scaled t ("t-distribution").
location: Location parameter for the analysis prior under the alternative hypothesis.
scale: Scale parameter for the analysis prior under the alternative hypothesis.
dff: Degrees of freedom for the analysis prior under the alternative hypothesis (if applicable).
model_d: Statistical model of the design prior under the alternative hypothesis: Normal distribution ("Normal"), Normal moment ("NLP"), or scaled t ("t-distribution").
location_d: Location parameter for the design prior under the alternative hypothesis.
scale_d: Scale parameter for the design prior under the alternative hypothesis.
dff_d: Degrees of freedom parameter for the design prior under the alternative hypothesis.
de_an_prior: Integer (0 or 1). If 1, analysis and design priors under the alternative are the same; if 0, they are not.
N1: Sample size of group 1.
N2: Sample size of group 2.
r: Ratio of the sample size of group 2 over group 1 (N2 / N1).
mode_bf: Integer (1 or 0). If 1, sample size determination; if 0, N1 and N2 are used for the calculation of probabilities of compelling and misleading evidence.
direct: If "h1", controls true/false positive rates (BF10); if "h0", controls true/false negative rates (BF01).

Examples

Run this code

BFpower.t.test_two_sample(
  hypothesis = "!=",
  e = NULL,
  interval = 1,
  D = 3,
  target = 0.8,
  alpha = 0.05,
  model = "t-distribution",
  location = 0,
  scale = 0.707,
  dff = 1,
  de_an_prior = 1,
  r = 1,
  mode_bf = 1,
  direct = "h1"
)

Run the code above in your browser using DataLab