Archive organized by date
This page is moving to a new website.
This page lists files created in calendar year 2008. Also look at the
archives for 2012, 2011,
2010, 2009, 2007, 2006, 2005, 2004, 2003, 2002, 2001, 2000, and 1999. You can
also browse through an archive of pages organized by topic. Some pages from early 2008 are not yet on this list.
December | November |
October | September | August |
July | June
December 2008 (3 entries)
- P.Mean: A false sense of frugality (created
2008-12-17). A while back I received a data set that was very well
documented, but there was one thing that I wish that the data entry person had
not done. The demographic data was listed as 45f, 52m, 22m, 21f, etc. This was
obvious shorthand for a 45 year old female, 52 year old male, and so forth.
- P.Mean: Viewing SPSS and SAS output without
buying SPSS and SAS (created 2008-12-17). For one of my consulting
clients, I asked for a copy of the output that the previous statistician had
used. I received two files, one with an SPV extension and another with an MHT
extension. It was obvious from the context that the first was an SPSS output
file and the second was a SAS output file. I have access to both SPSS and SAS
packages, but not conveniently at my laptop. I wanted to look at the files
quickly, so what could I do?
- P.Mean: Using a sub-optimal
approach in meta-analysis (created 2008-12-06). I am having difficulty
understanding the meta-analysis of ordinal data in a Cochrane systematic
review, and would appreciate advice and comments. One study in the
meta-analysis had an ordinal efficacy outcome with categories None, Some,
Good, and Excellent. The meta-analysis did 4 separate analyses, treating each
category as if it were a dichotomous outcome. Aside from the fact that this
generates (almost) more analyses than there are data, this approach seems
unnecessary and uninterpretable. The Cochrane Handbook says: "Ordinal and
measurement scale outcomes are most commonly meta-analysed as
dichotomous data." And "Occasionally it is possible to analyse the data using
proportional odds models where ordinal scales have a small number of
categories, the numbers falling into each category for each intervention group
can be obtained, and the same ordinal scale has been used in all studies."
What should the authors of the systematic review have done?
November 2008 (8 entries)
- P.Mean: What is the Lan-DeMets approach to
interim analysis? (created 2008-11-21). I read an article that talked
about a trial that ended early. They describe the approach as a
"O'Brien-Fleming stopping boundaries determined by means of the Lan-DeMets
approach.". Does anyone you know anything about this statistical technique to
determine if this is a valid approach?
- P.Mean: Ethics of research into
unscientific therapies (created 2008-11-15). What is a responsible
ethical position on research on complementary or alternative medicine that is
not based on "generally accepted" principles of science? For example,
redirecting energy fields in the body; or demonstrating the positive effects
of intercessory prayer (prayer on behalf of another person). It is one thing
for a scientist member to say "I don't think the proposed statistical
methodology is adequate to the task." It's quite another thing to say "I don't
believe that there is any scientific basis for the proposed research." What
then?
- P.Mean: IRB approval of studies with less
than adequate research integrity (created 2008-11-14). How should an
IRB view its job with respect to research approval if the study design is
"less than adequate" to provide valid results? Is the IRBs job only to assess
for patient risk? If a study's design is poor and will not yield useful
results should the IRB approve the study if there is minimal risk? Does it
matter if the study is minimal risk vs greater than minimal risk?
- P.Mean: Explaining CART models in simple
terms (created 2008-11-05). I need some help understanding and explaining Classification and Regression
Trees (CART). I am personally not familiar with this technique. When would
someone select this over linear/logistic regression model?
-
P.Mean: Reading abstracts instead of the full
article (created 2008-11-05). An interesting inquiry on the Evidence-Based
Health email discussion group generated a lot of responses. A busy
clinician has a limited amount of time to answer a clinical question. They
carry out a quick search and find 5 decent abstracts. They have two options:
1) Look at one full-text article. 2) Look at 5 abstracts. Which do people
think is preferable? It's a tricky question because both approaches have
problems. Here are my thoughts on this issue.
- P.Mean: Statisticians are not gatekeepers
(created 2008-11-04). A discussion of the proper role of statisticians
when presented with questionable data is raging in the MedStats discussion
group. I added some comments recently about the dangerous tendency for us
statisticians to view our roles as "gatekeepers". Here's the gist of my
comments.
- P.Mean: Teaching classes for a fee (created
2008-11-03). I was asked by someone at Children's Mercy Hospital how much
it would cost to continue teaching the short courses that I have offered.
Since I will also make those courses available to anyone else, I thought I'd
share my general comments about fees and logistics here.
- P.Mean: Comparing a set of microarray
experiments to a model experiment (created 2008-11-01). I have a matrix
of effect sizes from numerous microarray experiments. For example, in one
matrix I have 200 genes (rows) and 107 experiments (columns). In addition, I
also have a sort of “model experiment” which contains the values in which I am
most interested. For each gene, I am trying to determine which genes are not
statistically different from the “model experiment” value.
October 2008 (18 entries)
- P.Mean: How to design a new survey (created
2008-10-28). Someone wrote in with a question about how to design a
survey. There are entire books devoted to the subject. I couldn't do the
subject justice in a single email, but here's what I sent.
- P.Mean: Refusing to analyze a data set
(created 2008-10-28). An associate of mine has a problem. He has been
told by a statistician that they can't analyse his data because it is not from
a randomised trial. I personally feel that there is no problem with doing any
sort of analysis with this data group.
- P.Mean: A standard deviation that is too big for
its own britches (created 2008-10-22). I am a medical editor
(manuscript editor) at a peer-reviewed journal and have noticed that some
authors supply standard deviations (SD) with means even when their SDs are
more than half the value of their means. (Hypothetical example: patients
recovered function at a mean (+/- SD) of 220 days +/- 190 days after surgery.)
It is my understanding that an SD is meaningless when it is this large
(relative to the mean).
- P.Mean: Computing a difference between
the first and last variables (created 2008-10-20). Hello, I seen that
you have provided some answers to people's SPSS questions, and I was hoping
you could help me. I have a basic question that I imagine this can be done
quite easily through syntax, but I don't know how to do it. I have a data set
with multiple rows, where each row is data for an individual (pretty basic). I
have about 50 variables which are time points of data. I guess you could think
of it set up as individuals' scores on some measure at various time points
where the columns go, var1day1, var2day1, var3day1, var4day1, var2day1,
var2day2, var2day3, var2day4, var3day1, var3day2, var3day3,
var3day4.......var1day50, var2day50, var3day50, var4day50 What I need to do is
quite simple, I'm creating a variable which is simply the score of the first
time point of var1 (var1day1) minus the score of the last time point of that
same variable (var1dayX). However, it gets complicated because for some cases
the last time point is the variable day50, for some individuals the last time
point is variable day3 and there's everything in between. If all individuals
had the same time points (which ran up through day 50), I would simple say
COMPUTE NEWVAR=VAR1DAY1-VAR1DAY50. But that obviously won't work because not
every case has a day50. I also can't tell it to subtract the highest score,
because it's not always true that an individuals last score was their highest
score.
- P.Mean: What is an intervening variable
(created 2008-10-20). I'm familiar with dependent and independent
variables but I just heard about intervening variables. Please tell me what
are they, and how they deal with the other variables.
- P.Mean: Errors in statistical methodology
(created 2008-10-19). From those of you who review/appraise articles
regularly, I would like to hear what kinds or errors you find most often in
the statistical methodology. I will be training nurses to critically appraise
the statistical methodology sections of articles, and since their time and
knowledge of statistics are limited, I hope to focus mainly on errors they are
likely to find in real articles.
- P.Mean: Example of power calculation
for a repeated measures design (created 2008-10-19). I was asked how to calculate power for an interaction term in a repeated
measures design. There were two groups (treatment and control), and subjects
in each group were measured at four time points. The interaction involving the
third time point was considered most critical.
-
P.Mean: Defending Bonferroni (created
2008-10-18). I had someone argue with some advice that I gave, which is a
good thing. I had recommended the use of a Bonferroni comparison, and he
argued that Bonferroni should not be used when making "independent"
comparisons.
- P.Mean: Biostatistics or
Health Informatics programs in or near Kansas City (created 2008-10-18).
I'm looking to expand my knowledge base. I was wondering if there are any
programs in or near the Kansas City area that offer BioStats or Healthcare
Informatics types courses (things with a more math bent).
- P.Mean: What's the difference between
regression and ANOVA? (created 2008-10-15). Someone asked me to explain
the difference between regression and ANOVA. That's challenging because
regression and ANOVA are like the flip sides of the same coin. They are
different, but they have more in common that you might think at first glance.
- Normality assumptions for the paired
t-test (created 2008-10-14). I am confused about which data have to be
normally distributed on a paired t-test for testing that two data sets differ
significantly. Everitt-Hothorn "A handbook of statistical analyses using R",
page 33 says that the differences between the data should be normally
distributed without implying anything about if the original data should be
normally distributed, while Wiki t-test and Field "Discovering statistics
using SPSS" page 287 imply that both of the original data should be normally
distributed? Considering that I am a beginner in statistics, I am confused.
can you give me any clues please?
- P.Mean: T-test with 3 treatment values and 2
controls (created 2008-10-14). I received a question about how to run a
t-test when one group has 3 observations and the other group has 2
observations? It turns out that you use the same formula/program that you
would use with 30 observations in one group and 20 observations in the other
group. There are two things, however, that you need to watch out for.
- P.Mean: Using ANOVA for a sum of Likert scaled
variables (created 2008-10-09). I want to analyse data derived from a
questionnaire. The range of possible values that my variable can take goes
from 20 to 100. No evidence for rejecting the hypothesis of normality was
found. I would therefore apply an ANOVA, but I still have some doubts whether
this methods of analysis is valid, since the range of my dependent variable is
not [- infinity;+ infinity]. Is the ANOVA a valid method of analysis or are
there other approaches I can apply?
-
P.Mean: A simple example of overfitting (created 2008-10-08). A couple of the Internet discussion groups that I participate in have been
discussing the concept of overfitting. Overfitting occurs when a model is too
complex for a given sample size. I want to show a simple example of the
negative consequences of overfitting.
- P.Mean: What's new in SPSS 17 (created 2008-10-07).
I attended a web seminar on new developments in version 17 of SPSS. SPSS has
improved research and reporting tools, more accessibility for
non-statisticians and more accessibility for statistical programmers, and
easier integration, deployment, and management.
-
P.Mean: How do I fit a piecewise linear
regression (created 2008-10-07). I was asked to look at some data that involved monitoring glucose and
potassium levels before, during, and after a special infusion. You would
expect, perhaps, that there would be a flat trend before, and upward or
downward trend (possibly linear, possibly not) during administration, and a
different trend (possibly linear, possibly not) after infusion. There's a
simple regression model for this, which is sometimes called a piecewise linear
regression, segmented regression, join point regression, or elbow regression.
- P.Mean: Godwin's Law (created 2008-10-05).
There is a tendency in some discussions (mostly with a political focus, but
also in some with a medical focus), to invoke the name of Adolph Hitler in
criticizing one's opponents or to compare one's opponents to Nazis. This is
sometimes called the Reductio ad Hitlerum fallacy or the Argumentum ad Nazium
fallacy.
- P.Mean: Evaluating private conflicts of
interest (created 2008-10-01). The open source journal PLoS Medicine has
an interesting editorial that is worth commenting on. Making Sense of
Non-Financial Competing Interests. The PLoS Medicine Editors. PLoS Med 5(9):
e199 doi:10.1371/journal.pmed.0050199.
September 2008 (20 entries)
- P.Mean: Reporting Fisher's Exact test
(created 2008-09-26). I am a doctoral student trying to handle stats
analysis of very skewed data, so I think I've done all non-parametric tests by
the book. My time for writing the report was up when suddenly, I was told by
my advisor to change from Pearson chi square to Fisher's Exact test without
mentioning any details how to report it and none of the books I have carry an
example. Thank you so much for your very user/student friendly web page. I owe
you!
- P.Mean: Good examples of bad studies
(created 2008-09-26). Does any one have a good example of a fairly flawed therapy article for
a course that I teach in EBM. Seems most of the articles I find aren't too
badly designed. Students always want to see some bad articles to critique.
- P.Mean: Power calculations for repeated
measures designs (created 2008-09-25). I’ve been struggling with a
design/analysis question related to repeated measures design and power
analysis. Can you help?
- P.Mean: Can I ask you a
question? (created 2008-09-25). For several years now, I have enjoyed
reading your webpage and have recently purchased your text- thank for your
efforts in helping those of us attempting to learn biostatistics. I’ve been
struggling with a design/analysis question related to repeated measures design
and power analysis. I’m not sure if you are currently accepting questions of
this nature, and thought I would check to see if this would be OK. I certainly
understand if this is something that you would rather not involve yourself
with.
- P.Mean: Comparing pre and post
data with a parallel control group (created 2008-09-25). I am
retrospectively comparing pre and post treatment heart rates for two different
populations. I was going to use a paired t-test for comparison within each
population. Can I still use an independent t-test for comparison of the post
treatment differences between the two populations? If not, what would be the
most appropriate test?
- P.Mean: Processing
skip fields in SPSS (created 2008-09-25). How do I program skips on
SPSS so that data would not be entered on irrelevant questions?
- P.Mean: Venn diagrams with
proportional areas (created 2008-09-23). I was asked by someone to come up
with a graphic summary of a data set that includes three binary factors that
can be either present or absent in any combination. Typically this can be
illustrated with a Venn diagram, the intersection of three circles but I
wondered if you could do a Venn diagram with areas proportion to the actual
probabilities.
- P.Mean: Can an outcome with three levels be
used in logistic regression (created 2008-09-18). I had a quick
question about logistic regression. Is this the appropriate test to use when
your outcome variable has 3 levels? For example, we are looking at factors
associated with obesity in children. Our outcome variable is BMI percentage
and is classified as either normal, at risk, or overweight. I ran logistic
regression on SAS and then realized this may not be the right test to run.
- P.Mean: Links and references that I need
to add (2008-09-18). I often run across important links and references that I need to add to my
webpages. I don't want to add an undocumented link or reference, though. I try
to include a full description, including a summary or excerpt. This page will
be where I "park" those undocumented links and references until I get around
to documenting them.
- P.Mean: Jackknife applied to entropy
calculations (created 2008-09-15). I have been working with entropy for a
couple of different projects and one important question to ask is "How much
does the entropy change when a single observation is removed from the data
set. This process of removing one item from a data set and recalculating a
statistic based on the remaining (n-1) observations is called jackknifing. It
is a very simple but still very useful technique in a variety of statistical
settings.
- P.Mean: Use of entropy measures
for sperm morphology classification (created 2008-09-13). Entropy is a
measure used in quantum physics, communications, file compression, and
statistics. There are a variety of informal interpretations for entropy. A
high value of entropy implies a great deal of uncertainty, very little
regularity and limited predictability. High entropy describes a process that
is full of surprises. A low value of entropy implies limited uncertainty
substantial regularity, and very good predictability. The lowest value for
entropy is zero, which represents constancy, perfect regularity, and perfect
predictability. Entropy is a useful measure for sperm morphology
classifications, because it provides a quantitative way to assess the degree
to which different laboratory technicians will apply sperm morphology
classifications differently on the same set of sperm cell images.
- P.Mean: How do you compute a
continuity correction for a confidence interval? (created 2008-09-12). I
helped author a page on Wikipedia about confidence intervals for a binomial
proportion and a question arose on the discussion page about applying a
continuity correction.
- P.Mean: What does "population-based" mean?
(created 2008-09-12). I was reviewing an article that I am a co-author on
and the author described the dearth of "population-based" studies. I had to
think a bit, about what that term really meant.
- P.Mean: Comparisons involving distinct
groups collected at different times and with different methods (created
2008-09-12). I have a data set of 100 children with a specific health
problem. In this set I have medical histories of the children. In another
study, I have collected a data set of 65 children without that specific health
problem. In this set I also have medical histories of the children. Is it
possible to compare the two samples in some way to determine whether there are
significant differences in the medical histories in the two sets of children?
- P.Mean: Multiple email accounts (created
2008-09-09). I have created several email accounts to help separate my
personal life from my professional life and my new career as an independent
statistical consultant from my old career at Children's Mercy Hospital. You
can use any of these email accounts, and I will answer, of course. I will be a
bit more efficient, though, if you target the correct email address.
- P.Mean: Where did you get that formula
for the confidence interval? (created 2008-09-09). I sent someone a
confidence interval for a single proportion, and they asked how I computed it.
That's a fair question. It turns out that I used a classic formula that
everyone learns (and then forgets) in their basic Statistics class.
- P.Mean: The case of Ranjit Kumar Chandra
(created 2008-09-08). I've been working on developing a series of case
studies in research ethics and research fraud. Progress is slow for a variety
of reasons (my apologies), but I hope to have the first case study on this
website soon. I just ran across a series of webpages published at the Canadian
Broadcaster Corporation website associated with the allegations of fraud
associated with Dr. Ranjit Kumar Chandra. There is video associated with this
site and numerous pictures of the key participants.
- P.Mean: Comparing two proportions out of the
same multinomial population (created 2008-08-05). I am lucky enough to
be researching wine. Specifically I am exploring which components in wine
results in maximised preference. At the moment I am trying to compare
proportions from the same population. N = 68. 8 people most preferred wine 1,
25 most preferred wine 2, 1 most preferred wine 1 and 2, for 34 of
participants their most preferred wine was another wine. I want to see if the
proportion of people that chose wine 1 was significantly different from the
proportion that chose wine 2. I have been recommended to use McNemar's. But I
just don't know how. I found your website which is as close as I have got but
is slightly different. Just wondering if you had any thoughts? Cheers
-
P.Mean: The depths of anti-intellectualism (created 2008-09-05). My
brother-in-law is an avid conservative and often sends me political commentary
that would make Attila the Hun blush. That's actually a good thing, as it
makes me think things through more carefully. He recently made a sarcastic
comment about the lack of experience of Barack Obama ("the guy with the resume
consisting of two good speeches"). It made me think a bit more about a topic
of general interest to me and one that goes well beyond politics: the rise of
anti-intellectualism in the United States. Here's what I wrote back to him in
response.
- P.Mean: Applying the sequence logo
concept to data quality (created 2008-09-04). I am trying to adapt the
logo graph used in genetics to an examination of data quality. I am just
starting this, so the graphs are a bit crude. I took the 1973 NAMCS data set
and calculated entropy for each column of data. This is a massive data set
with 29,210 rows and 85 columns.
August 2008 (16 entries)
- P.Mean: Checks for data quality using
metadata (created 2008-08-28). I have been working on a series of webpages
discussing automated checking for data quality. I have been proposing
cumulative entropy as a measure to track sudden shifts in data entry
characteristics. Entropy can be thought of as the amount of underlying
heterogeneity in a data set, and if this value suddenly shifts upward or
downward, it may be an indication that a change in data entry practices has
occurred. These quality checks can also be applied to metadata (data about the
data itself). There are certain characteristics of a column of data that
usually stay constant. A name, for example, is usually all letters, while an
address is usually a mix of letters and numbers. A sudden shift in the
composition of the data may indicate a problem.
- P.Mean: Controversies with a test for ovarian
cancer (created 2008-08-27). A recent article in the New York Times raises
some interesting questions about diagnostic testing.
- P.Mean: Entropy as a measure of data
quality across multiple variables (created 2008-08-25). In a previous webpage, I discussed the use of cumulative entropy as a
measure of data quality. A sudden shift in cumulative entropy that is not
associated with a significant change in the research design is a possible
marker for a data quality issue. The advantage of entropy is that it can be
used for very large data sets where a context specific analysis of data
quality is difficult or impractical. Entropy can also be cumulated across
multiple columns of data, to look for global shifts.
- P.Mean: Where to look for information in
a controversial area (created 2008-08-20). I am currently researching
vaccinations and the negative effects they may cause. My ex-boyfriend is
against vaccines and I am for them. I was wondering if you could give me some
references to research supporting the autism link not being caused by vaccines
or actually any other theory regarding vaccines being harmful to the body. I
went to a seminar and they told us there is a problem with molecular mimicry
where the body may attack itself looking for a sequence similar to the
disease. Any help you can offer would be greatly appreciated. Desperately
seeking the truth.
- P.Mean: What resources are available
for fellows? (created 2008-08-20). I am on the Core Curriculum
Committee for a fellowship program in medicine (details omitted to protect
privacy) and we are in the process of updating our reading list for the
fellows in training. One of the many topics we are trying to update is some
basic info on statistics. I was going to reference your book, but they are
mostly looking for good review articles that the fellows are more likely to
look at. Specifically, they want info on: Test-performance characteristics:
principles of sensitivity, specificity, predictive value, and ROC analysis. I
was wondering also about referencing your web page as well.
-
P.Mean: Is there a scientific basis for EBM?
(created 2008-08-20). A pair of articles in Chest, along with two
rebuttals examines two sides to the debate over the validity of Evidence-Based
Medicine (EBM). Point: evidence-based medicine has a sound scientific base: P.
J. Karanicolas, R. Kunz, G. H. Guyatt. Chest 2008: 133(5); 1067-71, and
Counterpoint: evidence-based medicine lacks a sound scientific base. M. J.
Tobin. Chest 2008: 133(5); 1071-4; discussion 1074-7.
- P.Mean: Source for sample size formula (created
2008-08-20). Hello, I am looking at your page on sample size calculation,
and I'm curious as to where you got the equation shown there. I can't seem to
find that exact form in Cohen's book, not does it appear anywhere else that
I've looked. Would you happen to know its original source?
- P.Mean: Another inquiry about slash and burn
models (created 2008-08-20). In a binary logistic regression model, do
all variables including the constant need to be significant before you can
include them in the model or is it just the constant that has to be
significant?
- P.Mean: Social networking as a business
tool (created 2008-08-19). The New York Times had an interesting article
about using social networking tools for professional goals: The Social Network
as a Career Safety Net, Sarah Jane Tribble, August 13, 2008. This is a fairly
basic article, which was good for me because I have ignored social networking
sites until recently.
- P.Mean: Can I please skip the Bonferroni
adjustment? (created 2008-08-19). I ran multiple correlation analysis
for abundances, richness (species density), and diversity of different growth
forms in four different landscapes in Colombian Amazonia. My questions is: Do
I have to calculate a Bonferroni test to adjust for each probability?
-
P.Mean: New personal pages (created 2008-08-14). One issue at my old website was that I could not post personal information.
That's a fair policy, and I'm not complaining. But now that I have my own
website, I can add a few personal updates. The main focus of this website will
still be statistics, but since I'm paying for this URL, I will add a few fun
pages about myself. Since I'm a parent, I will be
spending most of my time bragging about my little boy, Nicholas.
- P.Mean: Cumulative entropy as a measure
of data quality (created 2008-08-11). I was talking to someone about some of my work with control charts, and
they asked a question out of the blue. A lot of data sources that might be
candidates for my control chart software has potential problems with data
quality. Did I have any thoughts about ways to screen for poor data quality?
-
P.Mean: Quote on
anti-intellectualism (created 2008-08-08). I want to write an article
about the growing mistrust of experts in our society. A working title is "The
Rising Tide of Anti-Intellectualism." Here's a quote that might be worth
starting out with.
- P.Mean: Interval scale for
count data? (created 2008-08-07). Some of my colleagues insist that the
variable: number of---,(say, quality distractors of an item) is not an
interval scale measure but I feel to the contrary. What do you say and why?
-
P.Mean: Harsh sanctions are only part of the solution (created 2008-08-06).
The Scientist Newsblog had an article about research fraud. Many of the people
who commented on the blog were surprised that there were not any serious
sanctions on the researcher who perpetrated the fraud. There were suggestions
to revoke the graduate degree and to block any future research funding for
this individual. I added a comment, because harsh sanctions are the simplest
suggestion, but there may be more effective approaches that should be used as
well. Here is what I wrote.
- P.Mean:
Resources from the Statistical Consulting Section of the American Statistical
Association (created 2008-08-02). I've talked with a lawyer about setting up an independent statistical
consulting practice and he had several valuable suggestions. One of the more
obvious ones, but one that I had not considered was to check out resources
available through your professional society. The American Statistical Association has a Statistical Consulting Section,
and this section provides quite a few resources. Here's a brief summary of
some of the ones that I found helpful.
July 2008 (23 entries)
- P.Mean: Naming conventions for variables
(created 2008-07-30). For almost all statistical software programs, you
can and should provide variable names for your data. Variable names are a
short descriptive explanation of what resides in each column of data. You
should choose a variable name that is short, concise, and descriptive.
- P.Mean: Business aspects of an independent
statistical consultant (created 2008-07-28). I'm learning quite a bit about the business issues associated with a new
career as an independent statistical consultant. Here are some of the issues
I've had to confront. My apologies to the many readers of this website from
other countries, but these issues are mostly specific to the United States.
- P.Mean: Reliable diagnosis of cataracts
(created 2008-07-28). Can you help with this question? Cataracts of the
eye may be difficult to diagnose, especially in the early stages. In a study
of the reliability of their diagnoses, two physicians each examined the same
1,000 eyes, without knowing the other's diagnoses. Each physician found 100
eyes with cataracts. Does this mean that the diagnoses are reliable?
- P.Mean: Should I abandon FrontPage?
(created 2008-07-24). I've used various iterations of Microsoft FrontPage
since the beginning of the StATS website (that site started in 1998, but the
first "surviving" page dates from 1999). I like FrontPage, but I'm wondering
if I should switch to an open source package.
- P.Mean: Plug for accrual research (created
2008-07-24). I received a request for use of material from my old website.
It's a bit tricky right now, but I hope to have things resolved soon. The
person inquiring was the owner of a company that specializes in clinical
research and clinical data management. I thought it wouldn't hurt to mention
some of the work that Byron Gajewski and I have done in accrual rates. Here's
what I wrote.
- P.Mean: What distribution does this data
come from? (created 2008-07-23). I'm very interested in assessing
distributional fits for empirical data and I've found tidbits of information
here and there but no real good source. Could you recommend a few good
sources?
- P.Mean: My very first meta-analysis
(created 2008-07-23). I am a research student embarking upon a
systematic review and possible metanalysis. I am currently in the process of
developing a protocol. I have been having difficulty understanding the
statistical issues especially since I am not very good at mathematics. Could
you kindly refer me to a source that would help me understand in a step by
step way the concepts needed in doing a meta analysis? For example
heterogeneity and the tests used for it which one is preferred and when; when
to use subgroup analysis and when to use metaregression. I have been reading
the Cochrane hand book for the purpose as advised by my supervisor but have
not been able to understand the concepts. Any help from you would be greatly
appreciated.
- P.Mean: Mail and calendar software I use in my
new job (created 2008-07-23). When I was working for Children's Mercy
Hospital, it made sense to use Microsoft Outlook for my email and calendaring
system. Working, independently, however, I have more choices (more degrees of
freedom, perhaps).
- P.Mean: Survey results from nine out of
thirty six employees (created 2008-07-21). Hi, hope you can help a
struggling grad student in health promotion and education. If I administer a
questionnaire to 9 out of 36 staff members, are the results statistically
significant or is the survey respondents number too small? This is a needs
assessment questionnaire-what the staff feels they need from an educational
standpoint. Or am I floating off course and hopeless? Thank you for your time
and help!
- P.Mean: Can I use some material from your
website? (created 2008-07-18). Someone wrote in asking for permission to
use some of the material on my old website, www.childrensmercy.org/stats. From
December of 2007 through June of 2008, that was an easy thing to do. I had
placed all of the content of my old website under an open source license. In
particular, I used the Creative Commons Attribution 3.0 United States License.
Unfortunately, someone at the hospital raised a question about this and I may
not have had the right, by myself, to declare that those web pages would be
made available under such a liberal license.
- P.Mean: Where can I learn more about Statistics?
(created 2008-07-18). Someone asked me how they could learn more about a
specialized topic in Statistics. They were willing to pay for this, though
they didn't have a lot of money.
- P.Mean: Software for bootstrap and
resampling (created 2008-07-18). Someone asked me what software they
should use as they were learning how to use the bootstrap and resampling
methods.
- P.Mean: Getting on and off various email lists
(created 2008-07-17). In my transition to a new email address, I am having
to sign off and re-sign on to a variety of email lists. Here are the details
of how to do this for those lists related to my work.
- P.Mean: Hire me as a consultant (created
2008-07-15). Starting on July 15, 2008, I am available for statistical
consulting. I have a PhD in Statistics, thirty years of experience, and many
satisfied customers. In the past, I have undertaken independent consulting
assignments using my evenings and weekends, but there was not enough time in
the week to handle this and keep up with my five year old boy. I am taking an
unpaid leave of absence from work in order to devote full time to develop a
career as an independent consultant.
- P.Mean: Undeclared missing code leads
to bad results (created 2008-07-15). I found this ticket in a computer
store many years ago and am just now getting around to showing it. It
demonstrates how failure to declare a missing value code can lead to laughably
incorrect results.
- P.Mean: A misleading bar graph (created
2008-07-15). A regular contributor on EDSTAT-L, found an interesting bar
graph on the DirecTV website.
- P.Mean: Using Google to search through my
website (created 2008-07-15). I want to set up a search function at this
website, but I will have to learn a bit more about Linux servers first. In the
meantime, it is very easy to search through my website (or any website) using
Google.
- P.Mean: How to report a one-tailed Fisher's Exact
test (created 2008-07-12). Thank you for your informative page about
the Fisher's Exact test. Can you please clarify how whether the test was 1 or
2-tailed affects the way that a significant result would be reported?
- P.Mean: Pearson correlation and ordinal
data don't mix (created 2008-07-11). I feel uncomfortable using a
Pearson correlation coefficient for two variables that are measured on an
ordinal scale (for example, 1=unaware, 2=aware, 3=fairly aware, 4=moderately
aware, 5=very aware). But I can't explain why I am uncomfortable with this.
Can you help?
- P.Mean: Where did that standard deviation
come from? (created 2008-07-09). Someone wanted some help with a power
calculation. I gave the standard spiel that you need three things: a research
hypothesis, an estimate of the standard deviation of your outcome measure, and
the minimum clinically important difference. This was for a study looking at
10 exposed patients (recent spider bites) and 30 control patients. I got an
article back in email very quickly, and while it was interesting to read, it
wasn't quite what I needed.
- P.Mean: Cytel software has developed a Poisson
model for predicting accrual (created 2008-07-09). I attended a web
seminar by Jeff Palmer, Cytel Corporation, about Bayesian methods in adaptive
clinical trials. It was a very good seminar, and I should try to summarize
some of the major points sometime. One of the figures, though, caught my
attention. It showed a projection of future accrual based on a Poisson
distribution.
- P.Mean: What does the FDA think about Bayesian
statistics (created 2008-07-08). The FDA is, in general, a cautious agency
(as it should be), but they are allowing newer approaches for establishing
efficacy and safety of new drugs. Many of these new approaches involve
Bayesian methods. A draft guidance "Guidance for the Use of Bayesian
Statistics in Medical Device Clinical Trials - Draft Guidance for Industry and
FDA Staff" is available in HTML format or PDF format.
- P.Mean: Distrust of a Bayesian
meta-analysis (created 2008-07-01). A regular correspondent on the
evidence based health email discussion group (BA) raised some questions about
the use of a Bayesian hierarchical model in a meta-analysis. He was worried
about whether this approach would be appropriate for this type of data.
June 2008 (1 entry)
- P.Mean: How many files are in this new
website? (created 2008-06-24). 1,192, I think. It's tricky to get an exact
count on the number of pages on this website. I could count the number of
files, but many of these files perform support operations and do not
themselves display directly on the page. Other files, of course, are images
used within a page.
This work is licensed under a
Creative
Commons Attribution 3.0 United States License. Need more
information? I have a page with general help
resources.
Files listed below were written by
Steve Simon while working at Children's Mercy Hospital. Although I do not hold the copyright for this material, I am providing these files at my new website as a service, as they are no longer available on the Children's Mercy Hospital website. For more details, consult the information page about my old website.
June
45. 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.
44. 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.
43. 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.
42. 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.
41. 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.
40. 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.
39. 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.
38. 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?
37. Stats: Why it's impossible to keep a
resume current (June 5, 2008). I'm updating my resume and found an
amusing coincidence. It serves as a cautionary statement to never make
assumptions.
May
36. 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.
35. 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.
34. Stats: Why does a Bayesian approach make
sense for monitoring accrual? (May 8, 2008). I'm working with Byron
Gajewski to develop some models for monitoring the progress of clinical
trials. Too many researchers overpromise and undeliver on the planned sample
size and the planned completion date of their research This leads to serious
delays in the research and inadequate precision and power when the research
is completed. We want to develop some tools that will let researchers plan
the pattern of patient accrual in their studies. These tools will also let
the researchers carefully monitor the progress of their studies and let them
take action quickly if accrual rates are suffering. We've adopted a Bayesian
approach for these tools. While a Bayesian approach to Statistics is
controversial, we feel that there should be no controversy with regard to
using Bayesian models in modeling accrual.
33. StATS: Slipped deadlines and sample
size shortfalls in a random sample of research studies (May 7, 2008).
There is a limited amount of data out there that suggests that many
researchers overpromise on the planned sample size and completion date and
underdeliver. About a year ago, I received a small grant to study the proportion of
studies at Children's Mercy Hospital (CMH) that failed to meet the proposed completion deadlines, that failed to
recruit the promised number of patients or both. Here is a brief summary of
these results.
32. StATS: Monitoring refusals and
exclusions in a clinical trial (May 1, 2008). Someone sent me an email asking about the work that Byron Gajewski and I
have done on monitoring accrual patterns in clinical trials. She had been
doing something similar at her job and wanted to see if we could collaborate. In her situation, the major issue was the number of patients who made an initial contact but did not keep
their first appointment, the number of patients who kept the appointment, but refused to sign the
consent form once they realized what the study was about, and the number of patients who did sign the consent form, but who did not
meet the inclusion criteria once the initial screening was done.
April
31. Stats: Directions to my new office (April
25, 2008). I have moved to a new office. It is a modular building just
north of Children's Mercy Hospital. It is between 23rd and 22nd street, just
off of Kenwood Avenue (Kenwood is a small north/south street just west of
Holmes). If you need to get from your office to mine, here are some
directions written by my Administrative Assistant, Judy Champion.
30. Stats: Nomination for the Kreamer Award
for Research Excellence (April 24, 2008). Every year, Children's Mercy
Hospital offers the Kreamer Award for Research Excellence. I plan to apply
this year. I wanted to outline the requirements for the award and offer an
overview of why I would be a good candidate for this award.
29. Stats: Upcoming topics in Poisson
regression (April 24, 2008). I get a lot of questions about Poisson
regression. I feel embarrassed when this happens because my pages on this
topic are woefully incomplete. Everything on my web pages is incomplete to
some extent, of course, but this is an area with the biggest gaps. I have
been planning for quite a while to write more about this topic, and here are
some of the areas I want to discuss.
28. Stats: Evidence Based Medicine for
patients (April 23, 2008). There was an interesting email exchange on the email discussion group
EVIDENCE-BASED-HEALTH@JISCMAIL.AC.UK. The first correspondent (TH) described a
series of workshops that are intended to help patients access and evaluate
health related websites.
27. Stats: I hate bad research examples (April
23, 2008). Someone wrote in asking if I know of any good examples of
research studies that illustrate problems of making false generalizations. I
had to mention my book, of course, which has lots of commentary of actual
publications, most of which are open source and freely available on the web.
For what it’s worth, I do have a pedagogical bone to pick. I believe it is
not a good idea to find a “bad” publication and tear it apart.
26. Stats: A brief overview of
instrumental variables (April 14, 2008). People will often ask me
questions that are outside my area of expertise. Yes, I know you're shocked
to hear this, but there are lots of areas of statistics where I only have a
vague understanding. One of these questions was about instrumental
variables. I could only offer a vague explanation, but I hope that is better
than no explanation at all.
25. Stats: What does one-third of the way
through a study mean? (April 6, 2008). Someone asked me a very good
question regarding interim analysis. If the call for an interim analysis is
specified as occurring one-third (and two-third) of the way through a study,
what does that mean. In a study with 60 patients lasting a full year, does
that mean until 20 (40) patients have arrived, or does it mean until 4
(8) months? Also, if you are counting discrete events, such as adverse
events, does it mean one-third of the expected number of adverse events?
24. Stats: Calculating statistics on whole
numbers (April 4, 2008). Dear Professor Mean, I have some data that
has a resolution of 1, meaning that these are whole numbers only and no
decimals. When I calculate the mean and standard deviation, should I report
the actual value or should I round it?
23. Stats: Why we need case studies of
research ethics in graphic novel format (April 3, 2008). I was asked to
provide some justification for a project I am working on, case studies in
research ethics using a graphic novel format. Here is what I wrote.
22. Stats: An alternative to the p-value
(April 3, 2008). A discussion on edstat-l concerned a statistic called
p-rep. I had not heard of this statistic before, but at least one journal is
calling for its use in all papers published by that journal.
March
21. Stats: What sort of statistical
training is needed for basic scientists? (March 29, 2008). Someone wrote
to a mailing list sponsored by the American Statistical Association asking
about what resources to use in a statistics class aimed at basic scientists
(as opposed to public health students and clinical scientists). I offered a
few general recommendations.
20. Stats: Statistics for Boards
(March 25, 2008). I was asked to give a talk to the medical residents
with the title "Statistics for Boards". Many health care professionals need
to take boards or other certifying examinations during their training and
afterwards to certify or re-certify their skill in an area. These boards
often ask some basic statistics questions. A common theme appears to be, what
statistic should I use in what situation. The answer often depends on what
the predictor variable is and what the outcome variable is.
19. Stats: When a client asks for a bad
analysis (March 24, 2008). I received an email from someone who was being
asked to perform a subgroup analysis that is likely to produce confusing and
counter-intuitive results. I was asked to help draft some language to convince
the client that this was a bad idea.
18. Stats: Presenting unadjusted and
adjusted estimates side by side (March 24, 2008). Someone on the Medstats
discussion group asked about reporting the analysis of a model without
adjustment for covariates along with the analysis adjusted for covariates.
What is the purpose of reporting the unadjusted analysis?
17. Stats: Presenting unadjusted and
adjusted estimates side by side (March 24, 2008). Someone on the Medstats
discussion group asked about reporting the analysis of a model without
adjustment for covariates along with the analysis adjusted for covariates.
What is the purpose of reporting the unadjusted analysis?
16. Stats: Simplifying repeated
measurements (March 12, 2008). I received an email inquiry about a project
that involved four repeat assessments on 10 different subjects. The question
started out as, is my sample size 10 or is it 40?
15. Stats: ROC curve for an imperfect gold
standard (March 12, 2008). Someone asked me about how to use an ROC curve
if you have more than two categories. Apparently the gold standard that the
researchers were using was known to be imperfect, so they wanted an
intermediate category (possible disease).
14. Stats: Calculating NNT for observational
studies (March 3, 2008). Recent discussion at the Evidence Based health
list centered on the calculation of NNT in a case-control study. While it is
indeed possible to do so, I have always been a bit curious why NNT and NNH are
computed almost exclusively for randomized studies and why they are rarely
used for observational studies. No one says this explicitly, but I suspect
that the reason is that the NNT and NNH lead to problematic interpretations in
observational studies.
February
13. Stats: Assessing the assumption of
an exponential distribution (February 25, 2008). The following 41
observations: 8, 2, 26, 29, 1, 2, 11, 8, 0, 5, 10, 1, 4, 9, 12, 3, 6, 5, 2,
12, 1, 5, 3, 5, 7, 0, 2, 8, 3, 3, 1, 0, 4, 8, 1, 8, 12, 0, 6, 1, 5, represent
waiting times that we suspect follow an exponential distribution. There are
several ways to examine this belief, and the simplest way to to draw a Q-Q
plot for the exponential distribution.
12. Stats: What is a critical value?
(February 22, 2008). Someone wrote in asking about the difference between
a p-value and a critical value.
11. Stats: Difficulties in generalizing
research (February 15, 2006). I found this information thanks to an email in the
Evidence-Based Health email discussion group. Someone asked if there was any empirical
evidence that the setting of a study (e.g., primary versus secondary care) could influence
the results of the research. Intuitively, you would suspect that this would be the case,
because the types of patients who show up at a primary care clinic are quite different than
those who show up at a secondary or tertiary care center.
10. Stats: Multiple methods for
computing percentiles (February 13, 2008). A recent discussion on the Medstats group highlighted some
of the confusion about computing percentiles. I use a simple formula. If you
want the pth percentile of a set of n observations, select the p(n+1) value
from the data. If p(n+1) is not a whole number then choose a value halfway
between the two adjacent values.
January
9. Stats: Does prevalence affect sensitivity (January 31, 2008). Dear Professor Mean, Does lowering the prevalence of a disease have an effect
on sensitivity?
8. Stats: A simple example of a Kaplan-Meier
curve (January 24, 2008). In response to a query, I wanted to write up a
simple example of how to calculate survival probabilities when you have
censored data. It is adapted from Chapter 6 of my book, Statistical Evidence
in Medical Trials. I have updated and simplified the example, for possible
use in a second edition of the book, if I am so lucky.
7. Stats: How do you analyze safety data
(January 22, 2008). Someone on the MedStats email discussion group asked
about how to analyze adverse event data. He noted that adverse event data is
not one of the primary or secondary outcome measures, and wondered if it would
be appropriate to provide statistical analysis of this data. Adverse events
(and safety data in general) represent a special type of analysis that does
not fit in well with the listing of primary/secondary outcomes. The main
reason for this is the number of possible adverse event categories is very
broad and it is not always possible to anticipate in advance what type of
adverse events are of greatest interest.
6. Stats: Finding only the important
studies (January 21, 2008). Someone wrote into the MedStats listserv
asking about a process that they had chosen to select "important" articles in
a particular research area. This was, I presume, a qualitative summary of
interesting results in a broad medical area rather than a quantitative
synthesis of all available research addressing a specific medical treatment.
The reason I suspect this is that the person mentioned that they had used the
statistical significance of the studies as a filter and eliminated any
negative studies from further consideration.
5. Stats: What have you changed your mind
about (January 18, 2008). A group called The Edge (www.edge.org) asks a
question each year of prominent scientists and a few select non-scientists.
This group represents from a broad range of backgrounds and disciplines
and provides an interesting variety of responses. Questions asked in the past
include: What are you optimistic about?, What is your dangerous idea?, and
What do you believe is true even though you cannot prove it? Just recently,
this group published responses to the 2008 question: What have you changed
your mind about? Why? Several of the responses touch directly or indirectly
on Statistics.
4. Stats: Too much power and precision?
(January 9, 2008).There was a discussion on EDSTAT-L about studies with too much power
and precision. You can indeed have too much
power/precision, and here is a pragmatic example.
3. Stats: The post-modern assault
on evidence-based medicine, part 3 (January 7, 2008). I have volunteered
to give a talk for a group of statisticians which discusses an article I am
preparing: "The post-modern assault on evidence-based medicine." I might give
this talk on the first Monday in February or the first Monday in March. Here
is a tentative abstract.
2. Stats: What statistic should I use
when? (January 4, 2008). Someone was asking about a multiple choice
question on a test that reads something like this: A group of researchers
investigating in patients with diabetes on the basis of demographic
characteristics and the level of diabetic control. Select the most
appropriate statistical method to use in analyzing the data: a t-test, ANOVA,
multiple linear regression, or a chi-square test. This is one of the more
vexing things that people face--what statistic should I use when.
1. Stats: Type III
error (January 3, 2008). Dear Professor Mean, What is the
definition of a Type III error?