estimate_log_odds: Estimate the log odds ratio and stanard error for counts of...
In wfforrest/maeve: Modeling Accessories Enabling Visualization of Experiments

Description Usage Arguments Value Author(s) References Examples

Estimate the log odds ratio and stanard error for counts of treatment relative to reference

estimate_log_odds(
  count_data_frame,
  group_name_char = "group",
  reference,
  number_on_first_day_char = "N_first_day",
  count_desc_char = c("PR", "EOS_CR", "PR_or_EOS_CR")[1],
  additive_offset = 0.5,
  return_original = TRUE
)

`count_data_frame`	data.frame with one row per group, total number per group, and number of positive counts out of the total
`group_name_char`	character name of column with group factor identifier.
`reference`	character name of level to use as a reference; must be in levels of factor designated by group_name_char.
`number_on_first_day_char`	character name of column with total number in a each group.
`count_desc_char`	character name of column with number of positive cases in each group.
`additive_offset`	numeric value to add to each count in a 2 x 2 table before computing log odds ratio. Default gives the Haldane-Anscombe-Gart correction of adding 0.5 to every count.
`return_original`	logical whether to return the original counts in the table with the estimated log_OR and SE_log_OR.

A data.frame with log odds ratio (log_OR) and its standard error (SE_log_OR) for each non-reference group relative to reference.

Bill Forrest <forrest@gene.com>

Bill Forrest forrest@gene.com

www.r-project.org

 test_log_odds <- data.frame( group = factor( c( 'group_01', 'group_02', 'group_03' ) ),
                              N_first_day = rep( 10, 3 ),
                              PR = c( 0, 2, 6 )
                             )
 log_odds_est <- maeve:::estimate_log_odds( test_log_odds, reference = 'group_01' )