Reader Comments

Post a new comment on this article

Nature of our DVs?

Posted by rjwpsy on 10 May 2013 at 01:29 GMT

Ultimately, the author intimates that the behavioral sciences are "fields where theories and methodologies are more flexible and open to interpretation, bias is expected to be higher." Isn't this just due to the nature of what the harder and softer sciences study? Studying human behavior is, in a lot of ways, is much more difficult (e.g., operationally defining and quantifying theoretical constructs, greater variability, no unifying theory). So, yes, the theory, method, statistics, and interpretation of our results will be more variable and open to criticism, so the strongest papers (positive results) will survive given the limited journal space; this is no surprise.

And what evidence is there that "scientists' prestige within the community, their political beliefs, their aesthetic preferences, and all other non-cognitive factors – play a greater role in all decisions made in research" for behavioral sciences? I did not see that any of these constructs were measured in the current study, and I think such intimations are, ultimately, unwarranted. I do not know of one scientific field that is immune from such factors.

Thirdly, how did the author conclude that "most of the research published in the social and behavioural sciences is qualitative, descriptive or speculative, and is published in monographs rather than journals, so it eludes the conclusions of this study"?

Fourth, the argument regarding statistical power seems flawed to me. I do not follow the reasoning that "Based on statistical power alone, therefore, the physical sciences should yield as many or more positive results than the other sciences". The references cited in support seem extremely limited given the broad conclusions made. Moreover, I would hypothesize that the physical sciences would publish MORE null results given that there is less "gray area" in conducting research and interpreting results?

Fifth, is searching journals with one Boolean phrase "“test* the hypothes*”" limited? Wouldn't including another Boolean phrase, such as "predict*", be more comprehensive?

Sixth, I think the complete dismissal of SLKoole's argument regarding journal space as "far from conclusive" can apply to some of the statements made in the current paper.

Regardless, this article confirms the publish-or-perish nature of academia, which is hardly touched on. Journals want positive results (88.3 vs. 81.4), regardless of discipline.

No competing interests declared.