In this lecture you will learn:
In many ways, bad science is even worse than pseudo and junk science. It often LOOKS so much like real science as it is dressed up to have the proper "trappings" of science. Often the ONLY way bad science is exposed is when somebody looks closely at the data and/or repeats their experiment.
There are things to look for, like going
shopping for a used car ....
vs objective tests
Subjective tests are often deceiving and inadequate. Our subjective experience tells us that a 50oF day in fall is cold, whereas a 50oF day in spring is warm. For example, an objective test is to use a thermometer AND define what hot and warm and cold are in terms of a thermometer. Good science strives to use objective tests whenever possible.
Controls are essential for science. Most experiments require an independent control. For example, if one battery manufacturer states that their batteries last longer than brand X, identical toy rabbits must be fitted with both brand X and the challenger batteries and turned on at the same time. The brand X batteries would be the control in this experiment. The lack of proper "controls" is often at the root of pseudo science and science fraud. It is also at the root of honest mistakes and misconceptions. They are also called "normal" in literature
Precision and standards
Not all thermometers are accurate or give a "true" measurement of conditions. Check a rack of thermometers in a store and you will find the readings vary. The equipment used for science is calibrated or set to perform within narrow limits called standards. Scientific thermometers are calibrated to show exactly 0oC when water freezes, and exactly 100oC when water boils at a set elevation above sea level.
In addition to being accurate, equipment must be precise. A thermometer, for example, must give the same reading every time it is placed in boiling water. Problems with precision are often seen in bathroom scales (there is often problems with accuracy as well).
and adequate sample size
Scientific experiments require repetition. Unlike the temperature of boiling water, repeated observations of the natural world show a great deal of diversity. The greater the natural diversity the larger the sample size and repetition of the experiment is required.
In order to deal with discrepant data, scientists use statistical analysis to assess whether the variation is within normal limits, or if the variation is outside of normal limits and the hypothesis fails.
Results are almost never 100% consistent. The difficulty comes in trying to determine whether the data fulfills or fails the prediction(s). This is where STATISTICS comes into play. Statistical analysis will answer the question "WHAT HAPPENS MOST OF THE TIME". Statistics helps to separate RANDOM events from natural variation found in the real world. For example, while men ON AVERAGE are taller than women in a society, there are always women that are taller than men. In some societies, almost all the women will be taller than men in other societies.
Actually, if the results were that consistent (like Mendel's experiments with peas) this indicates that the data has been tampered with, or data has been "smoothed out" to fit the hypothesis predictions. This is forbidden in science.
Each potential article is submitted to other scientists who examine the work. If they are unable to find weaknesses in the experimental methods, data, and analysis of the data, then it is recommended for publication. Publication puts it out into the science community for review by everyone. Others may then decide to duplicate the experiments, either skeptics or those interested in building on the research.