R/calculate-effect-size.R
In sarahyh/ckm-opencpu: CKM meta-analysis
Defines functions
calculateEffectSize
calculateEffectSize <- function(jsonData) {
library(tidyverse)
library(jsonlite)
# Read in data that users extract for each study.
data <- fromJSON(jsonData)
# data <- read_json("testFromExtraction.json", simplifyVector = TRUE)
# Force R to handle all numerical data as doubles to prevent type errors.
data <- data %>% mutate_if(is.integer, as.numeric) %>% mutate_if(~all(is.na(.)), as.numeric)
# helper function to add empty columns to df if they don't already exists
add_col_if_not_exist <- function(df, ...) {
args <- ensyms(...)
for (arg in args) {
if (is.null(df[[toString(arg)]])) {
df <- df %>% mutate(!!arg := NA)
}
}
return(df)
}
# Fill in dataframe with the summary statistics we need for meta-analysis.
# Handle cases differently depending on outcomeType and how information was reported:
# Sample size is relevant to all types of outcomes.
# Add columns if we need to.
data <- data %>% add_col_if_not_exist(N, nExp, nCtrl)
# Do we have separate sample sizes for our two groups, or do we need to assume a balanced design?
data <- data %>%
mutate(
N = case_when(
is.na(N) & !is.na(nExp) & !is.na(nCtrl) ~ as.numeric(nExp + nCtrl), # calculate overall sample size
TRUE ~ as.numeric(N)),
nExp = case_when(
is.na(nExp) & !is.na(nCtrl) & !is.na(N) ~ as.numeric(N - nCtrl), # calculate treatment group sample size
is.na(nExp) & is.na(nCtrl) & !is.na(N) ~ as.numeric(floor(N / 2)), # assume a balanced design
TRUE ~ as.numeric(nExp)),
nCtrl = case_when(
is.na(nCtrl) & !is.na(nExp) & !is.na(N) ~ as.numeric(N - nExp), # calculate control group sample size
is.na(nCtrl) & is.na(nExp) & !is.na(N) ~ as.numeric(floor(N / 2)), # assume a balanced design
TRUE ~ as.numeric(nCtrl))
)
if (any(data[["outcomeType"]] == "continuous")) {
# For continuous outcomes...
# Add columns if we need to.
data <- data %>% add_col_if_not_exist(meanDiff, meanExp, meanCtrl, tValue, pValue, nTails, fValue, stdErrMeanDiff, lowerBoundCI, upperBoundCI, confLevelCI, sdDiff, sdExp, sdCtrl, r, SMD, stdErrSMD, pbCor, stdErrPbCor)
# Calculate untransformed mean difference if we don't already have it.
data <- data %>%
mutate(
meanDiff = case_when(
is.na(meanDiff) & !is.na(meanExp) & !is.na(meanCtrl) ~ meanExp - meanCtrl,
TRUE ~ as.numeric(meanDiff))
)
# Calculate the standard error and pooled standard deviation with whatever information was extracted.
data <- data %>%
mutate(
tValue = case_when(
is.na(tValue) & !is.na(pValue) & !is.na(nTails) ~ qt(p = pValue / nTails, df = N - 2),
is.na(tValue) & !is.na(fValue) ~ sqrt(fValue),
TRUE ~ as.numeric(tValue)),
stdErrMeanDiff = case_when( # cases where we calculate stdErrMeanDiff from test statistics, etc.
is.na(stdErrMeanDiff) & !is.na(lowerBoundCI) & !is.na(upperBoundCI) & !is.na(confLevelCI) ~ 0.5 * (upperBoundCI - lowerBoundCI) / abs(qt(p = (confLevelCI + (100 - confLevelCI) / 2) / 100, df = N - 2)),
is.na(stdErrMeanDiff) & !is.na(meanDiff) & !is.na(tValue) ~ abs(meanDiff / tValue),
TRUE ~ as.numeric(stdErrMeanDiff)),
sdDiff = case_when(
is.na(sdDiff) & !is.na(sdExp) & !is.na(sdCtrl) ~ sqrt(((nExp - 1) * sdExp^2 + (nCtrl - 1) * sdCtrl^2) / (N - 2)),
is.na(sdDiff) & !is.na(stdErrMeanDiff) ~ stdErrMeanDiff / sqrt(1 / nExp + 1 / nCtrl),
TRUE ~ as.numeric(sdDiff)),
stdErrMeanDiff = case_when( # cases where we calculate stdErrMeanDiff from sdDiff and not the other way around
is.na(stdErrMeanDiff) & !is.na(r) & !is.na(sdDiff) ~ sqrt(2 * sdDiff * (1 - r) / N), # accounting for correlation (r) across timepoints for within-subjects contrasts
is.na(stdErrMeanDiff) & !is.na(sdDiff) ~ sdDiff * sqrt(1 / nExp + 1 / nCtrl),
TRUE ~ as.numeric(stdErrMeanDiff))
)
# Now, calculate standardized mean difference and point biserial correlation.
data <- data %>%
mutate(
SMD = case_when(
is.na(SMD) & !is.na(meanDiff) & !is.na(sdDiff) & !is.na(r) ~ meanDiff / sdDiff, # no correction for within-subjects measures (indicated by presence of r)
is.na(SMD) & !is.na(meanDiff) & !is.na(sdDiff) ~ meanDiff / sdDiff * (1 - 3 / (4 * (nExp + nCtrl) - 9)), # apply Hedges' correction
TRUE ~ as.numeric(SMD)),
stdErrSMD = case_when(
is.na(stdErrSMD) & !is.na(SMD) & !is.na(r) ~ sqrt(2 * (1 - r) / N + SMD^2 / (2 * N)), # accounting for correlation (r) across timepoints for within-subjects contrasts
is.na(stdErrSMD) & !is.na(SMD) ~ sqrt((nExp + nCtrl) / (nExp * nCtrl) + SMD^2 / (2 * (nExp + nCtrl))),
TRUE ~ as.numeric(stdErrSMD)),
pbCor = case_when(
is.na(pbCor) & !is.na(SMD) ~ SMD / sqrt(1 / ((nExp / N) * (nCtrl / N)) + SMD^2),
TRUE ~ as.numeric(pbCor)),
stdErrPbCor = case_when(
is.na(stdErrPbCor) & !is.na(pbCor) ~ sqrt((1 - pbCor^2)^2 * (N * pbCor^2 / (4 * nExp * nCtrl) + (2 - 3 * pbCor^2) / (2 * N))),
TRUE ~ as.numeric(stdErrPbCor))
)
}
if (any(data[["outcomeType"]] == "dichotomous")) {
# For dichotomous outcomes...
# Add columns if we need to.
data <- data %>% add_col_if_not_exist(pExp, countExp, pCtrl, countCtrl, riskDiff, stdErrRiskDiff, logRiskRatio, stdErrLogRiskRatio, logOddsRatio, stdErrLogOddsRatio, arcsineRiskDiff, stdErrArcsineRiskDiff)
# Calculate proportions and counts.
data <- data %>%
mutate(
pExp = case_when(
is.na(pExp) & !is.na(countExp) ~ countExp / nExp,
TRUE ~ as.numeric(pExp)),
pCtrl = case_when(
is.na(pCtrl) & !is.na(countCtrl) ~ countCtrl / nCtrl,
TRUE ~ as.numeric(pCtrl)),
countExp = case_when(
is.na(countExp) & !is.na(pExp) ~ floor(pExp * nExp),
TRUE ~ as.numeric(countExp)),
countCtrl = case_when(
is.na(countCtrl) & !is.na(pCtrl) ~ floor(pCtrl * nCtrl),
TRUE ~ as.numeric(countCtrl))
)
# Calculate untransformed risk difference and stdErrRiskDiff (default original units).
data <- data %>%
mutate(
riskDiff = case_when(
is.na(riskDiff) & !is.na(pExp) & !is.na(pCtrl) ~ pExp - pCtrl,
TRUE ~ as.numeric(riskDiff)),
stdErrRiskDiff = case_when(
is.na(stdErrRiskDiff) & !is.na(pExp) & !is.na(pCtrl) ~ sqrt(pExp * (1 - pExp) / nExp + pCtrl * (1 - pCtrl) / nCtrl),
TRUE ~ as.numeric(stdErrRiskDiff))
)
# Calculate standardized effect size as log risk ratio, log odds ratio, and arcsine risk difference
data <- data %>%
mutate(
logRiskRatio = case_when(
is.na(logRiskRatio) & !is.na(pExp) & !is.na(pCtrl) ~ log(pExp / pCtrl),
TRUE ~ as.numeric(logRiskRatio)),
stdErrLogRiskRatio = case_when(
is.na(stdErrLogRiskRatio) & !is.na(countExp) & !is.na(countCtrl) ~ sqrt((nExp - countExp) / (countExp * nExp) + (nCtrl - countCtrl) / (countCtrl * nCtrl)),
TRUE ~ as.numeric(stdErrLogRiskRatio)),
logOddsRatio = case_when(
is.na(logOddsRatio) & !is.na(pExp) & !is.na(pCtrl) ~ log((pExp / (1 - pExp)) / (pCtrl / (1 - pCtrl))),
TRUE ~ as.numeric(logOddsRatio)),
stdErrLogOddsRatio = case_when(
is.na(stdErrLogOddsRatio) & !is.na(countExp) & !is.na(countCtrl) ~ sqrt(1 / countExp + 1 / countCtrl + 1 / (N - countExp) + 1 / (N - countCtrl)),
TRUE ~ as.numeric(stdErrLogOddsRatio)),
arcsineRiskDiff = case_when(
is.na(arcsineRiskDiff) & !is.na(pExp) & !is.na(pCtrl) ~ asin(sqrt(pExp)) - asin(sqrt(pCtrl)),
TRUE ~ as.numeric(arcsineRiskDiff)),
stdErrArcsineRiskDiff = case_when(
is.na(stdErrArcsineRiskDiff) & !is.na(arcsineRiskDiff) ~ sqrt(1 / (4 * nExp) + 1 / (4 * nCtrl)),
TRUE ~ as.numeric(stdErrArcsineRiskDiff))
)
}
# if (any(data[["outcomeType"]] == "association")) {
# # For outcomes measured as correlations
# # Add columns if we need to.
# data <- data %>% add_col_if_not_exist(pearsonCor, stdErrPearsonCor, pbCor, stdErrPbCor, zCor, stdErrZCor, SMD, stdErrSMD)
#
# # Calculate standardized effect size as Pearson correlation, point-biserial correlation, Fisher's Z transformed correlation, and approximate SMD
# data <- data %>%
# mutate(
# # should already have pearsonCor
# stdErrPearsonCor = case_when(
# is.na(stdErrPearsonCor) & !is.na(pearsonCor) ~ sqrt((1 - pearsonCor^2) / (N - 2)),
# TRUE ~ as.numeric(stdErrPearsonCor)),
# # should already have pbCor
# stdErrPbCor = case_when(
# is.na(stdErrPbCor) & !is.na(pbCor) ~ sqrt((1 - pbCor^2)^2 * (N * pbCor^2 / (4 * nExp * nCtrl) + (2 - 3 * pbCor^2) / (2 * N))),
# TRUE ~ as.numeric(stdErrPbCor)),
# zCor = case_when(
# is.na(zCor) & !is.na(pearsonCor) ~ 1 / 2 * log((1 + pearsonCor) / (1 - pearsonCor)),
# TRUE ~ as.numeric(zCor)),
# stdErrZCor = case_when(
# is.na(stdErrZCor) & !is.na(zCor) ~ sqrt(1 / (N - 3)),
# TRUE ~ as.numeric(stdErrZCor)),
# SMD = case_when(
# is.na(SMD) & !is.na(pbCor) ~ pbCor / sqrt((nExp / N) * (nCtrl / N) * (1 - pbCor^2)),
# TRUE ~ as.numeric(SMD)),
# stdErrSMD = case_when(
# is.na(stdErrSMD) & !is.na(SMD) ~ sqrt((nExp + nCtrl) / (nExp * nCtrl) + SMD^2 / (2 * (nExp + nCtrl))),
# TRUE ~ as.numeric(stdErrSMD))
# )
# }
# Drop columns with nothing in them.
data <- data[,colSums(is.na(data)) < nrow(data)]
return(list(message = "success", data = toJSON(data)))
}
# Input variations to test.
# continuous:
# from sd per group
# from sd of difference (pooled standard deviation)
# from se of difference
# from p-value
# from t-value
# from f-value
# from confidence interval
# requires mean difference or group means
# requires sample size overall or per group
# requires repeated measures correlation for within-subjects
# dichotomous:
# from counts per group
# from proportions per group
# requires sample size overall or per group
# association:
# requires correlation coefficient (pearson or pb)
# requires sample size overall or per group
sarahyh/ckm-opencpu documentation built on Aug. 31, 2021, 8:47 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions calculateEffectSize
calculateEffectSize <- function(jsonData) {
library(tidyverse)
library(jsonlite)
# Read in data that users extract for each study.
data <- fromJSON(jsonData)
# data <- read_json("testFromExtraction.json", simplifyVector = TRUE)
# Force R to handle all numerical data as doubles to prevent type errors.
data <- data %>% mutate_if(is.integer, as.numeric) %>% mutate_if(~all(is.na(.)), as.numeric)
# helper function to add empty columns to df if they don't already exists
add_col_if_not_exist <- function(df, ...) {
args <- ensyms(...)
for (arg in args) {
if (is.null(df[[toString(arg)]])) {
df <- df %>% mutate(!!arg := NA)
}
}
return(df)
}
# Fill in dataframe with the summary statistics we need for meta-analysis.
# Handle cases differently depending on outcomeType and how information was reported:
# Sample size is relevant to all types of outcomes.
# Add columns if we need to.
data <- data %>% add_col_if_not_exist(N, nExp, nCtrl)
# Do we have separate sample sizes for our two groups, or do we need to assume a balanced design?
data <- data %>%
mutate(
N = case_when(
is.na(N) & !is.na(nExp) & !is.na(nCtrl) ~ as.numeric(nExp + nCtrl), # calculate overall sample size
TRUE ~ as.numeric(N)),
nExp = case_when(
is.na(nExp) & !is.na(nCtrl) & !is.na(N) ~ as.numeric(N - nCtrl), # calculate treatment group sample size
is.na(nExp) & is.na(nCtrl) & !is.na(N) ~ as.numeric(floor(N / 2)), # assume a balanced design
TRUE ~ as.numeric(nExp)),
nCtrl = case_when(
is.na(nCtrl) & !is.na(nExp) & !is.na(N) ~ as.numeric(N - nExp), # calculate control group sample size
is.na(nCtrl) & is.na(nExp) & !is.na(N) ~ as.numeric(floor(N / 2)), # assume a balanced design
TRUE ~ as.numeric(nCtrl))
)
if (any(data[["outcomeType"]] == "continuous")) {
# For continuous outcomes...
# Add columns if we need to.
data <- data %>% add_col_if_not_exist(meanDiff, meanExp, meanCtrl, tValue, pValue, nTails, fValue, stdErrMeanDiff, lowerBoundCI, upperBoundCI, confLevelCI, sdDiff, sdExp, sdCtrl, r, SMD, stdErrSMD, pbCor, stdErrPbCor)
# Calculate untransformed mean difference if we don't already have it.
data <- data %>%
mutate(
meanDiff = case_when(
is.na(meanDiff) & !is.na(meanExp) & !is.na(meanCtrl) ~ meanExp - meanCtrl,
TRUE ~ as.numeric(meanDiff))
)
# Calculate the standard error and pooled standard deviation with whatever information was extracted.
data <- data %>%
mutate(
tValue = case_when(
is.na(tValue) & !is.na(pValue) & !is.na(nTails) ~ qt(p = pValue / nTails, df = N - 2),
is.na(tValue) & !is.na(fValue) ~ sqrt(fValue),
TRUE ~ as.numeric(tValue)),
stdErrMeanDiff = case_when( # cases where we calculate stdErrMeanDiff from test statistics, etc.
is.na(stdErrMeanDiff) & !is.na(lowerBoundCI) & !is.na(upperBoundCI) & !is.na(confLevelCI) ~ 0.5 * (upperBoundCI - lowerBoundCI) / abs(qt(p = (confLevelCI + (100 - confLevelCI) / 2) / 100, df = N - 2)),
is.na(stdErrMeanDiff) & !is.na(meanDiff) & !is.na(tValue) ~ abs(meanDiff / tValue),
TRUE ~ as.numeric(stdErrMeanDiff)),
sdDiff = case_when(
is.na(sdDiff) & !is.na(sdExp) & !is.na(sdCtrl) ~ sqrt(((nExp - 1) * sdExp^2 + (nCtrl - 1) * sdCtrl^2) / (N - 2)),
is.na(sdDiff) & !is.na(stdErrMeanDiff) ~ stdErrMeanDiff / sqrt(1 / nExp + 1 / nCtrl),
TRUE ~ as.numeric(sdDiff)),
stdErrMeanDiff = case_when( # cases where we calculate stdErrMeanDiff from sdDiff and not the other way around
is.na(stdErrMeanDiff) & !is.na(r) & !is.na(sdDiff) ~ sqrt(2 * sdDiff * (1 - r) / N), # accounting for correlation (r) across timepoints for within-subjects contrasts
is.na(stdErrMeanDiff) & !is.na(sdDiff) ~ sdDiff * sqrt(1 / nExp + 1 / nCtrl),
TRUE ~ as.numeric(stdErrMeanDiff))
)
# Now, calculate standardized mean difference and point biserial correlation.
data <- data %>%
mutate(
SMD = case_when(
is.na(SMD) & !is.na(meanDiff) & !is.na(sdDiff) & !is.na(r) ~ meanDiff / sdDiff, # no correction for within-subjects measures (indicated by presence of r)
is.na(SMD) & !is.na(meanDiff) & !is.na(sdDiff) ~ meanDiff / sdDiff * (1 - 3 / (4 * (nExp + nCtrl) - 9)), # apply Hedges' correction
TRUE ~ as.numeric(SMD)),
stdErrSMD = case_when(
is.na(stdErrSMD) & !is.na(SMD) & !is.na(r) ~ sqrt(2 * (1 - r) / N + SMD^2 / (2 * N)), # accounting for correlation (r) across timepoints for within-subjects contrasts
is.na(stdErrSMD) & !is.na(SMD) ~ sqrt((nExp + nCtrl) / (nExp * nCtrl) + SMD^2 / (2 * (nExp + nCtrl))),
TRUE ~ as.numeric(stdErrSMD)),
pbCor = case_when(
is.na(pbCor) & !is.na(SMD) ~ SMD / sqrt(1 / ((nExp / N) * (nCtrl / N)) + SMD^2),
TRUE ~ as.numeric(pbCor)),
stdErrPbCor = case_when(
is.na(stdErrPbCor) & !is.na(pbCor) ~ sqrt((1 - pbCor^2)^2 * (N * pbCor^2 / (4 * nExp * nCtrl) + (2 - 3 * pbCor^2) / (2 * N))),
TRUE ~ as.numeric(stdErrPbCor))
)
}
if (any(data[["outcomeType"]] == "dichotomous")) {
# For dichotomous outcomes...
# Add columns if we need to.
data <- data %>% add_col_if_not_exist(pExp, countExp, pCtrl, countCtrl, riskDiff, stdErrRiskDiff, logRiskRatio, stdErrLogRiskRatio, logOddsRatio, stdErrLogOddsRatio, arcsineRiskDiff, stdErrArcsineRiskDiff)
# Calculate proportions and counts.
data <- data %>%
mutate(
pExp = case_when(
is.na(pExp) & !is.na(countExp) ~ countExp / nExp,
TRUE ~ as.numeric(pExp)),
pCtrl = case_when(
is.na(pCtrl) & !is.na(countCtrl) ~ countCtrl / nCtrl,
TRUE ~ as.numeric(pCtrl)),
countExp = case_when(
is.na(countExp) & !is.na(pExp) ~ floor(pExp * nExp),
TRUE ~ as.numeric(countExp)),
countCtrl = case_when(
is.na(countCtrl) & !is.na(pCtrl) ~ floor(pCtrl * nCtrl),
TRUE ~ as.numeric(countCtrl))
)
# Calculate untransformed risk difference and stdErrRiskDiff (default original units).
data <- data %>%
mutate(
riskDiff = case_when(
is.na(riskDiff) & !is.na(pExp) & !is.na(pCtrl) ~ pExp - pCtrl,
TRUE ~ as.numeric(riskDiff)),
stdErrRiskDiff = case_when(
is.na(stdErrRiskDiff) & !is.na(pExp) & !is.na(pCtrl) ~ sqrt(pExp * (1 - pExp) / nExp + pCtrl * (1 - pCtrl) / nCtrl),
TRUE ~ as.numeric(stdErrRiskDiff))
)
# Calculate standardized effect size as log risk ratio, log odds ratio, and arcsine risk difference
data <- data %>%
mutate(
logRiskRatio = case_when(
is.na(logRiskRatio) & !is.na(pExp) & !is.na(pCtrl) ~ log(pExp / pCtrl),
TRUE ~ as.numeric(logRiskRatio)),
stdErrLogRiskRatio = case_when(
is.na(stdErrLogRiskRatio) & !is.na(countExp) & !is.na(countCtrl) ~ sqrt((nExp - countExp) / (countExp * nExp) + (nCtrl - countCtrl) / (countCtrl * nCtrl)),
TRUE ~ as.numeric(stdErrLogRiskRatio)),
logOddsRatio = case_when(
is.na(logOddsRatio) & !is.na(pExp) & !is.na(pCtrl) ~ log((pExp / (1 - pExp)) / (pCtrl / (1 - pCtrl))),
TRUE ~ as.numeric(logOddsRatio)),
stdErrLogOddsRatio = case_when(
is.na(stdErrLogOddsRatio) & !is.na(countExp) & !is.na(countCtrl) ~ sqrt(1 / countExp + 1 / countCtrl + 1 / (N - countExp) + 1 / (N - countCtrl)),
TRUE ~ as.numeric(stdErrLogOddsRatio)),
arcsineRiskDiff = case_when(
is.na(arcsineRiskDiff) & !is.na(pExp) & !is.na(pCtrl) ~ asin(sqrt(pExp)) - asin(sqrt(pCtrl)),
TRUE ~ as.numeric(arcsineRiskDiff)),
stdErrArcsineRiskDiff = case_when(
is.na(stdErrArcsineRiskDiff) & !is.na(arcsineRiskDiff) ~ sqrt(1 / (4 * nExp) + 1 / (4 * nCtrl)),
TRUE ~ as.numeric(stdErrArcsineRiskDiff))
)
}
# if (any(data[["outcomeType"]] == "association")) {
# # For outcomes measured as correlations
# # Add columns if we need to.
# data <- data %>% add_col_if_not_exist(pearsonCor, stdErrPearsonCor, pbCor, stdErrPbCor, zCor, stdErrZCor, SMD, stdErrSMD)
#
# # Calculate standardized effect size as Pearson correlation, point-biserial correlation, Fisher's Z transformed correlation, and approximate SMD
# data <- data %>%
# mutate(
# # should already have pearsonCor
# stdErrPearsonCor = case_when(
# is.na(stdErrPearsonCor) & !is.na(pearsonCor) ~ sqrt((1 - pearsonCor^2) / (N - 2)),
# TRUE ~ as.numeric(stdErrPearsonCor)),
# # should already have pbCor
# stdErrPbCor = case_when(
# is.na(stdErrPbCor) & !is.na(pbCor) ~ sqrt((1 - pbCor^2)^2 * (N * pbCor^2 / (4 * nExp * nCtrl) + (2 - 3 * pbCor^2) / (2 * N))),
# TRUE ~ as.numeric(stdErrPbCor)),
# zCor = case_when(
# is.na(zCor) & !is.na(pearsonCor) ~ 1 / 2 * log((1 + pearsonCor) / (1 - pearsonCor)),
# TRUE ~ as.numeric(zCor)),
# stdErrZCor = case_when(
# is.na(stdErrZCor) & !is.na(zCor) ~ sqrt(1 / (N - 3)),
# TRUE ~ as.numeric(stdErrZCor)),
# SMD = case_when(
# is.na(SMD) & !is.na(pbCor) ~ pbCor / sqrt((nExp / N) * (nCtrl / N) * (1 - pbCor^2)),
# TRUE ~ as.numeric(SMD)),
# stdErrSMD = case_when(
# is.na(stdErrSMD) & !is.na(SMD) ~ sqrt((nExp + nCtrl) / (nExp * nCtrl) + SMD^2 / (2 * (nExp + nCtrl))),
# TRUE ~ as.numeric(stdErrSMD))
# )
# }
# Drop columns with nothing in them.
data <- data[,colSums(is.na(data)) < nrow(data)]
return(list(message = "success", data = toJSON(data)))
}
# Input variations to test.
# continuous:
# from sd per group
# from sd of difference (pooled standard deviation)
# from se of difference
# from p-value
# from t-value
# from f-value
# from confidence interval
# requires mean difference or group means
# requires sample size overall or per group
# requires repeated measures correlation for within-subjects
# dichotomous:
# from counts per group
# from proportions per group
# requires sample size overall or per group
# association:
# requires correlation coefficient (pearson or pb)
# requires sample size overall or per group
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.