How communication scholars see open scholarship

Figures & data

Table 1. Attitudes towards various open scholarship practices (see Abele-Brehm et al., 2019).

	Factor Loadings
Item	Support	Concern	Trust	M (SD)
Support for OSPs: M = 3.59, SD = .787, ω = .813 Eigenvalue = 3.68, 32.6% variance explained
It should be standard practice that research data are published.	.788			3.14 (1.29)
I can benefit from publishing my research data.	.707			3.43 (1.13)
Sharing data benefits scholarship.	.649			4.11 (.870)
Research should be open and transparent.	.548			4.40 (.832)
The publication of research data will soon be mandatory.	.544			2.87 (1.11)
Concerns over OSPs: M = 2.86, SD = .926, ω = .650 Eigenvalue = 1.61, 11.5% variance explained
I am afraid that I will place myself at a competitive disadvantage if I share my data for later use by others.		.867		2.93 (1.30)
I would gladly share data, but only when most other researchers are required to do the same.		.468		2.88 (1.17)
I am afraid that after publishing my data, others will find errors in my scholarship.		.452		2.79 (1.28)
Trust resulting from OSPs: M = 3.63, SD = 1.05, ρ = .711 Eigenvalue = 1.05, 7.3% variance explained
I trust research findings more when the research data are published.			-.986	3.79 (1.07)
I trust researchers who publish their data more than those who don’t.			-.617	3.48 (1.22)
Items removed from scale
The costs of publishing research data are outweighed by the benefits of publishing them.*			3.19 (1.09)
Publishing research data are generally problematic.*			2.44 (1.05)
Publishing research data are generally superfluous.*			2.36 (1.06)
In hiring decisions, applicants who share their data should be given preference in case of otherwise equivalent qualifications.+			2.34 (1.25)
In the current environment, data sharing is bad for my career.+			2.67 (1.21)

Notes: * = Item removed for weak primary loadings (< .400); + = Item removed for cross-loadings (< .300). Factor loadings less than .400 removed for ease of interpretation. Full analysis output included in supplemental materials.

Table 2. Respondents’ attitudes toward specific osps regarding submission to ICA journals and ICA conferences.

	ICA journals (n = 308)			ICA conference (n = 313)
Items	Component 1 loading	M	SD	Component 1 loading	M	SD
… should provide a preregistration option for manuscripts, where applicable.	.801	3.61	1.16	.842	3.29	1.19
… should require sharing datasets for manuscripts, where applicable.	.783	2.90	1.31	.829	2.31	1.21
… should provide the option for sharing datasets for manuscripts, where applicable.	.781	4.23	.904	.779	3.76	1.16
… should require shared research materials for manuscripts, where applicable.	.773	3.03	1.29	.781	2.44	1.27
… should provide the option for sharing research materials for manuscripts, where applicable.	.740	4.22	.949	.759	3.77	1.18
… should use Open Science Badges	.738	3.69	1.26	.749	3.39	1.32
OVERALL	ω = .835	3.61a	.883	ω = .843	3.17b	.970

Notes: Overall means differ significantly from each other using paired-samples t-tests, with journal OSP scores greater than conference OSP scores. (1 = strongly oppose, 5 = strongly support).

Table 3. Sample-wide frequencies of engaging with OSPs.

	M	SD	one-sample t-test (to 0.00)
sharing research materials online	1.85	1.20	t(317) = 27.5, p < .001, Cohen’s d = 1.54
sharing datasets online	1.05	1.12	t(316) = 16.7, p < .001, Cohen’s d = .935
sharing preprints online	1.54	1.30	t(304) = 20.1, p < .001, Cohen’s d = 1.18
preregistering a study protocol	0.67	1.01	t(316) = 11.7, p < .001, Cohen’s d = .661
submitting a preregistered report	0.30	0.70	t(315) = 7.51, p < .001, Cohen’s d = .423
using others’ shared materials	1.53	1.18	t(320) = 23.2, p < .001, Cohen’s d = 1.30
using others’ shared datasets	1.02	1.10	t(319) = 16.6, p < .001, Cohen’s d = .930
using others’ shared preprints	1.66	1.26	t(316) = 23.6, p < .001, Cohen’s d = 1.32

Table 4. Average engagement with OSPs across career stages.

	Student	Junior	Mid	Senior	ANOVA
sharing research materials online	1.90 (1.16)	2.08 (1.36)	1.92 (1.13)	1.88 (1.16)	F(3,224) = .311, p = .817
sharing datasets online	1.02 (1.15)	1.08 (1.16)	1.10 (1.14)	.93 (1.06)	F(3,222) = .294, p = .830
sharing preprints online	1.75 (1.28)	1.77 (1.29)	1.47 (1.25)	1.53 (1.26)	F(3,214) = .727 p = .537
preregistering a study protocol	.49 (.779)	.58 (.906)	.70 (1.03)	.69 (1.02)	F(3,224) = .548, p = .650
submitting a preregistered report	.41 (.865)	.24 (.635)	.21 (.681)	.34 (.692)	F(3,222) = .896 p = .444
using others’ shared materials	1.56 (1.21)	1.80 (1.11)	1.31 (1.21)	1.49 (1.14)	F(3,225) = 1.68 p = .172
using others’ shared datasets	1.00 (1.18)	1.12 (1.06)	.88 (1.03)	1.03 (1.11)	F(3,225) = .496 p = .685
using others’ shared preprints	1.83 (1.26)	1.82 (1.27)	1.47 (1.24)	1.65 (1.19)	F(3,222) = 1.02 p = .384

Table 5. Influence of methodological and epistemological focus on engagement with OSPs.

	Quantitative – Qualitative			Social Science – Humanities
	B	t	P	B	t	P
sharing research materials online	.005	1.67	.095	-.009	−2.79	.006
sharing datasets online	.004	1.34	1.80	-.008	2.59	.010
sharing preprints online	-.002	-.785	.433	-.005	−1.34	.183
preregistering a study protocol	∼.000	.185	.853	-.004	−1.61	.110
submitting a preregistered report	.001	.312	.756	-.002	−1.27	.205
using others’ shared materials	-.002	-.773	.464	-.001	-.303	.762
using others’ shared datasets	-.002	-.774	.440	-.002	-.583	.560
using others’ shared preprints	∼.000	.025	.980	-.004	−1.33	.186

Note: Bonferroni-corrected critical p-value of .003 is applied to these analyses.

Table 6. Support for potential ICA-specific OSP policies by career progression.

	ICA journal submissions		ICA conference submissions
	M	SD	M	SD
Student	3.60	0.85	3.17	0.92
Junior	3.53	0.93	3.19	1.02
Mid-	3.58	0.88	3.24	0.99
Senior	3.60	0.85	3.10	0.98

Note: Means are not significantly different from one another.

Figure 1. Comparison of attitudes towards open science practices in members of the German Psychological Society (DGP, 2018) and the International Communication Association (ICA, 2020). Note. Plotted scores are means of variables recoded to -2 = do not agree (DGP)/strongly disagree (ICA), -1 = slightly disagree (DGP & ICA), 0 = neither disagree nor agree (ICA), 1 = slightly agree (DGP & ICA), 2 = strongly agree (DGP & ICA). Item 5 (“Publishing research data is generally problematic.”) not included in DGP survey.

Table 7. Common themes associated with OSPs.

Theme	Definition/Interpretation	Freq. (%)
Open Access	References to open access, which referred to both open-access publishing as well as accessibility broadly	211 (68%)
Open Data/Code/Materials	Sharing or making study content freely available, with specific mentions of data, code, or materials	122 (39%)
Transparency & Integrity	Allowing research content to be viewed by anyone interested (broadly, or for verification purposes), including specific reference to questionable research practices (QRPs)	112 (36%)
Collaboration & Inclusivity	Increasing the number of individuals involved in the research process (including specific notes towards diversity and inclusion)	64 (21%)
Critical Concern	A collection of themes representing concerns over exploitation, dubious intentions, unnecessary surveillance, or ‘trendiness’ of OSPs	60 (19%)
Lack of Quality (Open Access)	Concerns about shallow review processes as well as ‘pay to play’ publication models (specific to open access publishing)	44 (14%)
Pre-registration	Posting (or submitting for review) study predictions prior to data collection	41 (13%)
Replication & Reproducibility	Specific mentions of replicating published research or reproducing shared analysis of in-progress work	33 (11%)
Progress	Notes about scientific progress	21 (7%)
Public Communication	Enabling more direct engagement with ‘lay publics’	21 (7%)

Note: Most answers were single words or short phrases, so definitions above are provided to contextualize each theme.

Table 8. Emergent themes regarding problems with OSPs.

Theme	Definition	Freq. (%)
Labor and Value	Concerns regarding the (often devalued) labor required to properly engage in many OSPs	79 (28%)
Misuse, Mistrust, and Theft	Concerns that shared study materials (especially data) could be used by third parties for ulterior motives	73 (26%)
Unclear Standards and Practices	Concerns that open science practices are not clearly defined or understood by all involved	61 (22%)
Legality, Privacy, and IRB	Concerns regarding the privacy of human participants in research when their data are shared (even in aggregate)	59 (21%)
Incompatibility and Exclusivity	Concerns that open science practices de facto exclude certain forms of scholarship (e.g. non-quantitative)	51 (18%)
Open Access Quality	Concerns about the perceived or actual quality of open access publishing (e.g. poor peer review)	48 (17%)
Fear and Attacks	Concerns about toxic cultures (such as #bropenscience) and aggressive ‘witch-hunting’ when sharing scholarship	35 (13%)
Control	Concerns that open science practices are too restrictive of creativity and incidental finding	22 (8%)
Resistance to Change	Concerns that scholars (e.g. senior scholars and journal editors) are unwilling to adopt open science practices	16 (6%)
‘Open Washing’	Concerns that open science practices are being used as a heuristic cue when evaluating scholarship	15 (5%)

Abele-Brehm, A. E., Gollwitzer, M., Steinberg, U., & Schönbrodt, F. D. (2019). Attitudes toward open science and public data sharing: A survey among members of the German psychological society. Social Psychology, 50(4), 252–260. https://doi.org/10.1027/1864-9335/a000384

 Web of Science ®Google Scholar