|
OK!
|
Nov. 24th, 2009 @ 04:28 pm
|
|---|
|
Over the past year I've spared y'all from around a dozen frantic posts explaining how I spent days preparing a huge data analysis presentation for the S---- study investigators, and then at the meeting they (very politely) berated me for 45 minutes over the first three slides, and declared they didn't want to see the rest. Not that I would have kept on talking about it, I mean that this happened on three or four separate occasions, with their unspoken and contradictory requirements changing each time. But I am going to post about this.
A couple of months ago, I insisted that we sit down and write out what the next set of analyses were going to be. Then I verified it by email, found out they'd changed their minds, wrote up an alternative, verified that, talked with a fellow postdoc about it, met with one of the investigators, sent it out for approval again, made the requested changes, sent it out again, made an interim presentation on exactly what I was going to do, and over the weekend, I did it.
Today I presented on it. It was a success and everyone was pleased! I can move on and maybe get something out of this at last.
I'm making the people in charge of this study sound like bad guys, but that's not really true (they do deserve some cussing out, but I've already vented elsewhere). This is a huge federal grant, the equivalent of several R01s rolled together. There are five primary investigators and two statisticians, each with their own approach to the data. It's hard to get them all in the room at the same time, let alone to get everyone to agree on an analysis. But I managed to clear things with the ones who really mattered -- who happen to be all the same ones who were at today's meeting. I triumphed over a really difficult situation, and it's to everyone's benefit!
A few months ago, I put together an online questionnaire study. UCSF doesn't have any undergrads, so ![[info]](http://l-stat.livejournal.com/img/userinfo.gif) dragonpaws agreed to help out and run it on students from her school. I sent her the IRB forms, and two days later it was approved! Then I sent her the website, and a day and a half later we had 100 responses! At UM that would have taken at least a month!
This afternoon John IMed me to show me an exam that had a series of Fermi problems... except that instead of estimating the answer, you were supposed to estimate your 94% confidence interval. I asked what it could possibly mean to be 94% confident about something -- of all the (probability-weighted) scenarios you can imagine, 94% of them give an answer within this range? John said maybe, but if you're thinking in bayesean terms then it just means that you're 94% confident. We got into another long conversation about bayesean statistics. I felt very clever because I kept coming up with ways to understand things in terms of the frequentist paradigm I'm already trained in, and kind of foolish, because part of the point is that with bayesean statistics, you don't have to mess about with all these imaginary distributions.
Then I showed our recruiter =concatenate(), saving her countless hours of cutting and pasting.
It's been a good day!
|
|
I've been surprisingly busy the last couple weeks, since being laid off. It hasn't been productive in the sense of earning an income, but I'm getting a number of things set up and developing connections with a number of people in the area.
I have taken this opportunity to set up my own business, as well. The website is still deeply under construction, but I am trying to develop a new blog which will include my various architectural writings, as well as posts about architectural projects and other personal work.
Since I am not renewing my LJ account, and am going to fall back to freeloader status (or whatever they call it here), more of my writing will probably be focused at the new blog. I haven't decided whether or not to port over all the old entries from here, or not. Some stuff will fall away at the end of the year, so I need to archive it someplace.
If you are one of the handful of people still reading this and want to keep following what I'm up to, I suggest you start checking in at psproefrock.wordpress.com. I think eventually most of my updates will be shifting over there. |
|
"But while the logit transformation yields the logarithm of the odds of event occurrence, the clog-log transformation yields the logarithm of the negated logarithm of the probability of event nonoccurrence." |
|
|
Nov. 22nd, 2009 @ 10:26 pm
|
|---|
|
Shorter Sopranos: Happy Families are all alike; every unhappy Family is unhappy in its own way. |
|
http://postsecret.blogspot.com/2009/11/hopeline.html
-----Email Message-----
Subject: thank you
Frank,
This past Friday night I found myself in a black hole of depression and I didn't know how I was going to make it through the night. Not knowing where to turn and feeling like I couldn't stop. I remembered seeing the Hopeline phone number in the front of your book. I talked with someone there for 2 and a half hours and I truly feel that they saved my life.
Thank you for the book, thank Hopeline for being there, and thank the people that send in their postcards so that others know they are not alone with their secrets.
-Casie (with permission)
|
|
|
Nov. 21st, 2009 @ 07:56 am
|
|---|
|
If you went by what I regularly call my cats, you'd think they were named "Little One" and "Mister Prettycat". |
|
"...Since the quotation marks are invisible, they are sometimes hard to see..." |
|
The problematic answer in real life tends to be, "Not yet, maybe, we don't know when."
Will I be able to get a Verizon plan for the iPhone? See above.
Will the apps available for Android phones ever rival those for the iPhone? See above.
(What I truly dread is getting an iPhone with a two-year AT&T contract in December, then having hearing them announce that they're adding Verizon in January.)
Something I don't recall: have any of y'all attempted to use an iPhone from within my apartment? That particular area of coverage is rather more important to me than many others. |
|
A partial inventory of people who don't like Malcolm Gladwell. The criticisms seem to fall into at least one of these three categories:
1) Everything Malcolm Gladwell writes is obvious.
2) Everything Malcolm Gladwell writes is wrong.
3) Everything Malcolm Gladwell writes follows an exceedingly predictable formula that's just so fucking readable...that crafty motherfucker.
The third point is definitely true, but I think the first two are only sometimes true. Says a blogger for the Village Voice:
"Whenever we see a piece by Malcolm Gladwell in the New Yorker, we know exactly what's going to happen. We'll be entertained and titillated as the frizzy-haloed essayist takes us once again into the trippy world of statistics and ideas, only to confront us with evidence that some axiom or law we thought was on solid ground is in fact misleading and counterproductive."
Exactly (and man oh man is his halo frizzy.) But that experience you just described is kind of neat, and he's only full of shit maybe 30% of the time, which is pretty good compared to most people. And that ketchup essay was the bomb. |
|
Anyone ever done it? I understand there's tranquilizers involved.
Someone at work recommended a good boarding place to me. I'm not sure whether, overall, they'd find it more stressful to do lengthy, drugged travel and a new place, but with lots of space and people paying lots of attention to them; or a short trip but confined to a small space with lots of other cats around for the duration.
My family would enjoy having them around, I have no doubt (plus, I asked.) I'll call the airport and vet tomorrow and figure out what the ups and down of it are.
Edit: US Airways policies basically make this a nonstarter. For starters, one passenger can't bring two pets. |
|
![[info]](http://l-stat.livejournal.com/img/syndicated.gif){kind=small}
