report (version 0.6.0)

report_sample: Sample Description

Description

Create sample description table (also referred to as "Table 1").

Usage

report_sample(
  data,
  by = NULL,
  centrality = "mean",
  ci = NULL,
  ci_method = "wilson",
  ci_correct = FALSE,
  select = NULL,
  exclude = NULL,
  weights = NULL,
  total = TRUE,
  digits = 2,
  n = FALSE,
  group_by = NULL,
  ...
)

Value

A data frame of class report_sample with variable names and their related summary statistics.

Arguments

data

A data frame for which descriptive statistics should be created.

by

Character vector, indicating the column(s) for possible grouping of the descriptive table. Note that weighting (see weights) does not work with more than one grouping column.

centrality

Character, indicates the statistics that should be calculated for numeric variables. May be "mean" (for mean and standard deviation) or "median" (for median and median absolute deviation) as summary.

ci

Level of confidence interval for relative frequencies (proportions). If not NULL, confidence intervals are shown for proportions of factor levels.

ci_method

Character, indicating the method how to calculate confidence intervals for proportions. Currently implemented methods are "wald" and "wilson". Note that "wald" can produce intervals outside the plausible range of [0, 1], and thus it is recommended to prefer the "wilson" method. The formulae for the confidence intervals are:

• "wald":

  $$p \pm z \sqrt{\frac{p (1 - p)}{n}}$$

• "wilson":

  $$\frac{2np + z^2 \pm z \sqrt{z^2 + 4npq}}{2(n + z^2)}$$

  where p is the proportion (of a factor level), q is 1-p, z is the critical z-score based on the interval level and n is the length of the vector (cf. Newcombe 1998, Wilson 1927).

ci_correct

Logical, it TRUE, applies continuity correction. See Newcombe 1998 for different correction-methods based on the chosen ci_method.

select

Character vector, with column names that should be included in the descriptive table.

exclude

Character vector, with column names that should be excluded from the descriptive table.

weights

Character vector, indicating the name of a potential weight-variable. Reported descriptive statistics will be weighted by weight.

total

Add a Total column.

digits

Number of decimals.

n

Logical, actual sample size used in the calculation of the reported descriptive statistics (i.e., without the missing values).

group_by

Deprecated. Use by instead.

...

Arguments passed to or from other methods.

References

• Newcombe, R. G. (1998). Two-sided confidence intervals for the single proportion: comparison of seven methods. Statistics in Medicine. 17 (8): 857–872

• Wilson, E. B. (1927). Probable inference, the law of succession, and statistical inference. Journal of the American Statistical Association. 22 (158): 209–212

Examples

library(report)

report_sample(iris[, 1:4])
report_sample(iris, select = c("Sepal.Length", "Petal.Length", "Species"))
report_sample(iris, by = "Species")
report_sample(airquality, by = "Month", n = TRUE, total = FALSE)

# confidence intervals for proportions
set.seed(123)
d <- data.frame(x = factor(sample(letters[1:3], 100, TRUE, c(0.01, 0.39, 0.6))))
report_sample(d, ci = 0.95, ci_method = "wald") # ups, negative CI
report_sample(d, ci = 0.95, ci_method = "wilson") # negative CI fixed
report_sample(d, ci = 0.95, ci_correct = TRUE) # continuity correction