Local resources
Advice
Modest Advice to grad students
When I started my Masters degree the ranking grad student in our group
handed me a many-generation photocopy of a pair of articles, Steve Stearns' Modest
Advice to grad students, and Ray Huey's reply
to Stearns.
Pdf versions of these (they deserve printing and posting above the
desk of every grad student) are available here (advice),
and here (reply).
More advice
The best academic advice I ever got was: "Spend at
least an hour every day on the manuscript closest to publication".
Do not burn yourselves out. Be as I am - a reluctant enthusiast, a
part-time crusader, a half-hearted fanatic. Save the other half of
yourselves and your lives for pleasure and adventure. It is not
enough to fight for natural land and the west; it is even more
important to enjoy it. While you can. While it's still there...
Enjoy yourselves, keep your brain in your head and your head firmly
attached to the body, the body active and alive, and I promise you
this much: I promise you this one sweet victory over our enemies, over
those desk-bound men with their hearts in a safe deposit box, and their
eyes hypnotized by desk calculators. I promise you this: you will
outlive the bastards. --Ed Abbey.
The best climber is the one having the most fun. -- Alex Lowe.
You are a fool if you don't climb Mt. Fuji; but you are a bigger fool if you climb it twice. -- E Matsuzawa.
Perpetual devotion to what a man calls his business is only to be sustained by perpetual neglect of other things. And it is not by any means certain that a man's business is is the most important thing he has to do. -- RL Stevenson.
Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius – and a lot of courage – to move in the opposite direction. -- E.F. Schumacker.
Administrivial denial of service attack / Declaration of Organizational Bankruptcy
"The effect of new technology has been to draw even senior managers
into minutiae. People paid to think and lead now spend much of their
time typing and responding to or sending an endless stream of
unnecessary messages, simply because communications technology invades
every second and every corner of their lives. This bureaucratization
of both the leadership and the creative process makes thought seem
irresponsible and clear action seem unprofessional. It provides a
sensation of activity while creating a broader sense of
powerlessness. This is what used to be called being nibbled to death
by ducks." -- John Ralston Saul
Hurd's Laws
Hurd's First Law: All good things are followed by laundry.
Hurd's Second Law: If you can't fit your rock into a van, you're not ready to tour. (inspired by the 118 tractor-trailers required to bring U2 to play Edmonton)
Hurd's Third Law: The dumber the organism, the better its behaviour will be predicted by game theory.
Some statistical functions in R
Little things here so I can find them
sem <- function(x){sd(x,na.rm=T)/sqrt(length(x[!is.na(x)]))}
combine.p(): Combine p.values from separate tests using one of the
Fisher, Winer, or Stouffer methods.
Rationale: A set of tests, all testing the same hypothesis, may
all approach significance without any of them being significant.
These methods calculate the probability that the null hypothesis is
true given the p values of a number of separate tests.
Bugs: The Winer and Stouffer methods don't deal with effects in
opposite directions correctly (neither does Fisher, but that's the way
Fisher's method is supposed to be).
Here's the combine.p function, combine.p.r,
and combine.p.Rd
Examples of combine.p() can be found in the combine.p.Rd file.
get.d(): Calculate effect size Cohen's d.
Rationale: Because it's not there.
Bugs: Once needed a signed version, but I cannot remember why now.
Here's the function (includes an option for dependent samples pooled standard deviation sent in by SylviaDKreibig), cohens-d.r.
boot.d.ci(): Bootstrap a confidence interval on Cohen's d.
Rationale: I seem to wind up doing this a lot, so I wrote a function...
Bugs: Purges NA by default, may throw hissy-fits with lots of NAs otherwise, includes old get.d() version.
The function, boot-d-ci.r, defaults to 95% confidence intervals with 1000 bootstrap resamplings, arguments override the defaults.
threeD.toi.test(): Tests of independence for three dimensional data matrices
Rationale: The native R chisq.test() and the log-likelihood
ratio test above do not handle three dimensional data matrices. This
function defaults to log-likelihood ratio, but adding the
argument 'method="c"' will invoke chi-squared.
Bugs: Omnibus tests only, no corrections, poor documentation,
inexplicably ugly output.
Here's the threeD.toi.test function, 3d-toi.r,
residuals.with.na(x,y): ugly kludge wrapper to return a vector of residuals with NAs where they belong
Rationale: Why isn't this default behaviour? Maybe there is an obvious option to do this that I have overlooked...
Bugs: Uses a for() loop, which is ugly bad.
Here's the function, resid-na.r,
latex2ps shell script for Nautilus
latex2ps, a
nautilus
script to latex & dvips a file.
1) Riding the crowd,
keeping an eye on the bouncer (late '80s DOA show?), photo Shawn Scallen 2) Pete's head, West
Coast Trail, Photo by KJB; 3-4) Sixpack Fleataxi; 5) Brian Kuchinka on
approach to Slesse's South Ridge Route; 6) Bouldering in Sollentuna
with T Ekefalk; 7) KRW in 24hr daylight of Lappland; 8) Short walk in
the snow, Photo GKH; 9) Why rush a good thing? traverse of Mt. Edith
Cavel in 24hr 15min, with SNW & Zoe, Photo Zoe. 10) Topping out on
the Grand Sentinel with Sisterijah (white helmets), Photo KRW. 11)
Pete Hurd on his first skateboard, 1976 12) Professor
Hurd on his new skateboard, 2004
Rtips,
(not StatsRus), a super-useful list of frequently asked
questions about R usage.
Quick-R for SAS, SPSS, Stata users.
