MIT Press

Table 6.

Results (percentage change coefficients) of 45 Poisson regressions with tags as dependent variables and different variants measuring disruption as independent variables each. The models have been adjusted for exposure time (different publication years): How long have the papers been at risk of being tagged and cited?

Tag	DI₁	DI₅	${D I}_{1}^{n o k}$	${D I}_{5}^{n o k}$	DeIn
Tags expressing newness (expecting positive signs)
Hypothesis	4.32	3.01	4.66	2.75	0.25
New finding	−2.71	−0.62	−2.13	−2.13	1.92
Novel drug target	6.89	6.74	14.85	15.19	−7.91
Technical advance	6.72	20.65	18.34	19.80	−18.11
Tags not expressing newness (expecting negative signs)
Confirmation	−5.08	−0.41	0.08	1.45	−1.88
Good for teaching	12.24	0.37	8.82	3.73	6.41
Negative/Null results	3.45	−11.46	−13.35	−5.20	−2.10
Refutation	−9.28	−9.59	−14.83	−10.37	8.06
Tag without expectations
Controversial	−0.33	4.42	1.46	1.46	−3.62

Tag	DI₁	DI₅	${D I}_{1}^{n o k}$	${D I}_{5}^{n o k}$	DeIn
Tags expressing newness (expecting positive signs)
Hypothesis	4.32	3.01	4.66	2.75	0.25
New finding	−2.71	−0.62	−2.13	−2.13	1.92
Novel drug target	6.89	6.74	14.85	15.19	−7.91
Technical advance	6.72	20.65	18.34	19.80	−18.11
Tags not expressing newness (expecting negative signs)
Confirmation	−5.08	−0.41	0.08	1.45	−1.88
Good for teaching	12.24	0.37	8.82	3.73	6.41
Negative/Null results	3.45	−11.46	−13.35	−5.20	−2.10
Refutation	−9.28	−9.59	−14.83	−10.37	8.06
Tag without expectations
Controversial	−0.33	4.42	1.46	1.46	−3.62

Note: The following abbreviations are used: different indicators measuring disruption (DI₁, DI₅, ${D I}_{1}^{n o k}$ ⁠, ${D I}_{5}^{n o k}$ ⁠, DeIn). Percentage change coefficients in bold are as expected, otherwise the results are not as expected (or, for the tag “Controversial,” there are no clear expectations).

View Large

This Feature Is Available To Subscribers Only