Friday, May 27, 2016

Learning About Athens

There's this huge new survey out about Athens-Clarke County (where I live). I hope to dig into its methodology and details on another day but while skimming the results I came across Question 14 in the appendix that asks:
Please indicate how much of a source, if at all, you consider each of the following to be for obtaining information about the Athens community and its activities, events and services:
Below are the rankings of those considered a "major source" by the respondents. The percentage who thought so is in parentheses (other responses were "minor source" and "not a source"). My comments are below the list.
  1. ACC web site (43%)
  2. Athens Banner-Herald online (36%)
  3. Athens Banner-Herald print (35%)
  4. Facebook (34%)
  5. Atlanta TV stations (33%)
  6. Flagpole print (32%)
  7. Flagpole online (28%)
  8. WUGA radio (27%)
  9. Emails from ACC Govt (25%)
  10. Other media outlets (23%) 
  11. Water bill inserts (22%)
  12. The Red & Black print (21%)
  13. Athens News Matters radio show (19%)
  14. R&B online (18%)
  15. Other radio talk shows (15%)
  16. (tie) radio ads and 102 FM's Magic Morning show (13%)
There are a few others, but you get the idea. Keep in mind these were provided to respondents, not volunteered by them. I'm shocked, frankly, that so many people consider the ACC site a major source. However, it seems to have been listed first in the questionnaire. Normally you'd randomly rotate the order in which these are presented but I can't find any suggestion that this was done. It's a subtle bias, if indeed it was not randomly handled. 

The R&B does surprisingly well given it's a student newspaper on the UGA campus.

As to methodology, I really want to dig into this. Briefly, 1,800 randomly selected households were sent a survey and 364 responded, which is a damn good response rate. You could fill it out in paper or online. That makes the margin of error here, if my math is correct, at 5.1 percent. So when you look at the results above, you'll find the MOE makes for very few differences and essentially the ABH and ACC sites are in a statistical tie for first. The survey undersamples blacks (17 percent in the survey, 27 percent in the county). There are other issues I'll write about at another time.

Tuesday, May 24, 2016

Hooker County, Nebraska

If you're of a sophomoric mind like myself, you see Hooker County and think, yeah man, when's the next plane? Just stop. The reason I'm writing about it is Hooker County, Nebraska, according to data I'm playing with, is ranked #1 when it comes to the proportion of people who belong to clubs or associations.

Keep in mind this is a county with barely more people than it has square miles -- under a thousand for each. According to the data there are six associations in the county with an association rate of 81.3 percent. #2 is 62.0 percent, Greeley, Kansas.

So what's up with this county? It's mostly white, it's mostly rural, with few people below the poverty level. It has no violent crime, a big fat "0" in the data, and overall looks like a nice place to live.

Oh, last place? There's a tie for a bunch of counties but we'll award it to Plute County, Utah, for no other reason than the alphabet says so.

Early Rankings

Like presidential polls this time of year, college football rankings in May are pretty much meaningless. So of course we pay attention to them. Below are five quickly grabbed early early rankings. Discussion below.

SB Nation
Bleacher Report
USA Today
Rant Sports
Sporting News
Florida State
Florida State
Florida State
Ohio State
Florida State
Ohio State


About the only consensus in these Top Five are Bama and Clemson. The final row shows the predictions for UGA (where I teach). The Dawgs don't look so hot, especially from USA Today. The beast of the SEC East, according to prognosticators, is Tennessee. That means UF will win it.

Flackery, UGA Style

UGA has a weekly PR release posing as news, called Columns, and just out is this piece about what is probably of no interest to anyone in the real world except college faculty: the question of gender equity in salaries. I get that most people could care less, but I'm briefly mentioning it here for a couple of reasons. First, I saw mention of it on a colleague's Facebook page and second, I'm on the University Council that discussed this. Also, I'm faculty and it's my blog, so I can care about and write about whatever I damn well please.

First, I often read Columns though usually I've heard it all before via Twitter. Helps me keep up with what's happening and it's usually a perfectly fine product. Still ....

... let me be clear -- this is a fluff PR piece that utterly fails to fairly represent the discussion at the Council and concerns about the study's methodology. At least give the other side a fair hearing, dammit. This is a university, not a corporation. And lemme guess -- the VP for Flackery is unavailable for comment, but she'll happily respond to an email in which she gets the questions in advance.

I don't want to get too PhDweeby here. It's summer and I have other stuff to do, like sipping drinks on a beach (except that I'm in my office, working, for no pay). The "news" article above glosses over any doubts about how the salary survey was conducted and offers not a single quote from folks with sophisticated statistical skills who raised serious concerns. I'm a stats nerd myself and though I don't do this kind of research, it's not all that terribly different from what I do every day.

I won't bore you with details other than to say there are reasons to doubt how this study was done or at least to question the lack of transparency by the consultant in how the study was conducted. Sorry, consultant folks, there's nothing proprietary about multiple regression. I do it every day. Wanna compare SPSS chops? Wanna go one-on-one in regression? You really don't.

There are far better people than I on the Council who raised this concerns, but you won't find this in the article. Not a hint. Though if you read between the lines, it's clear they're answering complaints during the Council meeting. Look at the quotes.

But you shouldn't have to read between the lines, UGA. C'mon. And stop calling it a newspaper. I don't think that word means what you think it means.

Friday, May 20, 2016

Rating Your Prof

If you go to Rate My Professor you can have a good time looking us up and amusing yourself with some of the caustic revenge comments by students about their teachers. Unfortunately (or, perhaps, fortunately) these ratings are not used for anything, especially as we're required after every class to have class evaluations. Those can count, but that's a different story.

(As an aside, I made a comment in a department meeting the other day about how the only faculty who don't like course evaluations are those with low course evaluations. Yes, I successfully trolled a faculty member who stepped right into it. Me for the win).

I dumped the Grady College faculty who are listed on the site and ranked them from first to worst, at least as far as Rate My Professor is concerned. A bunch of folks listed are no longer at Grady or UGA (or even living, for that matter). Some have retired, some have moved to teach elsewhere. Excluding them, here are the Top Five with the rating on a 1-to-5 scale and the number of ratings in parentheses (more on this after the list).

  1. Valerie Boyd, 5.0 (3 ratings)
  2. Joe Phau, 5.0 (2 ratings)
  3. Jeffrey Jones 5.0 (1 rating)
  4. Kirsten Strausbaugh-Hutchinson, 4.9 (4 ratings)
  5. Vicki Michaelis, 4.8 (6 ratings)
By the way, on the list above I'm 11th. As you methodologists out there know, so few ratings lead to enormous error. If we look only at those who have 10 or more ratings (and are still at Grady), a number I admittedly chose at random, we find a very different order.
  1. Jay Hamilton, 4.3 (31 ratings)
  2. Barry Hollander, 4.3 (12 ratings)
  3. Tom Reichert, 3.8 (22 ratings)
  4. John Soloski, 3.0 (43 ratings)
  5. Leara Rhodes, 2.9 (12 ratings)
Note that the list right above is all five folks who are still at Grady with 10 or more ratings, so making the Top Five is no biggie (hence I'm on the list). 

Are these predictive of anything? I'd love to correlate them with course evaluations but, unfortunately, I don't want to make an expensive public records request for all those evals. I can say my own 4.3 score is close to what I get on class evals, with those on a 1-to-5 scale usually falling around 4.5 or so, give or take. I've hit 4.9 as a high, 3.5 as a low.

Below, the complete list with faculty, living or dead, still at Grady or not, in order from first to worst.

Rating Last Name
5 Boyd 
5 Phua 
5 Jones 
5 Macias 
5 Ryan 
4.9 Brooks 
4.9 S.-Hutchinson 
4.8 Michaelis 
4.8 Jones 
4.8 Hillier 
4.7 Smith 
4.7 Sallot 
4.6 Fink 
4.5 Greenman 
4.5 Ahn 
4.5 Lin 
4.4 Dennis 
4.3 Hamilton 
4.3 Hollander 
4.3 Hudson 
4.3 Burns 
4.2 Suggs 
4.2 Parks 
4 Hume 
4 Thomas 
3.9 Weaver 
3.9 Hollifield 
3.8 Reichert 
3.7 Blackmon 
3.5 Benjamin 
3.5 Lee 
3.5 Deal 
3.5 Kwon 
3.5 Smith 
3.4 Dominick 
3.2 Becker 
3.1 Kreshel 
3 Soloski 
3 Middleton 
3 Murendehle 
2.9 Rhodes 
2.8 Tinkham 
2.7 Lariscy 
2.3 Griswold 

A Tale of Two Polls

Two presidential horserace polls went into the field at about the same time and the results reported within a day of each other. One was by Fox News, the other by CBS/NYTimes. They have significantly different results.

The Fox News poll has Trump ahead by 3 percentage points (within the margin of error). The CBS/NYT poll has Clinton up by 6 percentage points (beyond the margin of error). That's a 9 percentage point difference, enough to make you ask, WTF?

As Ron Faucheux notes in his lunchtime email blast, the difference may very well lie with how the two polls measure "independent" as a category. While the results of these two polls among Republicans and Democrats are similar, it's the Independents where we see differences. In the Fox poll, Trump beats Clinton 46-30 among Independents. In the CBS/NYT poll, it's a 40-40 tie.

What's going on?

Look hard at the number of "Independents" in each poll. For Fox News, only 16 percent of respondents classify themselves as such. In the CBS/NYT poll, 36 percent are classified as "Independents." How can there be such a huge difference in the number of Independents floating out there? One possibility is question wording.

Here's the Fox News question:
When you think about politics, do you think of yourself as a Democrat or a Republican?
Now, here's the CBS/NYT question:
Generally speaking, do you usually consider yourself a Republican, a Democrat, an Independent, or what?
See the difference?  The Fox question doesn't offer "independent" as a choice. You, the respondent, must volunteer it, making this more of a forced-choice question. Built into the CBS/NYT question is the opportunity to choose Independent, "or what." If the data from these questions are used in weighting the sample, it can have small effects, or even big effects, on the results.

Which is the better method? Hard to say. And keep in mind these are surveys of "registered voters," not "likely voters," so they're predictive power so early in the campaign are, honestly, meaningless -- especially when you consider the election is not nationwide but state-by-state. Still, there's much to learn in two polls, out at the same time, with different results.

Monday, May 16, 2016

UGA Buildings

In a moment of summer boredom (yes, it's summer, if you're an academic), I messed around with some data to see how many buildings UGA owns and where they're located. I know, I know ... sounds dull, and obvious. They're all in Athens, right?

Most of them are in Athens -- but only 39 percent. There's a map at the bottom of this post for your enjoyment. I think the data are right.

Surprising about Athens, eh? Out of the 1,690 buildings on the UGA database, 659 are listed as being in Athens. So, you ask, where are the rest of them? Glad you asked since I need something to do with this spreadsheet. Here ya go, the Top Five:

1. Athens (39.0 percent)
2. Eatonton (12.7 percent)
3. Tifton (10.8 percent)
4. Griffin (5.4 percent)
5. Watkinsville (4.3 percent)

Eatonton is the Rock Eagle camp, as anyone who has had kids knows all too well. Tifton and Griffin are obvious, UGA has campuses there, and Watkinsville is the location of a lot of ag stuff. Near the end of the list is the fun stuff -- a building in Costa Rica, Washington D.C., and my favorite, Oxford University (104 Banbury Road, to be exact).

The weirdest? UGA has 40 buildings listed in Aiken, South Carolina. Why? Mostly having to do with the Savannah River site, I suppose.

Again, the map is below. Expand it to see UGA stuff in Italy and Oxford, etc.