Study shows statistical methods seriously flawed
Pittsburgh PA (GSP). A recent study by Carnegie Mellon statisticians has concluded that nearly all statistical methods are flawed with heavy bias towards the views of the researchers and investigators.
The metanalysis examined over 20 000 scientific studies published by peer reviewed journals, ranging from basic biology research to environmental sciences to psychological articles. The analysis found that a statistically significant proportion of the studies could be explained by manipulation, or mining, of the data. Moreover, the researchers found that the manipulation of the data could not be accounted for by chance, but rather is due to a self-promoting agenda by the labs or principle investigators conducting the study.
Igemidalo Alosnugo, leader of this landmark study, has said, "there is really no surprise here. Someone once said there are lies, damn lies and statistics. We have only shown that in a scientifically sound way."
While this news is expected to rock the scientific community, there is likely to be little impact on the daily life most Americans. CNN Scientific Correspondent, Alfredo Perez has said, "People for the most part do not care. Who cares if an infant fears a visual cliff, or newborns have a conception of numbers and simple math? Does it matter to anyone if a protein is involved in staving off depression? Does it make a difference if scientists have been lying about these things from the start? These things are simply not important and should not be concerning the general population."
Response to the paper by the public has been surprisingly apathetic. One college graduate from a prominent university said, "What's new about this? I've been saying for years you can make statistics show what you want to show."
It appears like a day of triumph for those who have decried the use of government and university grants for useless research. One study profiled a sample of college students who responded favorably to the article. About 82% of those who liked the paper have said at one point in their life, "I just don't believe in all those claims (referring to science results)" and nearly 96% have once said, "that kind of stuff is just pseudo-science." Most of these students initially thought the analysis only looked at psychological studies and were surprised to hear it included biological, chemical and social sciences and economics studies too. The same study profiled a group of students who did not like the paper and found that most of them (erroneously) considered science an important part of life. The survey authors added that a statistically significant proportion of these subjects were of low intelligence and of poor social adjustment.
The article by Alosnugo and colleagues will be published in the August 2006 issue of the Journal of Research Sciences and Statistical Methods.
The metanalysis examined over 20 000 scientific studies published by peer reviewed journals, ranging from basic biology research to environmental sciences to psychological articles. The analysis found that a statistically significant proportion of the studies could be explained by manipulation, or mining, of the data. Moreover, the researchers found that the manipulation of the data could not be accounted for by chance, but rather is due to a self-promoting agenda by the labs or principle investigators conducting the study.
Igemidalo Alosnugo, leader of this landmark study, has said, "there is really no surprise here. Someone once said there are lies, damn lies and statistics. We have only shown that in a scientifically sound way."
While this news is expected to rock the scientific community, there is likely to be little impact on the daily life most Americans. CNN Scientific Correspondent, Alfredo Perez has said, "People for the most part do not care. Who cares if an infant fears a visual cliff, or newborns have a conception of numbers and simple math? Does it matter to anyone if a protein is involved in staving off depression? Does it make a difference if scientists have been lying about these things from the start? These things are simply not important and should not be concerning the general population."
Response to the paper by the public has been surprisingly apathetic. One college graduate from a prominent university said, "What's new about this? I've been saying for years you can make statistics show what you want to show."
It appears like a day of triumph for those who have decried the use of government and university grants for useless research. One study profiled a sample of college students who responded favorably to the article. About 82% of those who liked the paper have said at one point in their life, "I just don't believe in all those claims (referring to science results)" and nearly 96% have once said, "that kind of stuff is just pseudo-science." Most of these students initially thought the analysis only looked at psychological studies and were surprised to hear it included biological, chemical and social sciences and economics studies too. The same study profiled a group of students who did not like the paper and found that most of them (erroneously) considered science an important part of life. The survey authors added that a statistically significant proportion of these subjects were of low intelligence and of poor social adjustment.
The article by Alosnugo and colleagues will be published in the August 2006 issue of the Journal of Research Sciences and Statistical Methods.
posted by GSL Enterprises at
2:33 PM
0 Comments:
Post a Comment
<< Home