Table 4.
Categories of spin in preprints and Journal publications (n=67 studies)
| Spin categories and subcategories* | No spin N (%) |
Occurred in preprint and journal publication N (%) | Occurred in preprint only N (%) |
Occurred in journal publication only N (%) |
| Any category of spin† | 37 (55%) | 23 (34%) | 5 (7%) | 2 (3%) |
| Category | ||||
| Inappropriate interpretation given study design‡ | 55 (82%) | 7 (10%) | 4 (6%) | 1 (1%) |
| Subcategory | ||||
| Claiming causality in non-randomised studies | 62 (93%) | 4 (6%) | 1 (1%) | 0 (0%) |
| Interpreting a lack of statistical significance as equivalence | 66 (99%) | 0 (0%) | 0 (0%) | 1 (1%) |
| Interpreting a lack of statistical significance of harm measures as safety | 65 (97%) | 1 (1.5%) | 0 (0%) | 1 (1.5%) |
| Claim of any significant difference despite lack of statistical test | 67 (100%) | 0 (0%) | 0 (0%) | 0 (0%) |
| Other | 61 (91%) | 2 (3%) | 4 (6%) | 0 (0%) |
| Category | ||||
| Inappropriate extrapolations or recommendations | 52 (78%) | 13 (19%) | 2 (3%) | 0 (0%) |
| Subcategory | ||||
| Suggestion that the treatment or test is more clinically relevant or useful than is justified given the study design. | 60 (90%) | 6 (9%) | 1 (1%) | 0 (0%) |
| Recommendations made to population groups/contexts outside of those investigated. | 63 (94%) | 3 (5%) | 1 (1%) | 0 (0%) |
| (Observational) Expressing confidence in a treatment or test without suggesting the need for further confirmatory studies | 66 (99%) | 0 (0%) | 1 (1%) | 0 (0%) |
| (Observational) Making recommendations without stating a randomised controlled clinical should be done to validate the recommendation | 65 (97%) | 2 (3%) | 0 (0%) | 0 (0%) |
| Other | 63 (94%) | 3 (5%) | 1 (1%) | 0 (0%) |
| Category | ||||
| Selective focusing on positive results or more favourable data presentation | 54 (81%) | 8 (12%) | 2 (3%) | 3 (4%) |
| Subcategory | ||||
| Discussing only significant (non-primary) results to distract from non-significant (primary results | 66 (99%) | 0 (0%) | 1 (1%) | 0 (0%) |
| Omitting non-significant results from abstract/discussion/conclusion | 65 (97%) | 1 (1.5%) | 0 (0%) | 1 (1.5%) |
| Claiming significant effects for non-significant results | 67 (100%) | 0 (0%) | 0 (0%) | 0 (0%) |
| Acknowledge statistically non-significant results for the primary outcome but emphasise the beneficial effect of treatment | 66 (99%) | 1 (1%) | 0 (0%) | 0 (0%) |
| Describing non-significant results as ‘trending towards significance’ | 66 (99%) | 1 (1%) | 0 (0%) | 0 (0%) |
| Mentioning adverse events in the abstract/discussion/conclusion but minimising their potential effect or importance. | 64 (96%) | 2 (3%) | 1 (1%) | 0 (0%) |
| Misleading description of study design as one that is more robust | 67 (100%) | 0 (0%) | 0 (0%) | 0 (0%) |
| No considerations of the limitations of the study | 64 (96%) | 3 (4%) | 0 (0%) | 0 (0%) |
| Use of linguistic spin | 66 (99%) | 0 (0%) | 0 (0%) | 1 (1%) |
| Other | 62 (93%) | 1 (1%) | 2 (3%) | 2 (3%) |
*Subcategories of spin are not mutually exclusive; a preprint or journal publications could contain multiple subcategories of spin within a category. Preprints and journal publications could contain different subcategories of spin within a category.
†This row shows counts of at least one instance of spin in any category. Column category and subcategory counts add to greater than any occurrence of spin because multiple categories and subcategories of spin could occur within a preprint or article publication. Row percents do not add to 100 due to rounding.
‡Row percents may not add to 100 due to rounding.