My goal in writing this blog is to promote the questioning of existing beliefs and to promote prospective testing of hypotheses as a means of discovering the world around us and making decisions. I contend that this is an approach which has great value not only for science but for most all aspects of human activity including social, political and behavior issues.
Hypothesis testing involves the process of hypothesis formation, designing an experiment to test the hypothesis, conducting the experiment and evaluating the results.
Hypothesis formation involves reviewing existing data and theories and then formulating a hypothesis. There are many types of hypotheses. One hypothesis could be a proposed mathematical formula for the prediction of the behavior of elementary particles. Another hypothesis might be a prediction regarding the impact of a change in scheduling of work shifts in a factory. Another hypothesis might be a prediction regarding the impact of a change in health care policy.
A valid hypothesis must be falsifiable. That is, it must be possible to design an experiment which, in principle, could have an outcome which disproves the hypothesis. The actual outcome of an experiment will, of course, depend on the validity of the hypothesis.
Designing an experiment to test the hypothesis in essence involves planning a study which tests the predictive ability of the hypothesis under a specified set of conditions. All the details of the experiment need to be specified in advance including how the data generated by the experiment will be analyzed. This makes the experiment truly prospective in nature and not dependent on retrospective analysis of data. An experiment generally cannot prove a hypothesis to be true because most experiments test the hypothesis under only one specific set of circumstances where the hypothesis applies. It is always possible that when the hypothesis is applied to a different set of circumstances that the hypothesis might fail. So, experiments generally can potentially disprove hypotheses but not prove them to be always true.
The experiment must then be conducted fully according to the predetermined plan and the data analyzed also according to the prescribed plan.
In this blog I will attempt to address all aspects of this approach. The first step is a critical analysis of current beliefs and behaviors to see if they are consistent with existing data. Critical thinking is the first step in hypothesis formation. It is not always possible to prospectively test our hypotheses, but we should attempt to do so when feasible, especially if the potential impact of an action based on a hypothesis is likely to be large. We should attempt to determine if there are practical ways too conduct experiments in order to test a hypothesis.
I hope this blog will be engaging. I plan to comment on public affairs, hopefully from an evidentiary and not doctrinal perspective. Where possible I will try to cite experiments that have been conducted to test beliefs and hypotheses. In other case, much of what I say will likely be opinion directed towards hypothesis formation based on what will hopefully be a critical analysis of existing facts. I will try to bring in science whenever appropriate and to propose ways to test hypotheses when feasible.
Finally, this is a blog and not a research paper. I look forward to readers commenting to bring to light relevant facts and to share their analyses. I hope my views will be challenged and I expect that they will evolve as a consequence of the dialogue. John Maynard Keynes when he was challenged for having changed his view on an economic matter, replied “When the facts change, I change my mind. What do you do, sir?” I hope that will be the case here for all of us.
See also About this Blog.
See also About this Blog.
This comment has been removed by a blog administrator.ReplyDelete
Thx for your blog, it's very informative.ReplyDelete
Wish you good luck.
Toby, due diligence data room