Evidence

Decisions
Making a good judgment, or coming to a good decision, involves two steps: 1) asking the right question; and, 2) recognizing the right answer.

Asking the Right Question
The null and alternative hypotheses focus your attention on the relevant factor. When plants wilt and die, the obvious factor to investigate is moisture. The null hypothesis might be that rain is not required for plants to survive (the factor has no effect on plant survival). The appropriate alternative is that rain is required (the factor has the effect you think it does).

Recognizing the Right Answer
Coming to the right conclusion depends on obtaining information that can discriminate between the two hypotheses. Observing ponderosa pine trees growing in the forest on which you know rain falls, shows that a plant species survives in an environment that receives rain. But that is a correlation, i.e., a coincidence of two events which occur at the same time that, by itself, does not prove cause and effect. The success of those plants was influenced by many other factors which you did not measure and over which you had no control. You could just as well have said that God makes the trees grow, and you would have had as much evidence to support that conclusion but no further understanding of plant growth.

Experiment and Analysis
Plant Growth
Test the effect of moisture on plants by setting the environmental conditions for plant growth and controlling the factor being tested. In a greenhouse with known temperature and humidity, using identical soil for all plants, replicating plants, excluding disease and herbivores, apply several different amounts of moisture which are within the normal range for the species. All factors relevant for plant growth were controlled. Other factors, such as disease and herbivores that could confound conclusions by making interpretation of the results difficult, were eliminated. Greater growth in response to more moisture would not only show that the moisture factor was critical, but would allow you to calculate the amount of growth for each increment in moisture.

Hormone Replacement Therapy
Studies showed that women taking combined hormone replacement therapy (HRT) also had a lower-than-average incidence of coronary heart disease (CHD). The attractive conclusion was that HRT protected women against CHD. Subsequent controlled trials showed that HRT caused a small and significant increase in risk of CHD. Re-analysis of the data showed that women undertaking HRT were more likely to be from socio-economic groups with better than average diet and exercise regimes. Relevant factors, not considered initially, were correlated with both HRT and decreased CHD, leading to an erroneous decision that the medication produced an unanticipated beneficial effect.

Amanda Knox in court:http://abcnews.go.com/International/amanda-knox-prosector-blasts-dna-report/story?id=14195630&page=2Proper Protocol
Amanda Knox was convicted in December 2009 and sentenced to 26 years in jail for the 2007 murder of Meredith Kercher, a British student who shared an apartment with Knox in Perugia, Italy. Two experts were appointed by the appeals court to conduct independent DNA tests on crucial pieces of evidence – the alleged murder weapon: a knife, and the victim's bra clasp. The 145-page report stated that there was not enough DNA to retest on the two pieces of evidence. In addition, the analysis of the forensic data from the first trial found that the investigators had made a series of glaring errors in gathering and interpreting the evidence that could have resulted in its being contaminated. Strict protocols are also required to avoid contamination of DNA samples from fossils and prehistoric air samples from ancient ice.

Confounding Factors
Asking the right question gets you started in the proper direction. Analyzing the situation, obtaining evidence through controlled experiment, and eliminating or adjusting for confounding factors provide you reliable information to discriminate between hypotheses. Ideology clouds the mind with a bias toward a foregone conclusion, and it can cause one to imagine effects that do not exist or ignore ones that do. Use common sense instead, and account for all the evidence before reaching your decision.
Excerpted and adapted from: Correlation does not imply causation, Amanda Knox Prosecutor Blasts New DNA Report