any2any: any2any

Description Usage Arguments Details Value Note Author(s)

View source: R/fRedsRutils.R View source: R/C-3PR_ASCII.R

Description

Converts most common test statistics into most common (signed) effect sizes.

Converts most common test statistics into most common (signed) effect sizes.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
any2any(testInfo, df1 = NULL, df2 = NULL, N = NULL, n1 = NULL,
  n2 = NULL, esType = NA, var.lor = NA, CIcalc = TRUE, CL = 0.95,
  rID = 0, q = 1, alternative = "two", keepDirection = TRUE,
  keepSign = TRUE, keepSignNames = c("r", "l.r", "u.r", "fisher.z", "l.z",
  "u.z"))

any2any(testInfo, df1 = NULL, df2 = NULL, N = NULL, n1 = NULL,
  n2 = NULL, esType = NA, var.lor = NA, CIcalc = TRUE, CL = 0.95,
  rID = 0, q = 1, alternative = "two", keepDirection = TRUE,
  keepSign = TRUE, keepSignNames = c("r", "l.r", "u.r", "fisher.z", "l.z",
  "u.z"))

Arguments

df1

Degrees of freedom

df2

NULL or degrees of freedom of the denominator for the f-distribution.

N

Number of data points used in calculation of test-statistic.

n1

Number of data points in sample 1.

n2

Number of data points in sample 2.

esType

Type of test statistic. One of: "t", "lm.t", "f", "lm.f", "r", "X2", "Z", "lm.Z"

CIcalc

If TRUE (default) the Confidence Interval for the test statistic in x will be calculated using the "Confidence limits for noncentral parameters" functions in package (e.g., for type - "t": conf.limits.nct).

CL

Confidence Limit (default: .95).

rID

Correlation among predictor values in a linear model.

q

Number of predictors in the model.

alternative

Alternative hypothesis (defult = "two").

keepSign

Return effect size with sign of test statistic? (default = TRUE).

keepSignNames

Which effect sizes should keep the sign if keepSign = TRUE? Default is to keep the sign for: "r","l.r","u.r","fisher.z","l.z","u.z".

st

Value(s) of a test statistic.

st

Value(s) of a test statistic.

df1

Degrees of freedom

df2

NULL or degrees of freedom of the denominator for the f-distribution.

N

Number of data points used in calculation of test-statistic.

n1

Number of data points in sample 1.

n2

Number of data points in sample 2.

esType

Type of test statistic. One of: "t", "lm.t", "f", "lm.f", "r", "X2", "Z", "lm.Z"

CIcalc

If TRUE (default) the Confidence Interval for the test statistic in x will be calculated using the "Confidence limits for noncentral parameters" functions in package (e.g., for type - "t": conf.limits.nct).

CL

Confidence Limit (default: .95).

rID

Correlation among predictor values in a linear model.

q

Number of predictors in the model.

alternative

Alternative hypothesis (defult = "two").

keepSign

Return effect size with sign of test statistic? (default = TRUE).

keepSignNames

Which effect sizes should keep the sign if keepSign = TRUE? Default is to keep the sign for: "r","l.r","u.r","fisher.z","l.z","u.z".

Details

The procedure to calculate a variety of effect sizes is as follows:

The procedure to calculate a variety of effect sizes is as follows:

Value

The effect sizes calculated by compute.es corresponding to the test statistic(s), with either meta-analytic, or, exact CI.

The effect sizes calculated by compute.es corresponding to the test statistic(s), with either meta-analytic, or, exact CI.

Note

The prefix "lm" is currently disregarded, but will be implemented in future versions to indicate the test statistic is in fact a fixed factor in a linear model.

The prefix "lm" is currently disregarded, but will be implemented in future versions to indicate the test statistic is in fact a fixed factor in a linear model.

Author(s)

Fred Hasselman (inspired by RP:P function any2r by CHJ Hartgerink)

Fred Hasselman (inspired by RP:P function any2r by CHJ Hartgerink)


ManyLabsOpenScience/manylabRs documentation built on April 12, 2018, 8:22 p.m.