--no_bonferroni

Switch

--no_bonferroni

Description

Intended to turn off Bonferroni correction of p-values (setting bonferroni = False), separately from the overall p-value correction method controlled by --p_correction/--no_correction.

Argument and Default Value

No argument. Sets bonferroni to False; default (flag absent) is True.

Details

Known issue — this flag currently does not work at all. Right after argument parsing, dlatkInterface.py runs:

if not args.bonferroni:
  print("--no_bonf has been depricated. Default p correction method is now Benjamini, Hochberg. Please use --no_correction instead of --no_bonf.")
  sys.exit(1)

Since --no_bonferroni sets args.bonferroni to False, this check fires for --no_bonferroni itself, not just the older deprecated --no_bonf alias the message refers to. Verified: any command that includes --no_bonferroni prints this message and exits immediately with status 1, before running anything else in the command — no correlation, no feature extraction, nothing. Use --no_correction instead, which actually works and is the documented replacement.

Other Switches

Required Switches:

None (but see above — do not use this switch)

Example Commands

Do not use --no_bonferroni. Use --no_correction instead:

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id$16to16' \
--outcome_table blog_outcomes --outcomes age --correlate --no_correction