Think Twice About 'New Research': Are They Reliable?
An ironic headline once read, "New Research Finds – Don't Trust Research." Why? Because often, these studies aren't as reliable as they seem. People publish supposedly verified data, but proving them wrong is another story.
Every so often, we hear about a "new study" that claims to explain various wonders of creation in a simple and scientific way, such as how apes became humans or how the body fights diseases. But it turns out some believe that the ease of publishing "new research" is too much, leading to headlines like: "New Study Finds – Don’t Trust Studies." These studies, it seems, are simply not reliable. People publish allegedly verified data, and proving them wrong is no easy task.
Some diligent individuals have attempted to investigate this issue, as reported by the renowned newspaper, "The Guardian": "New research finds that..." is often stated in news editions, papers, and even advertisements. However, an in-depth examination of numerous published conclusions in psychology journals paints a grim picture of this science's state. An international team of researchers, including 270 scientists across five continents, repeated 100 experiments whose results were previously published in leading journals and succeeded in replicating only 36% of the original outcomes. "Of course, I would have been happy if the same results were confirmed again," says Brian Nosek, a psychology professor who led the study at the University of Virginia, "I am disappointed..."
The research claims results, but upon repeated examination, the vast majority of cases prove it just isn't true. The experiments are repeated, and such outcomes are not found...
Doron Halutz writes in "Haaretz": "Thirty years ago, a prestigious journal, 'Nature', published research that scientifically examined a question, collaborating between esteemed psychologists, statisticians, and astrologers. Each astrologer was given a birth chart of a random test subject and asked to determine which of three personality analyses, based on a standard psychological test, belonged to the chart's owner. The chance of guessing correctly by random: one-third. The success rate in the study: one-third. The science concludes: Astrology is nonsense."
"Late last summer, the parallel journal, 'Science', published an article that summarized a four-year research project, scientifically examining the accuracy of scientific research. The findings, alas, were not encouraging."
"'The Replication Project,' led by Professor Brian Nosek, a social psychologist from the University of Virginia, invited researchers to select an article published in one of three leading psychology journals in 2008 and to rerun the last experiment described. The 270 researchers who volunteered for the task had to adhere as closely as possible to the original research design."
"The bottom line: only about 40% of the replications, or 'replications' in the jargon, yielded similar results to those reported in the original articles. If less than half of the studies reported in the professional literature – the cornerstone of what is considered 'scientific truth' – yield consistent results in repeated executions, how much can they be trusted? And if they can't be trusted, what is this science really worth?"
"The question is good, the answer is complex, but one thing is clear: Don't rush to get excited about 'new research finds'."
