An epistemological question: Is there personal bias by scientists in analyzing a set of experiments? There is a recent book review [The Evidence for the Top Quark: Objectivity and Bias in Collaborative Experimentation by, Kent W Staley] at "physicsweb" by Bill Carithers discussing the interesting search for the sixth component of the Standard Model of particle physics at Fermilab.
"I was particularly interested in Staley's examination of possible bias in the methodology and how the CDF collaboration dealt with it. When particle physicists try to find a particular set of events among the trillions of collisions that occur in an accelerator, they have to focus their search by ignoring data outside a certain range. In the case of the top quark, the CDF physicists knew that they could select their data in two different ways. Although both approaches were valid, the one they chose turned out to produce a stronger signal."--Bill Carithers
"The top quark: an unbiased tale"
The Evidence for the Top Quark: Objectivity and Bias in Collaborative Experimentation by, Kent W Staley
And, in October of 2005 to cite one of many recent instances of scientists doing bad science was the dismissal of Luk Van Parijs of MIT who admitted "...to fabricating and falsifying research data in a published scientific paper and several manuscripts and grant applications."
"Five experts were asked to view results from some of these papers, which cast light on the detailed signals that prompt cells of the immune system to commit suicide. While they emphasise the need to see the underlying raw data, they say they have sufficient doubts about the authenticity of specific figures within the papers to merit an independent investigation into their authenticity."
Peer review is worthwhile and a necessary part of sound scientific methodology.
"MIT professor sacked for fabricating data"