Sharing in Science
The Survey Results
The population of geneticists used in the study was drawn from a stratified random sample of 3,000 respondents to a larger survey of university-based life scientists. The subsample of 1,820 members of genetics and human genetics departments used for this paper excludes people not currently active in research—clinical faculty members who had not published at least one article in the National Library of Medicine's MEDLINE database in the three years preceding the study. Of the qualifying sample, 50 percent are full professors, 28 percent associate professors, and 22 percent are assistant professors; 75 percent are male, and 25 percent are female. Respondents were administered an anonymous, largely closed-ended survey. The response rate for the overall survey was 64.5 percent. Key results are as follows:
Sharing research with other scientists
There is evidence of continuing normative ambivalence about openness and disinterestedness. Virtually all geneticists believe that scientists should share their results freely with all peers (49 percent "agree completely," and 42 percent "agree somewhat" on a 4-point scale) and should be motivated primarily by a desire for knowledge (73 percent "agree completely" and 24 percent "agree somewhat"). In addition, 68 percent indicate that they are personally "very willing" to share with other academic scientists. Nevertheless, a substantial proportion also believe that scientists should keep their newest findings secret to protect their priority (51 percent "agree completely" or "somewhat") and should also receive direct, personal benefits from their scientific discoveries (47 percent "agree completely" or "somewhat").
A significant proportion—approximately 30 percent—also report that, within the past three years, they have withheld research results from other academic scientists prior to publication on at least one occasion. The most common reasons given for withholding were to protect their own ability to publish (67 percent of those who withheld indicated that this was "very important") or that of their students (75 percent responded "very important"). In contrast, 21 percent indicated that it was "very important" to withhold to honor agreements with industry sponsors.
Practical constraints other than scientific priority or potential commercialization enter into the decision to share or not. More than two-thirds indicated that the extra effort involved in sharing prepublication results was a "very important" or "moderately important" factor in influencing their behavior. Distrust of other academics is also a salient issue for many. For example, 28 percent indicated that they did not share because they believed that others would not reciprocate ("very important" or "moderately important").
Sharing and relationships with industry
Most university-based geneticists—79 percent—have some financial relationship with industry. We asked respondents to indicate the industry roles that they had in the past three years that had potential for personal gain through financial remuneration (company ownership, service on a board of directors, consulting, etc.), and 35 percent reported one or more. We also asked about industry support for faculty work (grants, gifts, etc.), and 70 percent had received some professional support within the past three years. Nearly all of those who had personal-gain relationships with industry also had private research support.
Relationships with industry are weakly associated with attitudes about openness and personal interest. For example, there was no difference between those who have no industry relations and those who have them when asked about willingness to share "information, data or materials with other academic scientists," nor was there any difference between the groups regarding the importance of keeping their newest findings secret to ensure priority. The only area of difference is that those who had no industry relationships were more likely than their peers to "agree completely" with the statement that "academic scientists should be motivated by the desire for knowledge and discovery rather than by financial gain."
Relationships with industry are associated with more secretive behavior. We defined secretive behavior as self-reports that the respondent "usually" or "always" intentionally withheld pre-publication information from other scientists. Geneticists with industry relationships were somewhat (but significantly) less likely to share with their university peers and students than were scientists with no relationships or relationships that were limited to funding their university work. This includes venues such as seminars in their own departments, at other academic institutions and at professional meetings. We emphasize, however, that the percentage of faculty members that always or usually withheld information is a minority in all groups. For example, 12 percent of those with no industry relationships, 15 percent with professional funding, and 19 percent with personal gain relationships reported "usually" or "always" holding back at professional conferences.
The main disruption in scientific communication owing to secrecy and self-interested behavior is between university and industry scientists. The nearly 80 percent of faculty members who had any financial relationship with industry are less likely to share with their industrial counterparts than those who had no relationship (31 percent compared to 16 percent), and those who have personal-gain relationships are less likely to share than those with only professional support (34 percent contrasted with 29 percent).