A new study has found that the results of many psychological experiments can not be replicated and reinforced by future research. The scientific community is expressing concern that psychology research may not be accurate and reliable.
Scientists have long validated each other’s work by having different groups conduct the same experiment, using the same materials and the methods in order to reach the same conclusions. But if psychological studies fail to abide by this rule, can they still be trusted? The approach is said to confirm that the first team did not overlook something, misinterpret the data, or fabricate the results.
Many feel that a significant amount of non-reproducible results have managed to find their way into scientific literature and may be misleading everyone. If they are unchallenged and untested, they can act as a source of information for other studies, conducted by different researchers. And if these results are untrue, then the new studies will also fail to produce reliable results.
Concerned researchers say that if enough studies prove to be unreliable, inconclusive or non-reproducible, the image of the scientific community as a whole will suffer and people will stop taking important warnings seriously.
This is an especially dangerous scenario in our modern day society as climate change is threading to kill off many species of animals and change ecosystems that affect both animals as well as humans. But climate scientists already have trouble being taken seriously by certain groups who think that they are dramatizing and over-exaggerating.
Medical research and investments decisions can also suffer and lose credibility if the number of published studies with non-reproducible results is allowed to grow and become a part of the scientific community.
The new study saw over 270 scientists from across the planet banding together to try and replicate 100 social-psychology and cognitive-psychology experiments published in prestigious journals.
Even though 97 percent (97%) of the original studies had statistically significant results, only a mere 36 percent (36%) of them still had statistically significant results when the research team finished replicating them.
What’s more, Dr. Brian Nosek, Center for Open Science executive director and social psychologist from the University of Virginia, gave a statement saying that 83 percent (83%) of the successful 36 percent (36%) concluded with smaller effects than the initial studies.
The problem is that just because a set of results could not be replicated by a second group of scientists, this does not mean that the results are inaccurate, unreliable or wrong.
And the reverse is also true – just because the results of a specific study could be replicated by a second group of scientists, this does not necessarily mean that they’re right.
Many different variables can influence the outcome of a study – change in location, moment in time, duration of the study, to name a few.
Dr. Nosek doesn’t even know if the results of this new study could be replicated. He explained that “This should just be seen as a first step, an initial piece of evidence for establishing what reproducibility in general might be”. But he does believe that “the results suggest that there is a lot of room to improve reproducibility”.
The new study was published earlier this week, on Thursday (August 27, 2015).