Stats >> WebLog (Started on February 3, 2004)

This weblog is a collection of materials that I find interesting and useful. I hope you find these entries useful as well. The dates listed on the weblog are the dates when the original entry was written. I will make some minor editorial changes from time to time, but I'll note any major changes with a revision date. Some of these entries are incomplete because I was interrupted while writing them. I try to finish them up within the next day or so, but once in a while a weblog entry stays unfinished for some time. Please be patient, as I can only work on these web pages during quiet times at work.

Archive

Archive arranged by topic

The last ten weblog entries

  1. Stats: Bootstrap estimates of the standard error (June 20, 2008). A regular correspondent (JU) on the MEDSTATS email discussion group asked about using the bootstrap to estimate the standard error of the mean in a simple case with 9 data values. He wanted to know why the commonly used approach in the bootstrap community was to use n instead of n-1 in the variance denominator. It seemed to him that n-1 would produce an unbiased estimate of the standard error and wanted to know if that was true just in this special case or true in general. He quoted from the book by Efron and Tibshirani that they felt that for most purposes either method would work well.
  2. Stats: Running R on a web server (June 17, 2008). I'm working on a project for planning and monitoring accrual patterns in clinical trials. This will eventually lead, I hope, to a grant to support this work. I have some existing R scripts and want to examine the possibility of running those scripts on a web page.
  3. Stats: Can I run a quantitative analysis on this data? (June 17, 2008). I get lots of questions about how small a sample size can be before you can't perform a quantitative analysis and instead are forced to summarize the data in a qualitative fashion. The most recent question involved looking at infants with feeding disorders. There were 29 of these infants, but a subgroup of 5 had disorders so severe that they still required a feeding tube at 3 years of age. The researcher wanted to compare this group of 5 to the remaining 24.
  4. Stats: Criticism of random effects in a meta-analysis (June 14, 2008). There are two approaches to combining results in a meta-analysis. They are called the fixed effects model and the random effects model. The fixed effects model effectively weights each study by the sample size, or by a measurement that is closely related to the sample size, such as the inverse of the standard error of the estimate. A random effects meta-analysis, in contrast, will assume that an estimate from a single study has two sources of error. One error is the same as in the fixed effects analysis and varies by the sample size of the study. The other error is a random component that is independent of the sample size and represents uncertainties due to conditions in this particular study that differ from conditions in other studies.
  5. Stats: Modeling a declining count variable (June 14, 2008). I've been working on an interesting project that requires Poisson regression. A company sends out a mailing and gets a certain number of telephone calls back on each of the days following. The number of phone calls is typically (but not always) highest on the first day afterwards and declines rapidly on successive days. I wanted to develop a simple Poisson regression model for this data.
  6. Stats: Eliciting a prior distribution for rejection/refusal rates (June 7, 2008). I got a question about the Bayesian model for rejection/refusal rates. I had used three prior distributions in my calculations, a Beta(10,40), a Beta(45,5), and a Beta(25,25). The question was, how did I select those prior distributions.
  7. Stats: When does heterogeneity become a concern? (June 5, 2008). Dear Professor Mean, I have an ANOVA model and I am worried about heterogeneity--unequal standard deviations in each group. How should I check for this?
  8. Stats: How wide can you make a line of your web page? (May 27, 2008). When you are writing a web page, you do not have much control over how it is displayed at a remote site. If you really wanted this level of control, you should use a more rigid format, such as Adobe PDF files. But there are some serious advantages to the reader to let him/her control the display of a web page. As a web author, you should strive to make your web pages look good under a reasonable set of alternatives, such as differing screen sizes or differing font sizes.
  9. Stats: A simple Bayesian model for exponential accrual times (May 26, 2008). Here is a simple Bayesian model for exponential accrual times. This model will help researchers to plan the estimated duration of a clinical trial. The same model will also allow the researcher to monitor the accrual during the trial itself and develop revised estimates for the duration or the sample size.
  10. Stats: A short biography that can be used as an introduction (May 9, 2008). I'm giving a talk today, and I was asked to provide some material that could be used to introduce me.

The last ten interesting articles, interesting books, or interesting websites.

  1. Email Address Munger/Email Address Encoder Excerpt: Email address munging is the act of using ASCII, JavaScript, and scrambling of letters in your email address in order to hide your email address from spam bots, spiders, and spoofers. Our anti junk email tool protects your email address and helps prevent spam by avoiding spam bots and email address harvesters. This tool allows you to munge and mask your email address by using ASCII, JavaScript, and/or image links.
  2. The role of clinical uncertainty in treatment decisions for diabetic patients with uncontrolled blood pressure. Description: This article examines the process of care for diabetic patients with elevated blood pressure. Clinicians frequently did not intensify the therapy, mostly because of uncertainty about what the true blood pressure would be.
  3. The "3T's" road map to transform US health care: the "how" of high-quality care. Description: This article outlines the three major translational steps needed to apply research to actual clinical care.
  4. How much loss to follow-up is acceptable in long-term randomised trials and prospective studies?. Description: This article reviews current literature recommendations on how low a drop-out should be in order to be acceptable. The general consensus is that 5% or less is good and that 20% or higher is bad (though some authors will say that 50% or more is bad). The authors point out that the statistical consequences of drop-outs vary from study to study and that rigid adherence to any fixed cut-off is inappropriate.
  5. If we're so different, why do we keep overlapping? When 1 plus 1 doesn't make 2. Description: This article provides a simple explanation why two overlapping confidence intervals is not t he same as showing that the two means are not statistically different from one another.
  6. How to interpret figures in reports of clinical trials. Description: This article reviews several commonly used data display methods and explains what a non-technical reader should look for.
  7. What constitutes a "clinical trial"?: A survey of oncology professionals. Description: This article summarizes ths opinions of 66 oncology researchers on what constitutes a clinical trial. While the original responses were broadly inclusive, the responses became less inclusive when definitions of the Cancer Care Ontario and the Ontario Cancer Research Network groups were provided.
  8. Pediatric Drug Studies Seen as Obligation of Other Parents' Kids Description: This webpage summarizes the research of Davis and Matthew Davis, who surveyed parents about medical research in children. While most parents wanted to see research that insured safe medicines for children, most would not agree to let their own children participate in research studies.
  9. Overconfidence as a Cause of Diagnostic Error in Medicine. Description: This article proposes that a common source of misdiagnosis errors occur because of overconfidence and suggests strategies for reducing these types of errors.
  10. Missed and delayed diagnoses in the ambulatory setting. Description: This letter to the editor criticizes the use of malpractice claims to identify misdiagnosis rates.
  11. GRADE working group Excerpt: The Grading of Recommendations Assessment, Development and Evaluation (short GRADE) Working Group began in the year 2000 as an informal collaboration of people with an interest in addressing the shortcomings of present grading systems in health care. The working group has developed a common, sensible and transparent approach to grading quality of evidence and strength of recommendations. Many international organizations have provided input into the development of the approach and have started using it.