Feedback on book proposal.

I'd like some feedback on a book I am proposing to write. The working title is Statistical Evidence. This book would consist of an introduction, and at least three major sections.

• Apples or oranges? This section discusses the selection of the control group and whether this constitutes a fair (apples to apples) or unfair (apples to oranges) comparison.
• Who was left out? This section discusses the patients who were excluded from the study, who refused to join the study, who dropped out during the study, or who failed to comply with the conditions of the study. By examining who was left out, you can assess how generalizable the results are.
• Mountain or molehill? This section discusses the quality of the outcome measure. Was it valid and reliable? Was it precise enough? Did it measure an outcome important to your patients? Was the result large enough to be clinically important?

There is additional material that I might want to add.

I would like feedback from you about my proposed book. Any feedback is appreciated, but I would especially like comments on the following areas.

1. What is a good title for this book? Some possibilities I have considered are:

• Statistical evidence.
• Weighing the evidence--statistics in medical journals.
• How to read a medical journal article.

or possibly some combination of the above. What title do you think is most appealing?

2. Does the current organization seem to make sense to you?

3. What additional material would help make this a better book. I have several topics that I might want to include (see below). Which of these do you like? Are there any other topics of interest?

• Guidelines for meta-analysis. This section would discuss heterogeneity, publication bias, and the presence of a carefully written protocol prior to data collection.
• Interpretation of confidence intervals and p-values. This would cover the difference between statistical significance and clinical significance.
• Interpretation of the odds ratio and relative risk. This would also cover the number needed to treat.
• Objectivity in research (early draft). To what extent do we draw biased conclusions because of our preconceptions. To what extent do monetary ties influence the quality of research.
• The ethics of randomization and placebos in research. When it is ethical to randomize someone to an alternative treatment or possibly to no treatment?
• Interpreting and applying results of diagnostic tests. What are sensitivity and specificity. How can you use likelihood ratios to assess the probability of disease. What is an ROC curve?

4. I have substantial amount of material on blinding in research. This could stand alone as its own section or it could be combined with the "Mountain or molehill?" section, since it relates to the quality of the outcome measure (blinded assessment of outcome is superior to unblinded assessment).  Would you leave it as its own section or combine it?

5. Which of the current topics should be expanded. Should anything be reduced or removed?

This webpage was written by Steve Simon on (unknown date), edited by Steve Simon, and was last modified on 2008-07-08. Send feedback to ssimon at cmh dot edu or click on the email link at the top of the page. Category: Statistical evidence