Re: Ask the mineshaft

1

Write who's who on the back of them?


Posted by: parsimon | Link to this comment | 03-10-09 12:22 PM
horizontal rule
2

What is the least effort way to preserve photos?

Lure nattargramatt to a meetup, kidnap him and make him do it. If you decide to go that route, you might want to redact this comment.


Posted by: Jesus McQueen | Link to this comment | 03-10-09 12:24 PM
horizontal rule
3

1: but then store them how?


Posted by: heebie-geebie | Link to this comment | 03-10-09 12:25 PM
horizontal rule
4

2: I understand fried haggis with curry sauce forms an effective bait.


Posted by: LizardBreath | Link to this comment | 03-10-09 12:26 PM
horizontal rule
5

1 sounds like such a sound idea, and yet... What of the stacks of photos including people whose names you can no longer remember?


Posted by: Di Kotimy | Link to this comment | 03-10-09 12:31 PM
horizontal rule
6

3: Honestly, in boxes. That's if you want the least-effort method. If you're afraid of humidity, then you've got to put them in plastic somehow, and that's just going to be work.

Seriously, first thing first, you should really label them if nothing else. I went through my mom's photos recently and was very grateful to her for that.


Posted by: | Link to this comment | 03-10-09 12:31 PM
horizontal rule
7

You should just throw them away and start over. Really. I lost a bunch of boxes in a move, and after a couple of sentimental twinges, I haven't missed them. Now, whenever I'm faced with a pile of old junk I can't figure out how to organize, I toss it and pretend that it was in one of those lost boxes.


Posted by: jms | Link to this comment | 03-10-09 12:32 PM
horizontal rule
8

I'd go find a big craft store, with scrapbooking stuff, and talk to someone who works there and seems knowledgeable about stuff. I'm guessing there are obvious solutions like layering them in cardboard boxes interleaved with acid-free tissue paper or something.


Posted by: LizardBreath | Link to this comment | 03-10-09 12:32 PM
horizontal rule
9

6 was me.


Posted by: parsimon | Link to this comment | 03-10-09 12:32 PM
horizontal rule
10

4: Under a cardboard box propped up with a stick. Foolproof.

HG, are you looking to save them in their current form, or digitize them?


Posted by: Jesus McQueen | Link to this comment | 03-10-09 12:32 PM
horizontal rule
11

11: Like catching tigers with tunafish sandwiches.


Posted by: LizardBreath | Link to this comment | 03-10-09 12:34 PM
horizontal rule
12

Further to 7, I pretty much do the same thing with computer files that I've failed to organize. After a while, my hard drive becomes a swamp of unnavegable junk. Rather than try to organize it, I pretend I accidentally had a hard drive failure. Which happens to people all the time you know, so, it's no big deal.


Posted by: jms | Link to this comment | 03-10-09 12:36 PM
horizontal rule
13

10: Probably in their current form, but if someone argued convincingly that digitalizing them would be easier, I'd go for it.


Posted by: heebie-geebie | Link to this comment | 03-10-09 12:36 PM
horizontal rule
14

13: scanmyphotos.com does this.


Posted by: jms | Link to this comment | 03-10-09 12:37 PM
horizontal rule
15

And it would be easier, because you just mail them the photos and they do it for you.


Posted by: jms | Link to this comment | 03-10-09 12:38 PM
horizontal rule
16

The least-effort way to preserve the photos would be to drop them off at a shop that will digitize them.


Posted by: Chris Conway | Link to this comment | 03-10-09 12:39 PM
horizontal rule
17

Damn, pwned.


Posted by: Chris Conway | Link to this comment | 03-10-09 12:40 PM
horizontal rule
18

16: That actually strikes pretty close to what I was hoping for.

Does a place exist that will put all your photos in albums, with little blank labels so that you just have to go through and label them? That would be so wonderfully awesome.


Posted by: heebie-geebie | Link to this comment | 03-10-09 12:41 PM
horizontal rule
19

17: Fortunately I read threads from the bottom up, so I saw your comment first.


Posted by: heebie-geebie | Link to this comment | 03-10-09 12:41 PM
horizontal rule
20

I think you can get albums at scanmyphotos, but it is spensiver.


Posted by: jms | Link to this comment | 03-10-09 12:44 PM
horizontal rule
21

Are you bothered about the photos or the images? If it's the images you want to store, set yourself to scan a manageable number every day and establish a routine where this is something you do for, say 15 minutes before dinner. My SiL did this very successfully, and it took less time than you'd think.

(Don't trust Flikr or Picasa, keep them on an external drive and upgrade it from time to time.)

If you want to keep the actual photos, boxes is just about it, I fear. Could you persuade your alma mater to start collecting your archive for when you become rich and famous?

Like catching tigers with tunafish sandwiches.

I chipped a tooth eating a tunafish sandwich today. I'll never understand how.


Posted by: OneFatEnglishman | Link to this comment | 03-10-09 12:46 PM
horizontal rule
22

That's some crunchy celery


Posted by: TJ | Link to this comment | 03-10-09 12:47 PM
horizontal rule
23

20: Wouldn't that be: digitize, print, then albumize? Presumably, heebie would be content with analog albums.


Posted by: Chris Conway | Link to this comment | 03-10-09 12:48 PM
horizontal rule
24

and establish a routine where this is something you do for, say 15 minutes before dinner while breastfeeding.


Posted by: Jesus McQueen | Link to this comment | 03-10-09 12:48 PM
horizontal rule
25

Rather than try to organize it, I pretend I accidentally had a hard drive failure. Which happens to people all the time you know, so, it's no big deal.

Suddenly I have doubts about a collaborator's "I did the calculations and wrote my section of the paper, but then my laptop and notebook got stolen, so it'll be a while before I can re-do it all" excuse.


Posted by: essear | Link to this comment | 03-10-09 12:49 PM
horizontal rule
26

23: Yeah, unless you're going to pay someone else to do the digitizing then printing and albumitizing, it seems like the maximal rather than the minimal amount of work. I guess you could also digitize and put them into virtual albums organized on your computer, but again ...


Posted by: parsimon | Link to this comment | 03-10-09 1:00 PM
horizontal rule
27

Back when photos were analog, I put them in albums. Then I put the albums on a shelf and ignored them forever.


Posted by: essear | Link to this comment | 03-10-09 1:07 PM
horizontal rule
28

Then I put the albums on a shelf and ignored them forever.

I actually love going through old photo albums.

If nothing else, this thread is helping me refine my question: can I pay someone to put my photos into albums?


Posted by: heebie-geebie | Link to this comment | 03-10-09 1:20 PM
horizontal rule
29

Keep the photos in boxes. The effort/reward ratio just isn't good enough to put those things into albums.


Posted by: lemmy caution | Link to this comment | 03-10-09 1:30 PM
horizontal rule
30

28: I bet you can. While I find it completely alien, scrapbooking is a huge hobby industry; somewhere within a hundred miles of your house, there's a big store that sells 'archival quality' scrapbooking supplies with the other crafty stuff. And the people who work there will either do that sort of thing themselves, or will know some local person who'd do it if you paid her. I bet it'd be very expensive -- it's a timeconsuming process. But if I wanted photos put in albums, that's what I'd do.


Posted by: LizardBreath | Link to this comment | 03-10-09 1:34 PM
horizontal rule
31

can I pay someone to put my photos into albums?

I know this guy Craig. He keeps a list, see...


Posted by: Wrongshore | Link to this comment | 03-10-09 1:36 PM
horizontal rule
32

Or you can scan your photos into, say, iPhoto, and order albums of them. That way you'd have albums and digitized versions (and backups, with little extra effort, in case something happened to the originals).


Posted by: Jesus McQueen | Link to this comment | 03-10-09 1:38 PM
horizontal rule
33

Oh, my mom mentioned him the other day. In one of those wonderful cross-generation conversations, she says, "Well, you know my friend's son and his wife bought a lot of high-quality baby furniture off of a website called Craig's List" with that signature over-articulation that the old folks do.


Posted by: heebie-geebie | Link to this comment | 03-10-09 1:40 PM
horizontal rule
34

Did you mention that Craig apparently knows lots of hookers?


Posted by: Jesus McQueen | Link to this comment | 03-10-09 1:46 PM
horizontal rule
35

33: My mom once asked me how she might get started using these computer chat rooms, where she should go online to find them, whether I thought she'd like them.


Posted by: parsimon | Link to this comment | 03-10-09 1:47 PM
horizontal rule
36

In an unrelated matter, has anyone pointed out that even the ersatz liberal New Republic feels compelled to give Prof. Rauchway a shout-out when the subject is Amity Shlaes?


Posted by: politicalfootball | Link to this comment | 03-10-09 1:52 PM
horizontal rule
37

goddamit - meant to post that at EOTAW


Posted by: politicalfootball | Link to this comment | 03-10-09 1:52 PM
horizontal rule
38

Given the economy you can probably get a couple of Ph.D.s to do the scut work for some tuna sandwiches.

I scanned about a thousand slides and prints going back to the Fifties recently, one can surf the net, scan, and organize photos at the same time. Scanning itself is slightly less exciting than watching corn grow so having something else to do at the same time prevents madness.

I have the files on local hard drives and on Picasa, they won't disappear all at once.

There's plenty of info on photo preservation on photo.net.


Posted by: Biohazard | Link to this comment | 03-10-09 1:58 PM
horizontal rule
39

then store them how?

Parsimon is right. The simplest way to keep photos is make notes on the back and throw them in a box. Companies like Light Impressions and Gaylord sell archival photo storage supplies if you care about that sort of thing. (Since this is a Mineshaft question, obviously you should do your business with Gaylord.)


Posted by: Populuxe | Link to this comment | 03-10-09 2:06 PM
horizontal rule
40

OneFatEnglishman said: "Don't trust Flikr or Picasa, keep them on an external drive and upgrade it from time to time."

Also don't trust your spouse not to drop the external hard drive on the cement basement floor and lose all its contents.



Posted by: Nancy | Link to this comment | 03-10-09 2:24 PM
horizontal rule
41

Be careful writing on the backs of the photos, since certain ink/paper combinations will cause the writing to transfer onto the next photo in the stack (BTDT, bought the special cleaner at a camera store).

If your desire is just to keep the photos, then boxes are hard to beat - but they have no redundancy. You could get second prints and store them somewhere else (parents' house or similar). Acid-free paper degrades more gracefully than just about everything else.

If your desire is to keep and label the photos, archival albums are what you need - but again, no redundancy.

If your desire is to digitize them, scan for N minutes a day while you breastfeed, comment on Unfogged, whatever. If you digitize them you should: 1. store them on an external hard drive; 2. store them on a second external hard drive as well; 3. Upload them to flickr or something similar.

Assuming these are color prints they're likely to last 80-100 years if stored properly, and much less than that if not. Hard drives last about five years on average. CDs and DVDs supposedly will last up to 100 years if you buy the right media and store it properly, but I'd expect to re-archive everything digital every few years for, um, ever.

For the ultimate archive, I would: 1. Digitize everything. 2. Store the original prints in archival paper boxes with acid-free paper between them, and put the labeling on that acid-free paper. 3. Store a secondary set of prints offsite where a single natural disaster or Gojira attack will not wipe them out. 4. Make at least a second copy of the digital archive and store that elsewhere. 5. Upload everything to a third-party site. 6. Rotate my offsite and local backup copies every year. 7. Not bother with CDs or DVDs.

My dad likes to go through and create photo albums, and I've been known to look at them when I'm visiting my parents. This is the biggest argument against boxes, since boxes of photos are less inviting than albums that have been labeled.

I would assume unless you're really, really busy or your time is worth quite a lot that it's not worth paying somebody (grad student, scrapbooker, neighbor's kid) to put things into albums for you unless you're getting some other sense of satisfaction out of the deal (neighbor's kid saving up to go to Africa and put up mosquito nets or something like that).


Posted by: fedward | Link to this comment | 03-10-09 2:27 PM
horizontal rule
42

The 'photos in the shoebox' approach is totally underrated. And that way, people don't have to crowd around a silly album, either.


Posted by: LordKrishna | Link to this comment | 03-10-09 2:33 PM
horizontal rule
43

28: If nothing else, this thread is helping me refine my question: can I pay someone to put my photos into albums?

You know, thinking about this, if you do pay someone to do it, you're probably going to want to sort them to some extent anyway. Someone else is obviously going to have no idea what should go with what.

My photos are 'sorted' in a jumble in separate boxes roughly by era, like: before I was born; childhood; high school; college; etc.

If someone else puts them in an album, I'd at least want to them to be in the kind of album where I can remove them and swap them around. It'd drive me bonkers to find that Hey, those are all interspersed with those! But those people didn't even know each other at all, that's when I lived in x, but that's like 7 years later when I lived in y, after the horrible breakup with so-and-so, who hated the person whose picture his is right next to here!

And, uh. There may be some photos you want to remove altogether. I'm actually not sure I'd want someone else looking at my photos one by one.


Posted by: parsimon | Link to this comment | 03-10-09 2:35 PM
horizontal rule
44

28: Oh yes. Excellent points.

So it sounds like I'm not actually damaging the photos by leaving them in these boxes, since they're not in an attic or basement. That at least makes me feel better.


Posted by: heebie-geebie | Link to this comment | 03-10-09 2:39 PM
horizontal rule
45

4 sounds like tasty bait.

Actual paper photos keep pretty well. Colour photos not as well as black and white, though.

41 is pretty much bang on. Archival paper boxes with acid-free paper. Plus a set of scans with an off-site backup.

Scanning prints, btw, is easy and quick and doesn't require an expensive scanner. If you have the negatives you'll get better results from scanning the negs but it's more fiddly and you really need a slightly better scanner.

Standard 'archival' practice at the moment is to make 600dpi scans at 100%. Although you can go higher if you have the disk space and if the prints are good. Personally, I probably wouldn't bother.


Posted by: nattarGcM ttaM | Link to this comment | 03-10-09 2:40 PM
horizontal rule
46

Not 28, although I'm sure 28 was an excellent point as well. I meant to nod to 43.


Posted by: heebie-geebie | Link to this comment | 03-10-09 2:40 PM
horizontal rule
47

I am hugely in favor of paying people to do things that won't actually get done. There's a whole class of things that all my willpower and good intentions and lists won't achieve, and you can usually tell them because they're the same trivial but hovering project for years.

If money, especially one-time money, can get those done, I think that's a great choice.

Pay a scrapbooker, sit down with her, write out labels on the pictures she hands you and nurse your baby. Just having an appointment with her will help.


Posted by: Megan | Link to this comment | 03-10-09 2:45 PM
horizontal rule
48

although I'm sure 28 was an excellent point as well.

Sure it was, heebie. It was you.


Posted by: OneFatEnglishman | Link to this comment | 03-10-09 2:46 PM
horizontal rule
49

Oh, then it was RIGHT!


Posted by: heebie-geebie | Link to this comment | 03-10-09 2:47 PM
horizontal rule
50

Megan is eminently sensible. That is pretty much what I wanted to hear.


Posted by: heebie-geebie | Link to this comment | 03-10-09 2:51 PM
horizontal rule
51

Standard 'archival' practice at the moment is to make 600dpi scans at 100%.

For both prints and negatives? Or higher for negatives?


Posted by: Jesus McQueen | Link to this comment | 03-10-09 2:52 PM
horizontal rule
52

I am hugely in favor of paying people to do things that won't actually get done. There's a whole class of things that all my willpower and good intentions and lists won't achieve, and you can usually tell them because they're the same trivial but hovering project for years.

Isn't that why people go to grad school? I'm also reminded of a quote from Robert Benchley I've had since the days of .plan files: "Anyone can do any amount of work provided it isn't the work he is supposed to be doing at the moment."

I tend to get projects like this done in bits and pieces when I'm putting off doing something else. The photograph archive is an excellent choice for job-avoidance because the photographs themselves are already in distinct units (mine are all still in the folders/envelopes/packets from the lab that printed them). You can do precisely as many units as you need to fill the time you should be spending on something more urgent. It's the most satisfying kind of procrastination.


Posted by: fedward | Link to this comment | 03-10-09 2:55 PM
horizontal rule
53

Digitize. Easier to share. Put the goofiest ones in flickr so your friends can laugh at them.


Posted by: CharleyCarp | Link to this comment | 03-10-09 3:02 PM
horizontal rule
54

||

Comment number 1000000 is coming soon to an Unfogged near you.

|>


Posted by: fedward | Link to this comment | 03-10-09 3:04 PM
horizontal rule
55

Time to close the blog.


Posted by: apostropher | Link to this comment | 03-10-09 3:06 PM
horizontal rule
56

Wow? Is there a prize?

(horrifying apo link in 3, 2, 1...)


Posted by: Jesus McQueen | Link to this comment | 03-10-09 3:11 PM
horizontal rule
57

1. Buy a flatbed scanner. (Also, some extra memory storage for your computer, because, godloveyou, you will need it.)

2. Acquire or borrow some entertaining friends; alternatively, acquire or borrow a whole lot of movies you really want to watch.

3. Set up computer and scanner facing the television set, put in the first movie; alternatively, get your friends to gather round the computer and start talking entertainingly, because this is going to be long and boring.

4. Scan each photo as a hi-res jpg; name each jpg as you save it with a long filename that gives the year the photo was taken and the names of the people in the photo and any other information you want to have.

5. Repeat step 4 until you run out of movies or your friends cease to be amusing.

6. Repeat steps 3 to 5 until you run out of photos.


Posted by: Jesurgislac | Link to this comment | 03-10-09 3:17 PM
horizontal rule
58

I'm also reminded of a quote from Robert Benchley I've had since the days of .plan files: "Anyone can do any amount of work provided it isn't the work he is supposed to be doing at the moment."

I've been trying to work out how to trick myself into thinking something unimportant is what I'm supposed to be doing, so that I can procrastinate by doing the important things. I can't make it work.


Posted by: essear | Link to this comment | 03-10-09 3:19 PM
horizontal rule
59

re: 51

Prints, manuscripts, etc., yeah, 600dpi and full colour.

Negatives you'd need to go much higher. I'm not actually sure if there is a 'standard' res for that. Here we tend to scan at whatever resolution we need for a particular job. If, say, someone wants to produce a facsimile from a set of 1930s glass plate negatives then we'll scan them at whatever size the printer needs to print properly at the appropriate paper size. In general, the negatives themselves are treated as archival surrogates [rather than being objects for conservation themselves]*; so the process isn't quite like digital photography of manuscripts. However, as a general rule, if the negative is a negative of a quarto sized original, we'd scan at whatever res we'd need to produce a quarto sized print.

* except for, say, nitrate negs, that aren't archivally safe.


Posted by: nattarGcM ttaM | Link to this comment | 03-10-09 3:20 PM
horizontal rule
60

re: 57

Standard practice would be NOT to save as jpegs. Lossless formats like TIFF are preferred.


Posted by: nattarGcM ttaM | Link to this comment | 03-10-09 3:21 PM
horizontal rule
61

58: let me know if you figure it out.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 3:25 PM
horizontal rule
62

18: No, but you can pay M. LeBlanc enterprises to do it for you.


Posted by: m. leblanc | Link to this comment | 03-10-09 3:30 PM
horizontal rule
63

No, but you can pay M. LeBlanc enterprises to do it for you.

Ah, heebie-geebie's secret plan to start an Unfogged bidding war to get her photos organized cheaply begins to become clear.


Posted by: essear | Link to this comment | 03-10-09 3:33 PM
horizontal rule
64

Projects like this are totally the kind of thing that I would never, ever do for myself but it seems like fun when it's somebody else's life. Just like organizing closets and solving personal problems.


Posted by: m. leblanc | Link to this comment | 03-10-09 3:35 PM
horizontal rule
65

||

The University of Chicago vs. Westboro Baptist Church: apparently, God hates figs.

|>


Posted by: essear | Link to this comment | 03-10-09 3:38 PM
horizontal rule
66

I have a similar problem and have made not-yet-pinned-down plans to spend a long afternoon with my friend F., some photo corners, a white-ink pen, and a lot of albums. F. is a good choice for this because she appears in a lot of these photos and also has a good set of photo albums herself that she and I often look at together. She won't actually do the work, but sitting down with her and laughing over the photos together will make it happen. I doubt everything in the boxes will wind up in an album, but it will be a vast improvement over what I have now.


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 4:57 PM
horizontal rule
67

64 gets it exactly right.


Posted by: Robust McManlyPants | Link to this comment | 03-10-09 4:57 PM
horizontal rule
68

OT Bleg: I would like to learn some statistics, partly so that I can better understand medical research and economics. At some point, I would like to take a proper course, but right now that's not financially possible.

Can anyone recommend some books that I could work off of on my own? I was looking at Statistics for Dummies Online--and the Beginner one looked too easy while the Intermediate one looked too tied to one specific software package (Minitab).

I might have access to a university library, but I don't have access to the math and science one. The regional network isn't bad,so if people know of books likely to be available in a regular public library, I'd be grateful.

Bonus vocab Question: How does one pronounce the statistical term 'chi'? Is it chee as in tai chi or a hard sound similar to the second to last letter in the Greek alphabet?


Posted by: Bostoniangirl | Link to this comment | 03-10-09 5:11 PM
horizontal rule
69

I think it is the second to last letter in the Greek alphabet, isn't it?

Here's MIT's OpenCourseWare Statistics class; it's got lectures and problem sets, although I haven't looked at what software/books it's assuming you have.


Posted by: LizardBreath | Link to this comment | 03-10-09 5:18 PM
horizontal rule
70

68: You know about MIT's Open CourseWare, right? This course looks like it might be what you want.


Posted by: Josh | Link to this comment | 03-10-09 5:18 PM
horizontal rule
71

I would say pwned, but we pointed at different classes.


Posted by: LizardBreath | Link to this comment | 03-10-09 5:19 PM
horizontal rule
72

Chi in statistics is pronounced like the American English pronunciation of the Greek letter, yes. [kai] (It becomes clear when you see χ^2 written out, as in a "chi-square test")


Posted by: Otto von Bisquick | Link to this comment | 03-10-09 5:22 PM
horizontal rule
73

The course I linked to has prereqs, so I'd say you win. OTOH, the textbook for my course is cheaper.


Posted by: Josh | Link to this comment | 03-10-09 5:23 PM
horizontal rule
74

Thanks 69 and 70. I'll check to see whether I can get the books through a library. I wish that they posted video in addition to notes.


Posted by: Bostoniangirl | Link to this comment | 03-10-09 5:35 PM
horizontal rule
75

Scan them and use them as captchas for activating thoughtful gift cards received by friends and family.

If you do not mind lossy preserves, shred and mash them with your stand mixer, then use as fiber in the jelly / jam / marmalade of your foodie choice.


Posted by: Econolicious | Link to this comment | 03-10-09 5:40 PM
horizontal rule
76

Search on iTunes -- I think at least some of the OpenCourse lectures are up there. And try ABEBooks for the textbooks; lots of them are surprisingly cheap.


Posted by: LizardBreath | Link to this comment | 03-10-09 5:41 PM
horizontal rule
77

One drawback to putting all one's photos into neatly organized albums is that a malicious twerp of a roommate can the more easily grab and destroy your personal history.


Posted by: Jackmormon | Link to this comment | 03-10-09 5:42 PM
horizontal rule
78

68: If you can wait until fall I'll be taking statistics (an online course) then and will be happy to dish the media. If you haven't found what you've wanted, ask again in August.


Posted by: foolishmortal | Link to this comment | 03-10-09 5:43 PM
horizontal rule
79

77: Oh, no. Oatmeal boy?


Posted by: LizardBreath | Link to this comment | 03-10-09 5:44 PM
horizontal rule
80

77: When you've got something, you've got something to lose.


Posted by: Cryptic ned | Link to this comment | 03-10-09 5:45 PM
horizontal rule
81

Is oatmeal boy the spanker?


Posted by: A White Bear | Link to this comment | 03-10-09 5:52 PM
horizontal rule
82

Oatmeal boy=spanker=thief.

On another, more humorous note, the guy looking for New York writers to donate their time and work to a startup conservative website is ... still looking!


Posted by: Jackmormon | Link to this comment | 03-10-09 5:55 PM
horizontal rule
83

68: Do you know linear algebra? Most of the intro statistics techniques are special cases of linear regression, a fact that is impossible to guess from reading intro statistics books. Most economics papers use line regression, and medical papers frequently use linear-regression-like techniques, anyway. So if you're comfortable with linear algebra, I would recommend something that starts with regression, such as Bruce Hansen's intro econometrics notes (available online).


Posted by: Walt Someguy | Link to this comment | 03-10-09 6:07 PM
horizontal rule
84

How is linear regression related to the kind of regression whn you curl up in foetal position? Is it more autistic, and less infantile?


Posted by: minneapolitan | Link to this comment | 03-10-09 6:13 PM
horizontal rule
85

If you do not mind lossy preserves

I love that very much.


Posted by: Wrongshore | Link to this comment | 03-10-09 6:20 PM
horizontal rule
86

84: You don't curl up -- just lie there with your body in an absolutely straight line. Other than that, pretty much the same.


Posted by: LizardBreath | Link to this comment | 03-10-09 6:24 PM
horizontal rule
87

68 et al: you might also enjoy Andrew Gelman's blog and books.

Maybe a little farther down the line?

83: damnable linear regression!

84: no. It's about the same, except you're hiding from your data, rather than the world.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 6:25 PM
horizontal rule
88

Also, learning statistics without some recent (or at least ready-at-hand) calculus is going to be very frustrating. I don't know why they ever try to do that.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 6:26 PM
horizontal rule
89

Dang, this reminds me I have a question to ask Cosma. Can somebody offer a grant he needs to write a proposal for, or something?


Posted by: Sifu Tweety | Link to this comment | 03-10-09 6:29 PM
horizontal rule
90

89 I'll whip up a BAA on something machine learning something not actually power laws.


Posted by: TJ | Link to this comment | 03-10-09 6:38 PM
horizontal rule
91

Yeah that's probably about right. All I WAS really going to ask was "does this data mining class look awesome... or what?!?"


Posted by: Sifu Tweety | Link to this comment | 03-10-09 7:21 PM
horizontal rule
92

Was, not WAS.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 7:23 PM
horizontal rule
93

Wait, no, was, not Was.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 7:23 PM
horizontal rule
94

Was (not Was)?


Posted by: LizardBreath | Link to this comment | 03-10-09 7:37 PM
horizontal rule
95

Isn't that the name of some terrible band?


Posted by: TJ | Link to this comment | 03-10-09 7:37 PM
horizontal rule
96

Learning basic statistics seems to means many things. For the intro class I've already complained about on other threads, it meant learning which already formed equations are suitable to certain situations under specified assumptions, learning which numbers taken from problem statements and a set of fixed tables should be plugged into those equations, and learning which equation should be chosen for given problems.

Anything that involved learning why those equations get used, or what to do under other conditions, or anything else that would have required more than algebra fell into the category of "beyond the scope of this book."


Posted by: eb | Link to this comment | 03-10-09 7:44 PM
horizontal rule
97

95: bite your tongue. They're an awesome band.

Nah, actually, they're pretty terrible.

96: I took a class like that. It was a big bowl of what's-the-point.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 7:58 PM
horizontal rule
98

It seems regrettable that there aren't more statistics courses that are explicitly geared toward teaching people how to read statistics. I realize that learning about how to do experiment design and statistical analysis isn't entirely irrelevant to learning how to understand statistics when one encounters other people's deployment of them, but if the latter is what your students ultimately need to learn (and in many cases it is), courses surely could drive more deliberately and directly toward that goal.


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 8:02 PM
horizontal rule
99

It seems regrettable that there aren't more statistics courses that are explicitly geared toward teaching people how to read statistics.

Amen.


Posted by: Witt | Link to this comment | 03-10-09 8:07 PM
horizontal rule
100

68

OT Bleg: I would like to learn some statistics, partly so that I can better understand medical research and economics. ...

You didn't say what math background and aptitude you have. This is pretty important information for a reasonable recommendation.


Posted by: James B. Shearer | Link to this comment | 03-10-09 8:09 PM
horizontal rule
101

98: I'm not exactly clear how you could disentangle them.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:10 PM
horizontal rule
102

98: I'm not exactly clear how you could disentangle them.

I'm not suggesting disentangling entirely, but adjusting the focus. Imagine, if you will, that you are teaching this class to a bunch of people pursuing journalism degrees.


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 8:13 PM
horizontal rule
103

102: aieee!


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:14 PM
horizontal rule
104

I'm not exactly clear how you could disentangle them.

I'm totally talking out of my hat here, because I don't know beans about statistics, but if I were signing up for such a class I'd want the instructor to spend a lot of time saying "You can tell the story this way or that way or that other way. People choose P method when they want to highlight ___, and Q method when they want to hide ___. The thing to watch out for when you are reading X is to ask how they calculated the Y."

A stupid baby-level example would be the disingenuous lying way Michael Moore displayed murder statistics in Farenheit 9/11. (He used absolute numbers of deaths rather than murder rate per 100,000/pop.)


Posted by: Witt | Link to this comment | 03-10-09 8:16 PM
horizontal rule
105

Well, so, that'd almost be a different (but very interesting!) course: dishonest statistics, and how to avoid them. I could see that.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:18 PM
horizontal rule
106

But like, if you talk about the kind of stuff Cosma complains about (e.g. people doing Bayesian-ish analyses with embedded assumptions that don't get washed away) or the kind of stuff that Vul paper on neuroscience addressed (e.g. implausibly high correlations combined with tenuous cross-checking of small data sets) it seems like it'd be pretty hard to unpack without actually understanding the outlines of the theory.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:19 PM
horizontal rule
107

Figures somebody'd take a crack at it, although who knows who that dude is.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:22 PM
horizontal rule
108

104: Right -- not so much a statistics class, as a 'critical reading of statistics' class. More breadth, less depth; no real focus on the math necessary to actually do the stats yourself, instead focusing on different techniques and in what contexts they are and aren't going to produce useful evidence.

You know, while I see what you want, and it's a reasonable thing to want, I don't think there's any plausible way to do it past the baby level; if you can't do the math yourself, you're not going to be able to keep the pitfalls and possible misleading ways to present things straight. (On the other hand, the baby level's useful. And I say this as someone with only a baby level grasp of stats myself.)


Posted by: LizardBreath | Link to this comment | 03-10-09 8:22 PM
horizontal rule
109

105: This is where my ignorance enters in. From my perspective, there is no such thing as honest or dishonest statistics, just the tool and you can either use it to highlight or obscure things as your politics and ethics dictate.

The interesting question from my POV is how one goes about it. Because you have to use neutral language, but also you have to give at least some illustration of motivations, or your students will just be confounded (or misled!).

Lately I've been spending a lot of time making layperson's financial explanations to a frighteningly well educated finance whiz (he's not especially well versed in my field), and it's quite a challenge because it's so different from the usual kinds of explaining I do.


Posted by: Witt | Link to this comment | 03-10-09 8:23 PM
horizontal rule
110

102: Imagine, if you will, that you are teaching this class to a bunch of people pursuing journalism degrees.

103: 102: aieee!

Heh, imagine, if you will, that you are teaching a philosophy class to a bunch of people pursuing biology or physics or economics degrees. You struggle for a while, but you eventually decide it's worth figuring out how to do this.


Posted by: parsimon | Link to this comment | 03-10-09 8:25 PM
horizontal rule
111

You could do case studies for journalists: Lancet study, how to read beyond the press release page of polls, etc. There's always a risk that some people would just get better at cherry-picking, but that's not your fault.


Posted by: eb | Link to this comment | 03-10-09 8:27 PM
horizontal rule
112

You know, while I see what you want, and it's a reasonable thing to want, I don't think there's any plausible way to do it past the baby level

This, I feel like is the case. (Not that I'm some expert, here. I've taken a couple classes.) I also think it ties into this:

From my perspective, there is no such thing as honest or dishonest statistics, just the tool and you can either use it to highlight or obscure things as your politics and ethics dictate.

You can definitely use statistics dishonestly, and you there are definitely some statistical tools which are kind of red flags that say "this person is probably full of shit", but you're right that the bigger problem a lot of times is people deluding themselves, or things that are even harder to see, like trying a bunch of models until one fits right, that'll increase your likelihood of thinking there's a valid inference to be made.

I think you might have to know the math. But I dunno, maybe somebody who really knows their shit can correct me?


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:28 PM
horizontal rule
113

The median/mean distinction is widely abused.


Posted by: eb | Link to this comment | 03-10-09 8:29 PM
horizontal rule
114

110: surely also a frustrating thing, and there are probably some parallels, but not really the same as trying to teach somebody math without them learning the math.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:31 PM
horizontal rule
115

Obligatory.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:31 PM
horizontal rule
116

I think you might have to know the math.

I think there's a continuum. And I'm more convinced that there's room between the baby stuff and some medium-sized stuff than you and LB are. It sounds like you guys think it's binary. Either you don't have the math [and I don't] and so you can only do the babiest of baby stuff, or you do, and therefore you can do "real" stats.

It would be very nice if some real live statistician could show up here and explain which one it is.

From the outside it's always so hard to gauge a field -- whether a reasonably clear thinker can suss out the gist of what's going on, or whether that's a false certainty and you can do a lot of damage assuming a bit of research has informed you. I think that's why I can't read Yglesias regularly. He writes like someone whose experience has been that 92% of the time he can confidently reason his way to clarity, which makes the last 8% so resoundingly disastrous.


Posted by: Witt | Link to this comment | 03-10-09 8:36 PM
horizontal rule
117

I don't think there's any plausible way to do it past the baby level

I agree -- this would be a baby-level course. It's not as if intro stats courses for undergrads in, for example, psychology, turned out students with more than, or indeed even, a baby level understanding of the stats they've been taught.

I'm also actually not proposing necessarily not doing the math -- though for the journalists, maybe you would do it that way -- but if you do make the students work though the math in exercises that follow and flesh out case-study exercises in reading comprehension, I think many would remember and understand it better.


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 8:38 PM
horizontal rule
118

A stupid baby-level example would be the disingenuous lying way Michael Moore displayed murder statistics in Farenheit 9/11. (He used absolute numbers of deaths rather than murder rate per 100,000/pop.)

That doesn't strike me as misusing statistics so much as just basic innumeracy -- it's like saying "we gave a million dollars to the homeless! now they're rich!"


Posted by: essear | Link to this comment | 03-10-09 8:39 PM
horizontal rule
119

116: I don't think it's binary. I think either you don't have the math, and so you can only do the babiest of baby steps, or you have some calculus and maybe some linear algebra, in which case you can basically grasp what's going on, or you actually know what you're doing. There's definitely a continuum, but the part of the continuum where you can get what's going on without much mathematical background is waaaaay over on one end.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:39 PM
horizontal rule
120

Shorter 119: There's a continuum, except it looks very binary!

(Don't mind me, I'm punchy after having seen that ridiculously over-the-top old Indiana Jones story conference excerpted all over the place. Also, I had a long day with an unbelievably extraordinary set of stories and I'm kind of blitzed.)


Posted by: Witt | Link to this comment | 03-10-09 8:42 PM
horizontal rule
121

Some of it isn't so much the statistics as the assumptions behind the numbers that can be misleading. A lot of city comparisons aren't careful enough about picking which units to compare. So you get New York, reduced to Manhattan, compared to a metropolitan area including suburbs, or the Bay Area is treated as just San Francisco, or whatever. Figuring out which units to choose isn't always easy depending on what you want to measure - so it's not all deliberate gaming of the numbers - but it shouldn't be too hard to teach people at least to be aware of what's going on.


Posted by: eb | Link to this comment | 03-10-09 8:43 PM
horizontal rule
122

My sense is pretty much with Sifu, except that basic numeracy -- the baby-steps stuff -- will get you pretty far in dealing with the sort of stats you see in the news.


Posted by: LizardBreath | Link to this comment | 03-10-09 8:43 PM
horizontal rule
123

Basically I am thinking that it would be great if the undergrads who only learned the first little bit of what they were taught had some practice learning how to decipher one of the papers their other classes expect them to read. And it would have been good for me as an undergrad to have taken such a class.


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 8:45 PM
horizontal rule
124

(A related class for aspiring journos, which would be even baby-stepsier, would be good too. Journalism school offers no such thing.)


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 8:47 PM
horizontal rule
125

misusing statistics so much as just basic innumeracy

Come to think of it, this whole discussion seems to be divided between statistics talked about as numeracy and statistics talked about as a mathematical discipline.


Posted by: eb | Link to this comment | 03-10-09 8:49 PM
horizontal rule
126

A lot of city comparisons aren't careful enough about picking which units to compare.

Ha, this is so true. A while back I read this comparison of Philadelphia versus Detroit and Miami. You couldn't pick a better city than Detroit if you wanted to make Philadelphia look good.


Posted by: Witt | Link to this comment | 03-10-09 8:49 PM
horizontal rule
127

I just want undergrads to have a hope of doing something other than skimming over the "results" sections of articles.


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 8:53 PM
horizontal rule
128

I remember last summer reading a few articles comparing Houston with other cities (always favorable to Houston) that all appeared withing a short period of time and wondering if there was an active PR campaign behind it, perhaps on the part of Houston-based people, perhaps related to then-high gas prices. But that would never happen, so I dismissed the possibility.


Posted by: eb | Link to this comment | 03-10-09 8:55 PM
horizontal rule
129

127: whether or not they have any desire to being a different question.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 8:57 PM
horizontal rule
130

I just want Congress and the public undergrads to have a hope of doing something other than crossing their fingers that financial fancypants aren't selling them a bill of goods called Giving Hank $700 Billion Will Save the Free World skimming over the "results" sections of articles.


Posted by: Witt | Link to this comment | 03-10-09 8:58 PM
horizontal rule
131

wondering if there was an active PR campaign behind it

Obligatory.


Posted by: Witt | Link to this comment | 03-10-09 9:00 PM
horizontal rule
132

130: no chance.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 9:01 PM
horizontal rule
133

131: That piece is good, but all those comments at the end of how online writing is so much more honest make me wonder...


Posted by: eb | Link to this comment | 03-10-09 9:21 PM
horizontal rule
134

127: whether or not they have any desire to being a different question.

Yes! Yes, it is.


Posted by: redfoxtailshrub | Link to this comment | 03-10-09 9:48 PM
horizontal rule
135

116.3: This is well-put. Also possibly why I can't read Yglesias regularly.


Posted by: parsimon | Link to this comment | 03-10-09 10:15 PM
horizontal rule
136

I can't help but think that I've had a course that did a nice job of teaching me how to evaluate statistical claims. It was simply called Critical Thinking with a textbook of the same name, circa 91-92. It also built in quite a lot about different argument structures, common fallacies, etc. THis was later supplemented by a couple formal logic cources (the kind they do in the Philosophy Dept., none of this weird-ass CS stuff--although the one programming course I took probably helped as well).

Anyway, the experience made me convinced that critical thinking skills need to be taught from the age of 3 on. Obviously, Daughter #1 spends a lot of time annoyed with Dad's Socratic teaching of concepts she's not really ready for, but her checkers and Uno skills are coming right along--the little shit actually beat me straight up at checkers the other night (with much coaching on options from me, but consistently making the right choices and thinking things through). It'd be really fucking nice if even adults could identify post hoc ergo propter hoc arguments, much less kids, but I think the way you get there is starting from an early age, while the brain is still plastic enough to adapt to a new way of thinking.


Posted by: Chopper | Link to this comment | 03-10-09 10:17 PM
horizontal rule
137

I think that's why I can't read Yglesias regularly. He writes like someone whose experience has been that 92% of the time he can confidently reason his way to clarity, which makes the last 8% so resoundingly disastrous.

Holy moly you've got some high standards.

That dude is impressively prolific and interesting on a wide range of topics. Also, he's got a snarky streak that warms my heart.


Posted by: gswift | Link to this comment | 03-10-09 10:28 PM
horizontal rule
138

Yes yes yes to 136! Although if you're not started at an early age, you can still get there if you have teachers who can provide the tutelage, without necessarily requiring you to have x many years of math or philosophy or whatever training in advance. It's difficult to provide that kind of teaching, though, since the academy doesn't really encourage it.


Posted by: parsimon | Link to this comment | 03-10-09 10:31 PM
horizontal rule
139

I think either you don't have the math, and so you can only do the babiest of baby steps, or you have some calculus and maybe some linear algebra, in which case you can basically grasp what's going on, or you actually know what you're doing.

The average undergrad stats course designed to be accessible to people in the social sciences would not require calculus or linear algebra.

A course like this is still taught out of the math/stats department, but I think would be much better if it emphasized reading comprehension and "what sorts of questions should cross your mind when you read this paragraph?" type critical thinking.


Posted by: heebie-geebie | Link to this comment | 03-10-09 10:44 PM
horizontal rule
140

116: I too think there is some potential for a better way to structure courses for potential "readers" of stats (or quantitative reasoning in general) that can vary in level of mathematical sophistication (I could see calc and non-calc versions for instance).

In my experience intro stats (and actually almost all intro xyz) courses are taught as if they were the start of the progression leading to practitioner, and of course this is very understandable and appropriate for those on that path. However, for the many who are not, they get "this is how you would begin to do stats but of necessity we are going to stop so woefully short of the point where you could really do it yourself that what you learned is next to worthless". Better to equip these students with an understanding of the key concepts, assumptions and methods (and, yes, actually have them do the math at the level accessible to them to get the understanding, so it is not just "poets statistics", but rather a subtle shift of emphasis). What I would expect out of it is that someone would understand the key foundational ideas—distributions and tests and the assumptions behind them—well enough to be able to understand why a particular test was used based on the characteristics of the data and question being asked, or to engage a true practitioner in a constructive dialogue. You can get to some relatively sophisticated quantitative reasoning without the calc even (and, yes, that does beg the question of why not just learn the calc, and some of the answer to that rests with a similar observation on how intro calc is taught).

And somewhat pwned by Chopper. I do think a good grounding in Critical Thinking is key. A course at the high school level that I have toyed with designing would use assessing claims in ads and PR as the core.


Posted by: JP Stormcrow | Link to this comment | 03-10-09 10:47 PM
horizontal rule
141

The average undergrad stats course designed to be accessible to people in the social sciences would not require calculus or linear algebra.

And yet, if you ask the average undergrad what they got out of their average undergrad stats course, they will say "absolutely nothing". I'm not saying it's a bad idea, but I saw a class full of students blithely not their uncomprehending way through ten weeks of statistics taught exactly as described in 96. I imagine they got fuck-all out of it.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 10:48 PM
horizontal rule
142

a class full of students

Come now. That's not a valid sample size.


Posted by: Stanley | Link to this comment | 03-10-09 10:50 PM
horizontal rule
143

138: since the academy doesn't really encourage it.

Yes. My overly cynical assessment of some discussions at my high-school on the potential for courses like this; it won't help get the little suck-ups get into <insert your aspirational level of college here>. (And the best teachers and profs do incorporate it into their teaching no matter what the subject.)


Posted by: JP Stormcrow | Link to this comment | 03-10-09 10:53 PM
horizontal rule
144

142: the central limit theorem says you're wrong!


Posted by: Sifu Tweety | Link to this comment | 03-10-09 10:54 PM
horizontal rule
145

But a class full of Students would be another story.


Posted by: JP Stormcrow | Link to this comment | 03-10-09 10:56 PM
horizontal rule
146

144: Abbie Normal says he's right!


Posted by: JP Stormcrow | Link to this comment | 03-10-09 10:57 PM
horizontal rule
147

||
Hey Becks or w-lfs-n, if you're reading... I figured out what was causing those 403 errors I was seeing.
|>


Posted by: Josh | Link to this comment | 03-10-09 11:00 PM
horizontal rule
148

No javascript?


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:23 PM
horizontal rule
149

It's not that (or at least it's not quite as simple as that sounds). If you preview a comment, then attempt to submit it before the page has fully loaded, you get a 403 error. What tripped me up is that the browser's throbber won't go while the preview page is loading (as least in Firefox on OS X).


Posted by: Josh | Link to this comment | 03-10-09 11:29 PM
horizontal rule
150

Ah, yes. Same basic problem.


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:33 PM
horizontal rule
151

Way to put the comment submitting script in the footer even after preview has been pressed, guys. G'hyuh-hyuh-hyuh.


Posted by: Sifu Tweety | Link to this comment | 03-10-09 11:33 PM
horizontal rule
152

140: A course at the high school level that I have toyed with designing would use assessing claims in ads and PR as the core.

JP,* I'm surprised that you're talking about the high-school level here. I've been talking about college-age. In my experience, even the ivy league wannabes I've taught at had a lot of students who had a weak handle on critical thinking skills; we're talking ability to identify logical fallacies and so on. The students were far too focused on plugging the numbers into the right places, and were clueless if someone didn't tell them where to do that. More power to you, obviously, if you can head this off at the pass.

* I call you JP in order to be familiar, even though I think "Stormcrow" is the coolest name ever.


Posted by: parsimon | Link to this comment | 03-10-09 11:34 PM
horizontal rule
153

152: So true. I spend a lot of time teaching students how to build an argument in hopes that they might also learn how to deconstruct an argument. Well, I do both, but the lasting skill I want them to have is how to evaluate an argument so that maybe they have a chance to cut through all the bullshit they're fed (mine included).


Posted by: Parenthetical | Link to this comment | 03-10-09 11:37 PM
horizontal rule
154

That would be in the <foot> part of the document, you mean, right Sifu?


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:37 PM
horizontal rule
155

That would be "analyze" an argument, you mean, right Parenthetical?


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:38 PM
horizontal rule
156

Why's everybody gotta give me all this grief about being a dick today?


Posted by: Sifu Tweety | Link to this comment | 03-10-09 11:38 PM
horizontal rule
157

Yeah, yeah. I don't teach them grammar. (Thank god).


Posted by: Parenthetical | Link to this comment | 03-10-09 11:43 PM
horizontal rule
158

152: Well, for the high school, as I indicated, it would be a fairly simple thing, "Give me 4 reasons why the the claims in this advertisement might not be true", and basic numeracy per essear in 118, "Why $1 million dollars is not a lot in the context of the whole of the United States."

A former neighbor was an epidemiologist at Pitt. He mentioned that his first Critical Thinking course was one he took (a requirement?) his 1st year as a grad student at Stanford. He thought it was a great course, but said he was struck by how after the fact it seemed, and really wished he had had it earlier (he was an MIT undergrad).


Posted by: JP Stormcrow | Link to this comment | 03-10-09 11:46 PM
horizontal rule
159

155: Cough, choke.


Posted by: parsimon | Link to this comment | 03-10-09 11:49 PM
horizontal rule
160

Ok, I'm going to reveal genuine ignorance here. Does evaluate really not mean "to assess the value" of something? Does it only work for numbers?

I'd appreciate enlightenment and will accept gentle mockery along the way.


Posted by: Parenthetical | Link to this comment | 03-10-09 11:51 PM
horizontal rule
161

"analyze" insted of "deconstruct".


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:52 PM
horizontal rule
162

"evaluate" only works for code and symbols.


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:53 PM
horizontal rule
163

152: I call you JP in order to be familiar, even though I think "Stormcrow" is the coolest name ever.

Well, shucks...
It was basically one of two online gaming names I used and seized on it in the spur of the moment for my first blog comment (maybe at DailyKos). I sometimes wish I had used the other one, StableMagma, given the richer vein of double meanings it carries.


Posted by: JP Stormcrow | Link to this comment | 03-10-09 11:54 PM
horizontal rule
164

Oh. So what's wrong with deconstruct? Beyond the fact that it reveals I have been exposed to way too much academicese?

Are we just being nitpicky?


Posted by: Parenthetical | Link to this comment | 03-10-09 11:54 PM
horizontal rule
165

"deconstruct" has a (and arguably only a) technical meaning and you probably aren't teaching them to do that.


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:56 PM
horizontal rule
166

And of course we're being nitpicky. It's what we do.


Posted by: ben w-lfs-n | Link to this comment | 03-10-09 11:56 PM
horizontal rule
167

Are you *flirting*, w-lfs-n?


Posted by: Josh | Link to this comment | 03-10-09 11:57 PM
horizontal rule
168

166.2: Some of us.


Posted by: JP Stormcrow | Link to this comment | 03-10-09 11:59 PM
horizontal rule
169

I didn't realize that deconstruct only had a technical meaning.

And in class I use the word analyze, thank you very much. Well, that and "take apart." Generally, when speaking aloud, I remember to stick to the words I know the meaning of.


Posted by: Parenthetical | Link to this comment | 03-10-09 11:59 PM
horizontal rule
170

Some of us merely provide the fodder to be picked.


Posted by: JP Stormcrow | Link to this comment | 03-11-09 12:00 AM
horizontal rule
171

Just be careful not to call your course "The Secrets of Analysis Revealed."


Posted by: eb | Link to this comment | 03-11-09 12:00 AM
horizontal rule
172

167: If it's flirting, it's like the version where you pull the pigtail of the girl in front of you.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:00 AM
horizontal rule
173

172: Well yeah. This *is* w-lfs-n we're talking about.


Posted by: Josh | Link to this comment | 03-11-09 12:01 AM
horizontal rule
174

170: I think I'm in that boat.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:01 AM
horizontal rule
175

Baby steps, people.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:02 AM
horizontal rule
176

Actually, the OED kindly lists a nontechnical meaning ("take apart") for deconstruct, but the only example sentences it gives are for the technical meaning and they're all (as you'd expect) of recent vintage.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:02 AM
horizontal rule
177

ben, you can correct my writing all you want. When I have a draft, in fact, can I send it to you?


Posted by: Parenthetical | Link to this comment | 03-11-09 12:03 AM
horizontal rule
178

It's my understanding that "deconstruct" is not a transitive verb, but perhaps the meaning I learned so long ago is lost to time.


Posted by: Jesus McQueen | Link to this comment | 03-11-09 12:05 AM
horizontal rule
179

This relationship has turned exploitative faster than I thought.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:06 AM
horizontal rule
180

177: Second base! Woo hoo!!


Posted by: JP Stormcrow | Link to this comment | 03-11-09 12:06 AM
horizontal rule
181

How could "deconstruct" not be a transitive verb?


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:06 AM
horizontal rule
182

I thought you'd prefer nit picking to sex. My apologies if I was wrong.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:07 AM
horizontal rule
183

177: Second base! Woo hoo!!

I have always wanted to make love to a woman with punctuational nomenclature.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:08 AM
horizontal rule
184

182 to 179, of course.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:08 AM
horizontal rule
185

One doesn't deconstruct a text, it's a de-structuring inherent in the text itself.


Posted by: Jesus McQueen | Link to this comment | 03-11-09 12:09 AM
horizontal rule
186

How many deconstructionists does it take to change a lightbulb?


Posted by: Wrongshore | Link to this comment | 03-11-09 12:10 AM
horizontal rule
187

I just deconstructed a glass of wine all over my desk.


Posted by: Sifu Tweety | Link to this comment | 03-11-09 12:10 AM
horizontal rule
188

165: "deconstruct" has a (and arguably only a) technical meaning and you probably aren't teaching them to do that.

"Deconstruction" is a technical term for a specific method of analysis, a Derridean one, and no, you are quite probably not teaching them that, since it's a bit advanced and it doesn't sound like it's your field in the first place.

You meant that you wanted your students to analyze arguments.

I seem to be using ben's terms here, but that's because he's right. "Deconstruct" has been coopted, sadly, but it would be great if people stopped doing that.


Posted by: parsimon | Link to this comment | 03-11-09 12:11 AM
horizontal rule
189

Okay, Wrongshore, how many?


Posted by: Jesus McQueen | Link to this comment | 03-11-09 12:12 AM
horizontal rule
190

Deconstruct is a transgressive verb. Occupying the liminal space between transitive and intransitive verbs, it interrogates the problematics inherent in grammatical forms.


Posted by: eb | Link to this comment | 03-11-09 12:12 AM
horizontal rule
191

But deconstruction was the followup to the Destruktion der Metaphysik of Heidegger, which is most certainly something that has to be carried out. Quod erat bewiesen.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:12 AM
horizontal rule
192

I was cleaning out old junk

It's beautiful that you and Jammies are still getting it on despite the pregnancy related discomfort, but don't you think this is a slightly insensitive description?


Posted by: TJ | Link to this comment | 03-11-09 12:13 AM
horizontal rule
193

Okay, Wrongshore, how many?

For too long philosophers have merely described the lightbulb. The point, however, is to change it.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:13 AM
horizontal rule
194

Interpreted the lightbulb variously, that is.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:15 AM
horizontal rule
195

I guess in German you would wechseln lightbulbs, not verändern them?


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:16 AM
horizontal rule
196

188: Yeah, I know what it is, I've been exposed to the theory at great length. I'm not teaching it, no, because it's not my field. Unfortunately I have a brain like a raccoon that grasps at any shiny words and incorporates them into my vocabulary to be used willy-nilly without my rational mind's consent.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:16 AM
horizontal rule
197

I wish all of unfogged had been there when I had Blume explain deconstruction to me. You know, so you could all make fun of me.


Posted by: Sifu Tweety | Link to this comment | 03-11-09 12:17 AM
horizontal rule
198

Unfortunately I have a brain like a raccoon that grasps at any shiny words and incorporates them into my vocabulary to be used willy-nilly without my rational mind's consent.

All that and a fondness for mixed metaphors.


Posted by: Sifu Tweety | Link to this comment | 03-11-09 12:18 AM
horizontal rule
199

198: You know, I'm like the antithesis of ben.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:19 AM
horizontal rule
200

191: Not so fast, philosophy boy. While you've been reading that smartypants tomfoolery, I've had little more than Blueberries for Sal and Horton Hears a Who. I need a little Derrida refresher before I pursue this any further.


Posted by: Jesus McQueen | Link to this comment | 03-11-09 12:20 AM
horizontal rule
201

Four: one to change the lightbulb and another to point out the subordination of dark to light, off to on, and burnt out to functioning.


Posted by: Wrongshore | Link to this comment | 03-11-09 12:21 AM
horizontal rule
202

Remember! The thesis-antithesis-synthesis triad is not Hegelian vocabulary!


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:21 AM
horizontal rule
203

197: Isn't it formally like a Fourier transform only very different?


Posted by: JP Stormcrow | Link to this comment | 03-11-09 12:21 AM
horizontal rule
204

Four: one to change the lightbulb and another to point out the subordination of dark to light, off to on, and burnt out to functioning.

So what do the other two do?


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:22 AM
horizontal rule
205

200: when is Derrida not refreshing?


Posted by: Sifu Tweety | Link to this comment | 03-11-09 12:22 AM
horizontal rule
206

196: Unfortunately I have a brain like a raccoon that grasps at any shiny words and incorporates them into my vocabulary to be used willy-nilly without my rational mind's consent.

Okay, well, the Mineshaft is going to look at you askance if you do that.


Posted by: parsimon | Link to this comment | 03-11-09 12:22 AM
horizontal rule
207

Deconstruction is more specifically like finding Chladni patterns.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:23 AM
horizontal rule
208

205: When he is reading a long comment.


Posted by: JP Stormcrow | Link to this comment | 03-11-09 12:23 AM
horizontal rule
209

Parenthetical means she has a brain like a magpie.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:23 AM
horizontal rule
210

Deconstruction would need Borel resummation or something.


Posted by: TJ | Link to this comment | 03-11-09 12:24 AM
horizontal rule
211

206: Understood. Note, I wasn't actually complaining about ben's correction.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:24 AM
horizontal rule
212

204: I just put them in to fool people.


Posted by: Wrongshore | Link to this comment | 03-11-09 12:25 AM
horizontal rule
213

209: Raccoons like shiny things! They collect them too!


Posted by: Parenthetical | Link to this comment | 03-11-09 12:25 AM
horizontal rule
214

But deconstruction was the followup to the Destruktion der Metaphysik

I was about to point that out in one of those Liberal Fascism threads somewhere in response to the idiot author's crowing that he'd found the word dekonstruktion in some incriminating context, but thought better of it.


Posted by: Jesus McQueen | Link to this comment | 03-11-09 12:26 AM
horizontal rule
215

212: I thought they were representatives of the four people thus described.


Posted by: eb | Link to this comment | 03-11-09 12:27 AM
horizontal rule
216

The prevalence of colloquial "deconstruct" is actually something I've gotten more or less used to. (Not like "the reason is because", my sworn foe to the end of days.)


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 12:28 AM
horizontal rule
217

216: I have much smaller fish to fry. I'm still working on convincing students that novel is not a synonym for book.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:29 AM
horizontal rule
218

novel is not a synonym for book

Of course not. It's a synonym for "run!" Book is too slangy.


Posted by: Sifu Tweety | Link to this comment | 03-11-09 12:33 AM
horizontal rule
219

216: Yeah, I guess. It still annoys me when it comes from people in literary fields. I'd like them to remember that they don't know what they're talking about.


Posted by: parsimon | Link to this comment | 03-11-09 12:39 AM
horizontal rule
220

||

Hey, the Idiocracy novelization's online!

|>


Posted by: Sifu Tweety | Link to this comment | 03-11-09 12:40 AM
horizontal rule
221

219: yeah, because what do literature types know about Derrida?


Posted by: Sifu Tweety | Link to this comment | 03-11-09 12:41 AM
horizontal rule
222

They're always already wrong, Sifu. You can point out the error of their ways, the ways of their error.


Posted by: Jesus McQueen | Link to this comment | 03-11-09 12:43 AM
horizontal rule
223

220: But where are the electrolytes?


Posted by: JP Stormcrow | Link to this comment | 03-11-09 12:47 AM
horizontal rule
224

218: Dear god, it took me far too long to get that.


Posted by: Parenthetical | Link to this comment | 03-11-09 12:47 AM
horizontal rule
225

||
224: Here is one I puzzled on for a while earlier tonight,

Instructions on setting up mouse traps (classic variety):
Engage Locking Bar (S=Sensitive/F=Firm)

All I could think of was, "You can set it to kill them 'sensitively'?"
|>


Posted by: JP Stormcrow | Link to this comment | 03-11-09 12:53 AM
horizontal rule
226

221: 219: yeah, because what do literature types know about Derrida?

If literature types use the term deconstruction to mean analysis, when they've read and studied Derrida et al., then I imagine I have to yield.


Posted by: parsimon | Link to this comment | 03-11-09 1:02 AM
horizontal rule
227

225: WHEREOF WE CANNOT SQUEAK, WE MICE BE SILENT.


Posted by: OPINIONATED *SNAP*!! | Link to this comment | 03-11-09 1:09 AM
horizontal rule
228

Had I been drinking wine at the time, 187 would have caused me to spit it out all over my desk.


Posted by: Sir Kraab | Link to this comment | 03-11-09 6:19 AM
horizontal rule
229

A course at the high school level that I have toyed with designing would use assessing claims in ads and PR as the core.

You might consider using How To Lie With Statistics as the centrepiece.


Posted by: Ginger Yellow | Link to this comment | 03-11-09 6:46 AM
horizontal rule
230

Standard 'archival' practice at the moment is to make 600dpi scans at 100%.

This is very good to know. AB has, historically, been a conscientious photo and album person, but there's a few shoeboxes that have been forlorn in the film-to-digital transition. But a mass scanning will surely be in her future, and it'll be good to know how to scan them.

Should do the same with the scattered, random, and unrepresentative* photodocumentation of my side of the family.

* Best pic from my childhood? Circa age 3, I fell asleep on the kitchen floor next to the trash can. The others still remember the rush to get the camera and snap it before I awoke.


Posted by: JRoth | Link to this comment | 03-11-09 8:10 AM
horizontal rule
231

225: I must be using the cheap traps; I see no sensitivity setting on them.


Posted by: JRoth | Link to this comment | 03-11-09 8:13 AM
horizontal rule
232

Umm. ... My math background. I took AP Calculus AB in highschool. (My school was dumb; they made us take that instead of BC, because they wanted to spend an entire month reviewing to help us get higher scores. So, no Taylor series for me.) I didn't take multivariable calculusin college, because I was very focused on my concentration AND mainly because I was afraid that I wouldn't understand the Russian TFs.

I don't know any linear algebra. Should I learn that first?


Posted by: Bostoniangirl | Link to this comment | 03-11-09 8:20 AM
horizontal rule
233

Back on the statistics discussion: Speaking as one data point, I was always very good at math, got a 5 on the B/C AP exam for calc, but cannot grok statistics. I never understood permutations, even on the level taught in HS. I don't know what the block is, but it's nothing to do with basic numeracy or exposure. Based on the distinctive skills shown by card-counters and the like, I wonder is statistics/probability is actually a different bit of the brain from other pieces of math.

I do have enough numeracy to recognize the common mistakes made in the popular press, but that's beside the point. Given that Joe Six-or-so-pack is lucky to get unit pricing, it's hardly surprising that he easily fucks up issues related to rates and large numbers. I'm not sure we can assume that he'd be able to get it if only he had a course once.

Apologies if this has been covered - I'm just skimming through a bit (as 230 would indicate).


Posted by: JRoth | Link to this comment | 03-11-09 8:21 AM
horizontal rule
234

I don't know any linear algebra. Should I learn that first?

Sure, it's easy. You've got line A, length X. Another line B, length Y.

If X=5 and Y=3, which line is longer?

No looking it up on Wikipedia.


Posted by: JRoth | Link to this comment | 03-11-09 8:22 AM
horizontal rule
235

231: This is the first time I've seen that setting myself. My wife and I have compromised on having the no see/no touch traps in the "public" places you encounter first thing in the morning. But "the ratio of people dollars to cake dead mice is too big" for my taste, so in the out-of-the way places I go medieval. I can't kill the deer, so I take it out on the mice.


Posted by: JP Stormcrow | Link to this comment | 03-11-09 8:24 AM
horizontal rule
236

I have always wanted to make love to a woman with punctuational nomenclature.

There's that song from the guy who was in Georgia Satellites.

FWIW, "deconstruct" is now a term of art for demolition that involves taking apart a building to salvage the usable bits, as opposed to just smashing it all up and burying the debris. There are fewer puns about this than you might expect.


Posted by: JRoth | Link to this comment | 03-11-09 8:25 AM
horizontal rule
237

Deconstruction does have a technical meaning, but I don't know what it has to do with this discussion.

(Did no one ever use the word to mean "take apart" before Derrida? That seems hard to believe.)


Posted by: essear | Link to this comment | 03-11-09 8:25 AM
horizontal rule
238

235 makes me feel better about living in the city; I always figured you suburbanites at least were vermin*-free.

Fuck! 10:30.

* For values of vermin not including white-tailed deer.


Posted by: JRoth | Link to this comment | 03-11-09 8:30 AM
horizontal rule
239

re: 230

I think those numbers are based on the sorts of scan size you need to reliably print a facsimile, although in practice you can get away with lower res scans if you aren't too fussy.

Anyway, storage is cheap, and scanning prints at 600dpi is fast anyway.

Scanning negs is MUCH slower [and if you had a lot too do, it'd start to get cost effective to outsource it].


Posted by: nattarGcM ttaM | Link to this comment | 03-11-09 8:30 AM
horizontal rule
240

237: We construct renormalizable, asymptotically free, four dimensional gauge theories that dynamically generate a fifth dimension.

Laydeez.


Posted by: JRoth | Link to this comment | 03-11-09 8:31 AM
horizontal rule
241

237: Me too. But that's because I'm still smarting from ben's nit picking.

(Though, really, as parsimon points out, I should know better. I did read Derrida et al. And then promptly put them out of my mind, because, well. Obvious, I'm sure).


Posted by: Parenthetical | Link to this comment | 03-11-09 8:45 AM
horizontal rule
242

Anyway, storage is cheap, and scanning prints at 600dpi is fast anyway.

Exactly. I would have guessed/feared that the recommended dpi was significantly higher.


Posted by: JRoth | Link to this comment | 03-11-09 8:54 AM
horizontal rule
243

229: We used How To Lie with Statistics in a unit in my high school algebra class. It definitely a good place to put it.


Posted by: Nathan Williams | Link to this comment | 03-11-09 8:55 AM
horizontal rule
244

238: I always figured you suburbanites at least were vermin*-free.

Many probably are. However, my semi-remuddled circa 1900 "farm" house* (of which I am but the latest in a long line of undercapitalized owners) is more "porous" than most and provides a nook and cranny-rich environment for critters various.

*Twice the problems with half the charm! But a big yard, and schools tailored for the make/molding of offspring into fully realized consumero-normative citizens.

</Free-floating grumpy Wednesday morning cynicism>


Posted by: JP Stormcrow | Link to this comment | 03-11-09 8:55 AM
horizontal rule
245

The reason that I'm asking about this is that I'm in the 2nd round of interviews for a second part-time job on a research project. My role would not be research oriented. I'd be one of the educator/support/navigator types whose interventions they'd be studying, but I was thinking that I would like to suck up as much as possible, i.e. express my interest in the research, so that I might get good recommendations later. I am thinking that I might at some point want to do some sort of law and health career by getting an MPH--or even a doctorate if that meant that I could get funding.

Anyway, it would be a long haul,but I thought that I would test the waters by trying to learn some statistics.


Posted by: Bostoniangirl | Link to this comment | 03-11-09 8:59 AM
horizontal rule
246

I'm trying to evaluate comment 162, but I'm coming up empty.


Posted by: essear | Link to this comment | 03-11-09 9:00 AM
horizontal rule
247

BG, if you're in it for the long haul, learning something about fundamental properties of the most common statistical distributions might be worthwhile. You wouldn't be interested in the rest of the book, but I found the explanations in Reif's Fundamentals of Statistical and Thermal Physics clearer than anything I had read in statistics books. I can't find a good start-from-scratch discussion of the Poisson distribution offhand, but that might be worthwhile to look for. Separately, for basic Bayesian methods, this page is IMO a good starting point.


Posted by: lw | Link to this comment | 03-11-09 9:23 AM
horizontal rule
248

re: 242

Nah. But it is important that it's 600dpi @ 100%. So, for a 6 inch by 4inch photo you are scanning at 3600 x 2400 pixels, 24 bit colour. That's easily enough to print an identical copy.

Most of the time, anything printed at more than about 300dpi is good enough for the human eye at normal viewing distances. So scanning at twice that res gives a bit of leeway for resizing/sharpening, etc.

In practice, we usually shoot at more than 600dpi for smaller items [we just shoot at maximum resolution for the cameras, basically, since it doesn't really cost us anything in terms of time], but that's overkill for a lot of material.

For things where people might to print large enlargements of a small segment -- manuscript illumination -- then we'd shoot the illumination separately and at a higher res. That might be relevant for scanning home prints if you wanted to crop out just the part of the picture containing Auntie Jean, or whatever, but most of the time would be overkill.

Also, with most prints you're going to be hitting the maximum resolution of the paper pretty quickly. "Drug-store" 6" x 4" prints aren't going to exhibit the same level of detail as 8" x 10" silver-prints hand-printed with an enlarger.


Posted by: nattarGcM ttaM | Link to this comment | 03-11-09 9:41 AM
horizontal rule
249

219: yeah, because what do literature types know about Derrida?

Not as much as they think, frequently.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 11:51 AM
horizontal rule
250

Not as frequently as they think, temporally.


Posted by: JP Stormcrow | Link to this comment | 03-11-09 12:09 PM
horizontal rule
251

A course at the high school level that I have toyed with designing would use assessing claims in ads and PR as the core.

I knew a (public) high schooler a few years ago who was taking such a class. I was so jealous that someone was actually trying to teach her critical thinking skills.


Posted by: Sir Kraab | Link to this comment | 03-11-09 2:31 PM
horizontal rule
252

Der Tempuskontrast is eine transzendentallogische Aussageform; man versteht ihn nicht durch zufällige Merkmale der Morphologie bestimmter Sprachen, sondern sofern man nur überhaupt anschauungsbezogene Aussagen macht. Wir werden sehen, daß dasselbe für den Aspektkontrast gilt.


Posted by: ben w-lfs-n | Link to this comment | 03-11-09 2:40 PM
horizontal rule
253

God I fucking hate Bayesian statistics.

BG, on the off-chance you check this thread: linear algebra is probably harder than elementary statistics, so it depends on your goal. If you want to be able to read medical or economic papers, you'll need to know something about regression, which relies on linear algebra. (You could probably get by with an intuitive notion of what's going on.) But it does allow you to see how statistics could be generally useful to answer questions beyond the kind of "What's the average? How do we know it's really different from zero?" questions that they teach you in elementary statistics.


Posted by: Walt Someguy | Link to this comment | 03-12-09 10:54 PM
horizontal rule
254

So, I shoudl teach myself linear algebra first.


Posted by: Bostoniangirl | Link to this comment | 03-13-09 1:57 PM
horizontal rule