Skip to main content

Table 5 Post-hoc comparison of reporting in publications, and changes in reporting from abstract to publication, according to journal impact factor

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

Variable

Prevalence in lower impact n = 32

Prevalence in higher impact n = 30

p value of comparison

Change from abstract (A) to publication (P)

Randomized

12 (38%)

12 (40%)

0.99

 

 Change from A to P

8 (25%)

4 (13%)

0.25

All 12: from non-R in A, to R in P

 Method

0 (0%)

0 (0%)

–

–

 Allocation concealment

0 (0%)

0 (0%)

–

–

Blinding (possible)

0 (0%)

4 (13%)

0.049

 

 Change from A to P

0 (0%)

4 (13%)

0.049

All 4: from no mention in A, to blinding of some outcomes in P

Sample size calculation

0 (0%)

0 (0%)

–

–

Primary outcome stated

3 (9%)

0 (0%)

0.09

 

 Change from A to P

3 (9%)

6 (20%)

0.24

All 9: state in A, to not stated in P

Numbers with denominators

8 (25%)

13 (43%)

0.13

 

 Change from A to P

6 (19%)

10 (33%)

0.19

All 16: no denominators in A, to denominators in P

Main outcomes positive

32 (100%)

30 (100%)

–

 

 Change from A to P

4 (13%)

1 (3%)

0.19

All 5: negative in A, to positive or excluded in P

Number of animals stateda

17 (53%)

18 (60%)

0.59

 

 Change from A to P

6 (19%)

7 (23%)

0.66

From A to P the number was smaller in 3 (9%) and larger in 10 (29%)a

Statistically significant result of main outcomes

30 (94%)

28 (93%)

0.95

 

 Change from A to P

13 (41%)

10 (33%)

0.55

All 23: not significant (or not stated) in A, to significant in Pb

Months to publication

27 (SD 17)

18 (SD 16)

0.02

–

  1. Comparisons made using Fisher’s Exact or Chi square test, or independent samples student t test, as appropriate
  2. A abstract, P publication, R randomized
  3. a In the 13 that changed in animal numbers from A to P: in the lower and higher impact P the number was smaller in 1 (by n = 4) and 2 (by n = 3 and 6), and larger in 5 (by n = 4, 14, 36, 52, 54) and 5 (by n = 4, 5, 7, 22, 25) respectively
  4. b In the 23 that changed in statistical significance from A to P, in the lower and higher impact P respectively: the animal numbers did not change in 4 and 1 (p = 0.19) [these numbers did change in 2 and 2 (larger number in 2, and smaller number in 2), and change could not be determined in the rest because numbers were not stated in P in 6 and 3, and were not stated in the A in 1 and 4]; the main outcomes changed in 2 and 0 (p = 0.16); and denominators changed in 1 and 2 (p = 0.52) [change could not be determined in 19 others because denominators were not reported in 11 and 8; thus we could be sure that denominators did not change in 1]