Skip to main content

Table 3 Post-hoc comparison of oral versus poster abstracts and publications

From: Publication bias in animal research presented at the 2008 Society of Critical Care Medicine Conference

Abstract variables Oral (n = 16) Poster (n = 84) p value
Research location in North America 14 (88%) 57 (68%) 0.14
Methodological quality
 Randomized 1 (6%) 26 (31%) 0.06
 Primary outcome given 2 (13%) 24 (29%) 0.23
 Numbers with denominators 0 (0%) 6 (7%) 0.59
Ethical quality
 Highest species rodent 13 (81%) 48 (57%) 0.07
 Number of animals stated 3 (19%) 44 (52%) 0.02
 Number of animals used 28 (SD 17) 29 (SD 12) 0.27
Outcomes
 Main outcomes positive 16 (100%) 74 (88%) 0.36
 Statistically significant result 9 (56%) 46 (55%) 0.95
Type of animal model
 Sepsis 8 (50%) 52 (62%) 0.41
 Drug used 8 (50%) 53 (63%) 0.40
 Surgery 11 (69%) 51 (61%) 0.59
Publication variables Oral published (n = 14) Poster published (n = 48) p value
Methodological quality
 Randomized 1 (7%) 23 (48%) 0.006
 Blinded 0 (0%) 4 (8%) 0.57
 Primary outcome given 0 (0%) 3 (6%) 0.99
 Numbers with denominators 3 (21%) 18 (38%) 0.35
Ethical quality
 Number of animals stated 4 (29%) 31 (65%) 0.03
 Number of animals used 46 (SD 55) 26 (SD 19) 0.53
Outcomes
 Main outcomes positive 14 (100%) 48 (100%) 0.99
 Statistically significant result 13 (93%) 45 (94%) 0.99
Journal factors
 Journal impact factor 5.1 (SD 3.2) 5.3 (SD 5.3) 0.92
 Months to publication 31 (SD 23) 20 (SD 14) 0.12
  1. Comparisons made using Fisher’s Exact or Chi square test, or independent samples student t test, as appropriate