test_means_within: Test two means against each other for a "within-subjects"...
In darrellpenta/uxstats: Common data analysis tools for UX researchers

Description Usage Arguments Details Value See Also Examples

View source: R/test_means_within-function.R

test_means_within() compares two sample means obtained from the same group of users and performs a two-sided statistical test to determine if the means differ from each other.

test_means_within, ab_test_within, and test_times_within are synonyms.

test_means_within(.x, ...)

ab_test_within(.x, ...)

## S3 method for class 'numeric'
test_means_within(.x, .sd, .n, ..., .alpha = 0.05)

## S3 method for class 'data.frame'
test_means_within(.x, .var1, .var2, ..., .alpha = 0.05)

`.x`	A single numeric value (i.e., the difference of two means) or a long-format data frame with named columns of numeric data corresponding to the variables being tested. See Details.
`...`	(Optional) If `.x` is a long-format data frame, you can pass the name of one or more grouping variables as unquoted, comma-separated column names (without naming the `...` argument) to compute stats by groups.
`.sd`	If `.x` is a single numeric value representing the mean of the variable differences, `.sd` should be a single numeric value representing the standard deviation of the diferences. See Details.
`.n`	If `.x` is a single numeric value representing the mean of the variable differences, `.sd` should be a single numeric value representing the total number of trials. See Details.
`.alpha`	(Optional) A positive number (where 0 < `.alpha` < 1) specifying the significance level to be used. Defaults to `.alpha = 0.05`. To set a different significance level, the argument must be named (i.e., `.alpha=0.001`) or else the function may yield unexpected results.
`.var1`	If `.x` is a long-format data frame, the (unquoted) name of a data frame column containing the values representing the first sample.
`.var2`	If `.x` is a long-format data frame, the (unquoted) name of a data frame column containing the values representing the second sample.

If .x is a single numeric value representing the mean of differences from the sample scores, .sd should be a single numeric value representing the standard deviation of the difference scores.
If .x is a single numeric value representing the mean of differences from the sample scores, .n should be a single numeric value representing the number of pairs that derived the difference scores.
If .x is a data frame, .var1 and .var2 should be the unquoted names of the columns containing the raw scores obtained.
You can modify the alpha level to adjust confidence intervals by including .alpha as a named argument and providing a numeric value: e.g., .aplha = 0.001.
If you're passing a data frame to .x, you can optionally pass one or more grouping variables as unquoted, comma-separated column names (without naming the ... argument) to compute stats by groups.

Note that NAs are automatically dropped in all calculations.

A tibble with the results of a test of two means, with confidence interval information, and other information.

Other tests comparing means: test_means_between()

# When you have computed difference metrics in advance:
test_means_within(45,9.2,18)

.ux_data <-
 data.frame(
  "id" = rep(seq(1,20,1),2),
  "task" = c(rep(1,20),rep(2,20)),
  "score1"  = sample(1:7,40,TRUE),
  "score2"  = sample(1.5:7.5,40,TRUE)
  )

test_means_within(.ux_data, score1, score2, task, .alpha=0.1)