Skip to contents

Given the results of the splitting and testing algorithm, report on the blocks where the null of no effects could be rejected at level alpha. Currently calculates rejections using an FWER style criteria (p of a node = max of all previous nodes) if the final alphas are all the same as the scalar alpha OR if fwer=TRUE.

Usage

report_detections(
  orig_res,
  fwer = TRUE,
  alpha = 0.05,
  only_hits = FALSE,
  blockid = "blockF"
)

Arguments

orig_res

results data.table output from the findBlocks function.

fwer

(default is TRUE) means that a block is detected (or not) using the maximum p-value associated with the block (or the groups containing that block). fwer=FALSE to detect blocks (or groups of blocks) using FDR control.

alpha

Is the false positive rate used for detecting an effect if it is constant (i.e. not an FDR-style approach).

only_hits

(default FALSE) returns only the detected blocks instead of all of them

blockid

Name of block variable (the blocking variable is a factor)

Value

A data.table adding a column hit to the res data.table indicating a "hit" or detection for that block (or group of blocks)