Validation is always good..

From Flowing Data:

“The lesson here isn’t about global warming. It’s that you shouldn’t take data lightly. When you’re dealing with data, you have to look past the numbers.”

I feel like this has been my mantra since I started my current job (November 2004), which it seems surprises more people than it should. From my blog, written in September 2005 (emphasis added):

“As I sit and ponder this confluence of things fibery, specifically knitting, though I imagine the same would hold for weaving as well, and number crunching, there’s a certain .. something that connects the two. I’ll try to explain, at least how it works in my head, but it might get tangled. Essentially, when I look at data, I’m trying to draw out the pattern, or the story, the data tells. Another way of thinking of this is trying to create the picture of what’s happening by finding the strands and threads that weave the whole together. A single data item, like a single strand of fiber, may be beautiful, but is rarely complex. Only by combining it with others and teasing out the patterns does the larger picture, in all its glory, become clear. Sometimes, like when working a delicate lace motif in a fuzzy mohair, the pattern stays diffuse and difficult to perceive until you set it off against some contrasting background. Sometimes, like when working cables in 100% cotton, you have to exert a fair amount of effort and a not-insignificant amount of force to bring the pattern out, but once it’s there, you can’t miss it. And sometimes, like when working with hand-dyed variegated yarn, you start out expecting a certain pattern only to find a completely different one emerging as you go. Of course, there are also the times, like when working with a luscious and soft wool in stockinette, when the pattern comes out exactly as you expected without an undue amount of effort. So, when I think about it that way, it makes a certain sense that data geeks are also commonly fiber geeks. It is, after all, sometimes easier to just follow the pattern than it is to find it.”

It’s nice to have some external validation of this approach, especially lately. My life has been consumed by work this summer.. sorry for the radio silence, but sometimes that’s just how things go. Major data systems conversions will do that to you (but I can now add the development and delivery of a two day training workshop on PeopleSoft Query to my resume). *shrug*

My garden is growing; I ate my first peas yesterday (yes, they’re very late) and have several squash and cukes and tomatoes and even a  pepper, and more onions than I’ll know what to do with (not really) happily ripening. I think the watermelon vine has bit the proverbial dust, however, and the birds are eating all the strawberries and not leaving any for me, but for now I can deal with that.

I’m also knitting some, but you’ll have to wait until I replace my camera to get updated pictures, which will have to wait until someone more knowledgeable about cameras can accompany me to a store to check some out. I finished the back of Celtic Icon, though, and the right front. So that just means the right front, the two sides, the two sleeves, and the hood and seaming.. *sigh* Baby steps for now. (Though I’d love to have this one done by fall.)

I’ve also been reading and listening to books (always am really). I’m about an hour and a half from finishing listening to the third book in Karl Schroeder‘s Virga series, which has been amazing, though I have to admit I’d expected him to follow different characters for the second and third books than he did.

f(ruit salad)

This will pr’bly only make sense to a few of you out there, but it was amusing to my cold-addled brain, so I thought I’d share.

Part of my job involves some nominal oversight for creating samples for electronic surveys. It’s not a big part, but because it’s something that we’ve been seeing a lot more of, it’s sort of been getting a lot of attention lately, so I’ve been more or less muddling my way through the process of trying to create a way to figure out how much electronic surveying we’re actually doing on our campus.

In the process, we’re trying to provide guidance to those conducting surveys about how large of a sample they reall need. It’s very easy to just say “Send it to everyone!” because with an email distribution, there’s no additional marginal cost for increasing the sample size. Unfortunately, when everyone starts doing that, recipients get fatigued from constantly being asked to respond to surveys and stop responding. So it’s one of those “tragedy of the commons” type problems where it’s really in everyone’s best interest to use an appropriately sized sample, even if any individual *could* (and usually does) take the easier option of just hitting up everyone.

So, the question of sample size is one that is covered in every basic research methods class, right? So it should be easy to figure out – run a little power analysis, figure out expected response rate, and you’re good to go. So I decided I needed to actually put myself through the paces yesterday for an institutionally sponsored survey and immediately realized that I couldn’t do a power analysis because I didn’t know what comparisons would be made with the data. So I emailed a bunch of friends in the faculty, some from our Psychology department and some from our Sociology department, and explained. And the Psych folks were scratching their heads along with me, but the Soc folks got it immediately and straghtened me out.

In summing up for all of us this morning, a dear friend made the following statement (edited somewhat to remove names and such): we were talking apples and oranges.. the question was how many apples do we need to represent the orchard within a confidence level. The Psych folks were talking about how many apples and oranges do you need to compare them to each other.” My response (again, remember – cold is eating my brain): “Having been brought up in an orange orchard, my thinking was more in line with the Psych folks, so I had to learn how to make fruit salad.. or.. something.”

This isn’t it either..

.. the general update that is. *smile*

Instead this is a call to the general ether for suggestions, but it needs a little background.

I’m having a rather lovely, glorious morning. I got to meet with a faculty member this morning and talk about what I could do to help them with some planning they’re trying to do for their department and it’s the part of my job that I tend to like once I start it because most people start into meetings with me a little nervous. I’m a statistician and like it or not, that comes with a lot of stereotypes and expectations that make people won’t don’t live and die by the numbers a touch anxious. But I, personally, am also both a social worker and a policy analyst, which I like to think allows me to put data into context, though in truth any good institutional researcher can and will do just that. The numbers can tell any story.. it’s the people with the context who make the numbers tell the *correct* story. My role is to help tell stories, help provide some framework to display the intricacies of what really happens. So it makes me happy when, somewhere around the middle of meetings like the ones I had this morning, the trepidation starts to fade away and the conversation becomes about the story and how we can build the framework to bring out the most important pieces.

I’m now looking at the rest of the day – a day with no other meetings, a day where I can kick off my shoes, close the office door, and turn up my iTunes and .. hrm. I’ve heard everything in my iTunes a million times or more and while some of it definitely fits well with my current mood, a lot of it is stale. So.. here’s the call for suggestions.

I need new music. I’m open to pretty much anything but rap and overtly religious stuff; most of my music currently runs toward folk-ish (Indigo Girls, Cowboy Junkies, Barenaked Ladies, Moxy Fruvous, The Paperboys, etc.), classic rock (The Grateful Dead, Tom Petty, Traveling Wilburies, etc.), and a bunch of random bits. I’ve got a couple Pandora stations tuned to the sorts of stuff I like, but I’ve also found those to be pretty stale lately. So.. what am I missing? What’s good and new and not to be missed? I have a preference for buying actual CDs instead of electronic music (and I’m fine with outlets like CDBaby – it’s the only way to get some of the stuff I like), but if it’s only available electronically, I’m okay with that, too.

I don’t care if you know me or not – if you’ve stumbled across this randomly or lurk or anything – drop me a comment with a suggestion.

Fence-sitting

I had to laugh this morning at Scott McLemee’s piece in Inside Higher Ed proclaiming the difference – or rather, what he’d define as such as there appears to be no truly common consensus – between a geek and a nerd:

As a nerd, my bias is towards paper-and-ink books, and while I do indeed use information technology, asking a coherent question about how any of it works is evidently beyond me. A geek, by contrast, knows source code….has strong opinions about source code….can talk to other geeks about source code, and at some length. (One imagines them doing so via high-pitched clicking noises.) My wife understands network protocols. I think that Network Protocols would be a pretty good name for a retro-‘90s dance band.

While the content of the article is not really about the difference between geeks and nerds, I find this distinction somewhat.. interesting all the same. I would claim, by Scott’s definition, to be neither properly a nerd nor a geek, but rather some combination of the two* – while the printed word holds value and mystique, I’m not so wed to dead-tree format that I don’t understand code, but I’m also not adept enough with code to speak in high-pitched clicking noises. Taking this a step further, I would say this appears to be yet another divide in which practitioners of institutional research sit the fence.

(More beyond the jump.. things got a little long!) Continue reading

Things I keep meaning to look into..

Thanks to Sara, I stumbled across Flowing Data a few weeks ago. It’s got me thinking a lot about data visualization and, as it’s the time of year when those sort of pipe dream projects get a little more chance of getting explored, I’ve been keeping track of the things related to data visualization and reporting and such that I want to look into more when I have some time. Below is a quick list, annotated more for my own reference.

PolicyMap – this one’s neat and I want to figure out if what it does is what I think it does and whether it might be worth paying for in order to pull demographic information about the areas from which we get our students. Knowing, for instance, if we get applications but not enrollments from particular areas and being able to dig into additional local and regional demographic information might be able to help us figure out if we’re not attractive as an option for students from particular backgrounds, which could then lead to further research about why.

blist – I’m not at all sure I understand what this is or maybe just not how it works, but I’m intrigued by the idea of sharing data structures across users. Seems like this might be terribly useful for IR folks using data from some of the larger student information systems (PeopleSoft, Banner, Datatel, etc.) to report out to the Common Data  Set and IPEDS..? As a more or less totally random aside, I’ve also wondered if the name originated from a slurring of “b-list” or as a slang word a la “blissed out”.

BigTable – not data visualization but rather data storage, this one breaks all kinds of commonly accepted database design and efficiency rules, and possibly because of that is terribly intriguing. No idea what possible application it might have for me, but it’s neat and therefore worthy of being checked out more.

Tableau – this was a vendor from the conference I was at last week and I’ll admit that I’m a little skeptical that it really can do everything it claims. There’s also the piece about sharing online through *their* servers that makes me twitch a little.. but.. I’m intrigued, so I’ll check it out further.

For what it’s worth, I’m also trying to figure out if I can carve out time this summer to learn more about Fuzzy Set Theory and how some lessons from econometrics and forecasting models, particularly  ARIMA, could be applied to enrollment projections.. yeah.. I tend to get optimistic about summer projects, but I think if you don’t, you just get stuck doing what you’ve always done and I guess I get bored with that too fast. *smile*

It’s not Wednesday yet, is it?

I have a love/hate relationship with this week. I love that the semester is finally over, that things are actually *slow* (or what counts for it around here), that campus is nearly deserted and very quiet. I hate that I lose just about all motivation to do anything in my office (there’s some twisted logic in there related to finally feeling like I don’t have to be running all the time, so I just want to stop altogether or something), that once I clean up the few things of interest all that’s left are the things that got put off because I needed more than two minutes in a row to figure them out, that I can’t really stop running because I’ll be out again all next week at a conference.

This year I also hate that my dad and Jack are ripping out a wall in our house and I don’t get to help. I mean.. remodeling sucks, really, I get that.. but they’re doing the fun part today and all I’ll get to see is the rubble and the dust once it’s done. (We’re making the (actually rather large) closet in our bedroom more accessible; it’s currently accessed by a standard door, but runs the length of the wall (10 feet?) so we’re ripping out the wall and putting in sliding doors and a shelving system to make it a real, functional closet.)

Even Celtic Stompgrass on Pandora is not helping much today, largely because it seems that they don’t really have all that much that fits that description in their catalogue so I’m to the point where I’ve heard all the songs they’re coming up with several times in the last couple days. Don’t get me wrong, I still like them (most of them anyway), but I want to find more new stuff. (And I’m really trying not to be too upset that I’ll miss the Paperboys by a little over a week..)

Life’s been.. interesting lately, but most of it is not really suitable for here, so you’ll have to take my word for it. *shrug*

Random Thursday

Tragedy.

As an Institutional Researcher, this is about 90% of what I do daily. Because Institutional Research is one of those more or less obscure careers – most people have no idea what it is, even a fair number of the rank and file in academia – there’s no predominant path people take to get here. In the three and half or so years I’ve been in this profession, I think that my background as a social worker has been far more useful than my background as a researcher or a statistician. Social workers are trained to see person-in-environment – to explicitly see the context of a particular set of circumstances. More than anything else I do, being able to set the context around any particular piece or set of data is by far the most valuable (and also most often overlooked by users of data). (If you want an example, ask me about the IPEDS Graduation Rate Survey sometime.. *smile*)

I apparently smell like jelly beans today. *shrug*

The Peterson’s Graduate Survey has become officially known as “The Evil Survey(tm)” by my student survey monkey. This amusing me for several reasons, not the least of which being that it is truly an evil survey.

I promised pictures of Ornette; here are a couple from when I hit the 3 inch mark a few days ago (row 36 or the 40 row cable pattern):

As usual, all pictures are clickable thumbnails.

I’m knitting the large size of the pattern and started with the right sock. I’m using Cherry Tree Hill Supersock Merino yarn on US1 needles. I thought I had a set of Harmony US1s, but I can’t find them so I’m using Clover bamboo needles, which are just not quite pointy enough for the cabling and causing the yarn to be splittier than I think it should be. Despite the heavily cabled appearance, the sock is fairly elastic – relaxed, it measures about 6.5 inches, but it stretches enough not to be tight or luck stretched out around my 10.5 inch lower calf. (I’d show you a picture, but it’s rather difficult to take a picture of your own calf, particularly when the interesting part of the pattern is on the outside!)

At Cookie’s request, I washed and blocked it this morning (I’m now through row 40, one full repeat of the cable chart), but apparently can’t take a non-blurry picture today; it also appears darker because it’s still damp.

The yarn did bloom a little in washing, but I didn’t lose as much stitch definition as it appears from the pictures. I might go outside at lunch and try to get some better pictures.

A caveat: I’m test knitting the pattern along with a couple dozen other folks, so it’s possible that the pattern I knit will be different from the final released pattern. There have been one or two minor revisions already, but for cosmetic purposes; the pattern is well written and easy to follow if you understand how to read charts. I’ve been doing the cabling without a cable needle, which is a bit of a new trick with the small gauge (and is another reason I’d really like some pointier needles).