Unfortunately as undergraduates, we can have the false impression that all of our research will be published, as long as it was a good study and well written. Well, it seems we have been misled. It seems that within science, and particularly psychology, research is generally only published when a significant result is found. This form of publication bias is known as the file drawer problem. Named aptly because those psychologists who have slaved months and potentially years of their lives on a piece of research but don’t find a significant result, often file these away, never for it to be read again because they will not be published.
The publication bias in publishing more statistically significant results than non-significant has been shown to occur even though the significant papers have a poorer research design than the null result ones (Easterbrook, Berlin, Gopalan, Matthews, 1991). This level of publication bias can have detrimental effects on psychology. With papers finding a significant result, three times more likely to be published over non-significant ones, psychology is potentially missing on many crucial papers (Dickersin, Chan, & Chalmers, 1987). Even though a paper may not find a significant result, this does not mean that these results are enlightening. For example, results finding that say the attentional spotlight doesn’t exist and this was carried out in a very well designed study with accurate findings are likely to not be known and thus not informing psychology that a theory is disproven.
On the other hand, not publishing non-significant results does hold some benefit. With regards to say treating a psychological disorder; if a patient researches treatment and finds that many papers say there is no treatment or it doesn’t work well, this can have a detrimental impact on the wellbeing of the individual. Also, one study questioned researchers as to why some papers were not published and they themselves said that they did not send it to be published, with many reporting a change in interest to another topic as theirs now held little significance.
It appears therefore that the file drawer problem can seriously hinder psychological research by not publishing non-psychological research, despite the chance it could change ideas on particular theories. It is apparent though that not publishing non-significant research can protect individuals from being aware of this research in treatment, and that sometimes, research isn’t published due to the researches own choosing. I guess this explains why we find it hard to find opposing research for our essays!
http://www.scientificexploration.org/journal/jse_14_1_scargle.pdf
dy bannee diu 
I do agree with you in relation to if a patient looks up on potential treatments and how detrimental it could be for them personally, but I think we have to look at this from a couple of different prospective.
One reason for publication of research is for peer review, allowing others to see our work and criticise the work we have produced. If our results are not significant, maybe our peers can point out any problems with our methodology, and analysis that could then, result in our findings being significant. From this we could then repeat the study, to then hopefully finding something.
Also, maybe research that is not signifiant can give us an insight to what we need to look at. Using your example of treatments; how do we know if a treatment does not work if a paper has not been published on it, how do we then improve the treatment to eventually support that we have found a treatment that works?
Pingback: Homework for my TA :) « te9192
I do believe that the file drawer problem is a problem, as it skews our perception and knowledge of a certain area of research – as the journal editors in a way sensor what we see and are allowed to know. I do understand that sometimes research is unsignificant and doesn’t even show anything, and they can’t publish every research paper, but when they is an obvious bias,(eg it goes against the journal editors opinion or institution bias) i think it is wrong.
Pingback: Homework for my TA (week 8) « psud63
I agree with the argument that it is very problematic that only research which has a significant result is published. There will be a lot of research out there significantly supporting specific ideas or theories, however there may be more that do not significantly support the theories. This means that there could be 100 studies which show that a certain type of medication does not help with the treatment for depression, however there may be only 10 studies which show a significant result from the same drug. It is the 10 studies which are published, however is this representative of the truth?
The following article ‘ Unpublished results hide the decline effect’ talks about the decline effect and how unpublished research hid this from the public eye. The decline effect is when scientific claims receive decreasing support over time and it was found that ‘some effects diminish when tests are repeated’. These declining effects, however, have not been published and therefore the decline effect has been hidden.
http://www.nature.com/news/2011/110223/full/470437a.html
This is an example of how people can be misled if all the research is not presented, whether it be statistically significant or not.
I also agree that this is a serious problem in the field of psychology and the publishing of researcher’s work. I particularly agree with your point about how some non significant results could present a whole new perspective and evoke a whole new set of studies and may even provide evidence that supports a theory. I also wonder if the publication if these results would save resources because surely it’s possible for one researcher to unwittingly carry out the same experiment as another only to find that the results are non significant, which they could have found out before completing the research in the first place if the previous research results had been published. In addition I know that some journals use peer reviews to decide on which papers to publish and which ones to reject which I would have thought would reduce this problem. However I came across some research by Mayo et al. (2006) that suggests otherwise. This paper discusses how there can be quite a difference between the 2 main researchers and the other researchers on the review panel. After reading this and how this issue can also have an effect on grants/funding I feel the problem is even greater than originally thought.
Ref
Mayo et al. (2006). Peering at peer-review revealed high degree of chance associated with funding of grant applications. http://www.sciencedirect.com/science/article/pii/S0895435606000059
Pingback: Homework for my TA week 8 « fr4nw
I am always fascinated with psychology that is why i took a Bachelors Degree on psychology. `
<a href="Current write-up from our own blog site
http://www.prettygoddess.com