Glass Research Raw Data

gistfile1.txt

* For the (6332 + 1026) Google+ postings whose authors had first names assigned to a gender group, TK% (6332/(6332 + 1026)) were guessed to be male and TK% (1026/(6332 + 1026)) female. Including 517 uncertain cases, which somewhat generously throw out anything guessed to be less than 95% probability, the numbers water down to TK% male (6332/(517 + 6332 + 1026)) and TK% (1026/(517 + 6332 + 1026)) female. Among the uncertain cases, TK% (396/517) where thought to be 'likely male' and TK% (121/517) 'likely female'. TK% (2907/(6332 + 1026 + 2907 + 517)) of the overall postings were deemed ambiguous.

array(8) {
  ["all"]=>
  int(10782)
  ["male"]=>
  int(6332)
  ["gendered"]=>
  int(10265)
  ["likely-male"]=>
  int(396)
  ["uncertain"]=>
  int(517)
  ["ambiguous"]=>
  int(2907)
  ["female"]=>
  int(1026)
  ["likely-female"]=>
  int(121)
}


* For the (8407 + 2117) Twitter tweets whose authors had first names assigned to a gender group, TK% (8407/(8407 + 2117)) were guessed to be male and TK% (2117/(8407 + 2117)) female. Including 807 uncertain cases, the numbers adjust to TK% male (8407/(807 + 8407 + 2117)) and TK% (2117/(807 + 8407 + 2117)) female. Among the uncertain cases, TK% (581/807) where thought to be 'likely male' and TK% (226/807) 'likely female'. TK% (10353/(10353 + 8407 + 2117 + 807)) of the overall postings were deemed ambiguous.

array(8) {
  ["all"]=>
  int(21684)
  ["male"]=>
  int(8407)
  ["gendered"]=>
  int(20877)
  ["female"]=>
  int(2117)
  ["ambiguous"]=>
  int(10353)
  ["likely-male"]=>
  int(581)
  ["uncertain"]=>
  int(807)
  ["likely-female"]=>
  int(226)
}