Sunday, August 28, 2005

The Harris Poll - Analysis & Conclusion

First off, some caveats about conflicts: I recognize that there is no way to accurately display direct connection between the conflict list in the previous post and actual biases. Some conflicts are much closer than others. For example, Bill Yeoman coached for 25 years at Houston and is still a legend there, while Larry Smith has 6 conflict schools and didn't leave with a good relationship at several of them. A long career at the same school may mean a bias toward that school, but also a bias against that school's rivals. An administrator that has worked at 6 schools may have dozens of biases toward and against schools. It's easy to say that since there are countless biases inherent to every voter, "they all cancel each other out", but I don't think the evidence supports that. Also, whether a conflict is in favor (Jack Thompson at WSU) of or against (Mackovic at Arizona) a particular school equally brings the poll into question.

I. Conference Categorization

For this section, I'm considering all conflicts equal (though I recognize that some are stronger than others and that the positive/negative dynamic exists). Also, if a particular voter has more than one conflict in the same conference (Dick Bestwick in the ACC - GT, UNC, UVA) , that voter only counts once for the conference.

SEC: 25 conflicts
Big XII: 22 conflicts
Big Ten: 20 conflicts
ACC: 17 conflicts
Pac-10: 17 conflicts
Conference USA: 14 conflicts
Big East: 12 conflicts
WAC: 11 conflicts
Independents: 10 conflicts
MAC: 9 conflicts
MWC: 9 conflicts
Sun Belt: 7 conflicts

Analysis: It is what it is. Some conferences are probably overrepresented (Independents, SEC, Conference USA). Some are probably underrepresented (ACC, Big East, MWC).

II. School Categorization

Same caveats - any conflict counts, though some are bigger than others. Categorized by number of conflicts.

6: UNC

5: Alabama, Notre Dame, Oklahoma

4: Auburn, Cincinnati, Kansas, LSU, Ohio State, Southern Cal, Tennessee, Wisconsin

3: Arizona, Fresno State, Hawaii, Illinois, Indiana, Iowa, Iowa State, Kansas State, Maryland, Memphis, Michigan, Missouri, NC State, Northwestern, Oregon State, Rice, South Carolina, Syracuse, Texas Tech, Tulane, Vanderbilt

2: Army, Ball State, BYU, California, Clemson, Colorado, Colorado State, Duke, Ga. Tech, Houston, Louisiana-Monroe, Louisiana Tech, Miami (OH), Middle Tennessee State, Mississippi State, San Jose State, Stanford, Temple, Texas, Tulsa, UCLA, UConn, UNLV, Utah, Virginia, Wake Forest, Washington, Washington State

1: Arkansas, Arkansas State, Baylor, Boise State, Bowling Green, Central Michigan, East Carolina, Georgia, Kent, Kentucky, Louisiana-Lafayette, Louisville, Marshall, Miami (FL), Michigan State, Minnesota, Mississippi, Navy, New Mexico, New Mexico State, Pitt, Purdue, Rutgers, SMU, Southern Miss, South Florida, Texas A&M, TCU, UAB, UTEP, Va. Tech, Wyoming

0: Air Force, Akron, Arizona State, Boston College, Buffalo, Central Florida, Eastern Michigan, Florida, Florida Atlantic, Florida International, Florida State, Idaho, Nebraska, Nevada, Northern Illinois, North Texas, Ohio, Oklahoma State, Oregon, Penn State, San Diego State, Toledo, Troy, Utah State, Western Michigan, West Virginia

Analysis: Several points to make:

  • Florida, FSU and Miami all are perennial Top 10 teams, and therefore are most likely to be involved with a BCS controversy than just about any othe trio of teams. Yet there is only one voter with even a nominal connection to any of these schools (Bokamper). Florida is the most poorly represented state in this poll. And this is inexcusable. Unless the rules have changed, the ACC relies on the BCS rankings in part to decide which teams go to which bowls. If I were a FSU or Miami fan, I'd be very upset.
  • Some schools are significantly better represented than in-state rivals. Example: Arizona vs. Arizona State, Oregon vs. Oregon State.
  • Two of the three main Northeastern programs have no representation at all - Penn State and Boston College. It isn't like there aren't any writers in the Northeast.
  • Nebraska, Florida State, Penn State, Florida: all traditionally impressive programs, all involved in National Title chases in the last 20 years. Combined: No voters tied to the programs. I'd even add West Virginia in here.
  • UNC, Kansas, Cincinnati: among the schools with the most conflicted voters. Is the Harris Poll going to be used for basketball as well?
  • If the state of Florida is underrepresented, the state of Louisiana is clearly overrepresented. 12 conflicts among 5 schools. And they said the Ice Cream Distributor guy from ULM was needed for geographical inclusion? I call BS. This ought to make USC fans still sore about LSU's 2003 BCS Championship pretty mad.
  • Several successful mid-major programs are underrepresented. Mid-majors may have more to gain/lose by BCS poll bias. Boise State, Bowling Green, Southern Miss, TCU, UAB, UTEP, Toledo, Northern Illinois, and Marshall all have one or fewer related voters. Watch one (or more) get screwed.

Just at this level, there are plenty of questions which should be raised and answered by the Harris Poll. And these problems, to me, bring the entire poll into disrepute.

III. Age Distribution

In a previous post I discussed how a slight slant of this poll toward older voters could throw off the entire thing. I view older voters as "pre-Baby Boomers". My take is that there is a generational gap in college football. The younger generation thinks that the Florida schools are the standard bearers, that Boise State and Fresno State can hang with anyone, that Louisville, Virginia Tech and Kansas State are traditional powers not bottom feeders. The older generation still thinks that Notre Dame and Penn State are dominant (or even relevant), that Alabama is the only team south of Ohio, that Boise State is a punchline not a threat. Yes, I know this is stereotypical. Yes, I know that individual voters are able to do the research which should be required no matter how old he is. But this slant is a potential pitfall for the poll, in my opinion.

Finding voters' birth dates online isn't exactly easy, so my data isn't as accurate as I'd like. However, at least 45 voters (out of 109) can be considered in that pre-baby-boom generation. Nearly half. I see this as a problem.

Some other interesting facts about some of the voters' ages:

  • Eddie Crowder retired from Colorado as AD nearly 20 years ago. Retired as coach 31 years ago.
  • Bump Elliott coached his last game 37 years ago.
  • Bob Hammel published his last column 9 years ago. (Might not seem so long, but writing isn't exactly as physically taxing - imagine Furman Bisher + a decade)
  • Fred Jacoby has "53 years experience" in administration.
  • Mike Lude last coached in the 60's and his bio uses the Kent State shootings as a frame of reference.
  • Ken Shipp's coaching highlight was 30 years ago.
  • Homer Rice started his career in coaching or administration 54 years ago.
  • Bill Yeoman retired 19 years ago, but started coaching 43 years ago.

IV. Other Sports

Several of the voters were not football players or coaches, but rather were primarily involved in other sports. I don't think football experience necessarily must be a requirement, but it's worth looking at.

  • Gene Bartow, basketball
  • Charlie Cavagnaro, basketball
  • Bob Frederick, basketball
  • Irwin Smallwood, primarily golf
  • Lou St. Amant, baseball coach
  • Frank Windegger, baseball

V. The Problem of "Lifetimers"

In the above categorizing there is a clear flaw in my reasoning, which I readily admit. Specifically, an individual may have a tie to 8 different schools because of a career that has had him bouncing around. While all that moving around affects the categorization, it also tends to dilute an individual's loyalties. For example, Dick Bestwick worked or played for UGA, GT, UVA, USC, UNC, and Missouri. It's hard to be a Bulldog homer when you also worked for two close rivals. On the other hand, there are several voters who spent their entire (or nearly all) career at one school, and naturally would have much more loyalty to overcome in order to vote without a bias intruding. While that might not show up in the categorization above, it is indeed a problem. Unfortunately, these individuals must be addressed individually.

  • Jake Crouthamel (Syracuse). 27 years as AD. Built nearly every facility on campus.
  • Peter Dalis (UCLA). Student, nearly 50 years at UCLA.
  • Don Fambrough and Bob Frederick (Kansas). Fambrough was coach twice, spent nearly 40 years at Kansas and has "the deepest and strongest ties to KU football." Frederick has 3 degrees from KU, worked a total of 23 years at KU and is currently on the faculty (pdf).
  • Jim Grabowski (Illinois). One of the most beloved players in Illinois history, member of several halls of fame, commentator for 26 years for Illini football.
  • Bob Grim and Steve Preece (Oregon State). Grim was a player for the Beavers and worked as an OSU football broadcast commentator for 15 years. Preece also was a player and broadcast the postgame interviews for Oregon State's radio network.
  • EJ Holub (Texas Tech). One of the most egregious conflicts. Red Raider legend, longtime donor to the program. Current fundraiser for the program (Associate Executive Director for Development).
  • David Housel (Auburn). 40 years on the Plains.
  • Rocket Ismail, Jim Morse, Roger Valdiserri (Notre Dame). Let the criticism begin, since Notre Dame has 3 of the closest conflicted voters. Ismail is the best known Irish player over the last 20-30 years. Valdiserri's job for 30+ years was public relations for the Irish. Jim Morse is probably the biggest conflict. "Major benefactor" has the athlete's tutoring building named after his family because of his donations, serves (for 25 years) on the Athletic Alumni Development Council (read - fundraiser), has endowed athletic scholarships and provided "aircraft assistance" to the athletics department (to fly Kevin White around trying to hire coaches?).
  • Charley Johnson (New Mexico State). Another egregious example. Legendary player at NMSU (the only number ever retired by the school), but also he's served on the faculty at NMSU (head of the Chemical Engineering department) and served as the Assistant to the President of NMSU for Athletic Progress.
  • Pat Richter (Wisconsin). This one is almost parody. 20+ years with Wisconsin as a student and administrator. Longest serving AD in the Big 10 when he retired. His biography is titled "Always a Badger". Seriously.
  • Paul Roach (Wyoming). Coach twice, AD. "Loves Wyoming deeply".
  • Jim Ray Smith (Baylor). Player, President of the Baylor Lettermen's Club.
  • Lou St. Amant (La.-Monroe). Coach, broadcaster, current president of the L Club (booster).
  • Jim Sweeney (Fresno State). Longtime coach, field named for him. Current fundraiser for the athletic department. Read this article to see how he'd vote (yes, I know he's kidding, but he kind of proves the point).
  • Jack Thompson (Washington State). Legendary player (1 of 2 retired numbers). Father of a current WSU player. "...near and dear to my heart for three decades."
  • Frank Weedon (NC State). 45 years at NC State. Current consultant. Longtime SID and Assoc. AD.
  • Frank Windegger (TCU). 40 years as AD, baseball coach. Only number retired (baseball).
  • Bill Yeoman (Houston). 25 years as coach.

VI. Conclusion

The biggest problem in my opinion with the poll is voters with personal interests in the results of the poll. The Harris Poll, as part of the BCS directly affects bowl selections, and therefore the amount of revenue a particular school may bring in. Many of these voters are directly involved (at times employed) in fundraising for schools which may benefit from increased revenue resulting from inflated poll positions.

Example: Notre Dame has a singular arrangement with the BCS. If they finish in a particular position in the BCS poll, they are guaranteed a BCS at-large position. Three of the voters have direct ties to the university, one of which is directly involved in fundraising. Is there an incentive to fudge the votes on the poll in favor of a 8-3 Notre Dame to push them into an automatic berth (and $12M+ unshared)? I think so.

Another example: Fresno State runs the table, is on the verge of a top 6 BCS finish, but just outside. Should Jim Sweeney (or any other WAC-related voter) fudge the vote in order to get the Bulldogs into a BCS bowl game (and bring in $12M+ into the conference coffers)? He's a fundraiser, wouldn't this raise funds?

And this problem isn't just limited to the top of the poll, though that's definitely where the most deleterious effects can be felt, economic-wise.

Consider this hypothetical: The ACC has used the BCS standings as a tiebreaker in terms of bowl selection slotting. Assume Boston College, NC State and Georgia Tech all finish 8-3, and tied for 3rd-5th in the conference. All are toward the bottom of the BCS standings. Is it possible that Frank Weedon and Dick Sheridan rate NC State a little bit higher than the others in order to ensure that NC State gets that 3rd place finish and a high profile bid to the Peach Bowl (which pays out a little more), while unrepresented BC gets stuck with an expensive (most schools lose money going there) trip to Boise State. Sound fair? And would an issue like this fly under the radar? Possibly.

The fact is that people have biases. It's natural. I spent 7 years in Athens, and I will never be able to put aside my love for the Dawgs. I can't imagine spending 40+ years at a school, the amount of loyalty that would engender.

Can voters put aside biases? Maybe. I believe people can think rationally and objectively. But when individual voters have external pressures, such as job related benefits and detriments, I think it's much more difficult to be objective.

Where this takes us, I'm not prepared to say. Some will say that this poll is just another reason to go for a playoff. I'm not prepared to go that far. Whether we like it or not, this is the system in place. And that's why the system needs to be scrutinized, to ensure that it's the fairest system in place. I cannot say that the Harris Poll's component provides for a level playing field for all teams. When some schools have multiple ad hoc representatives voting for them, while others have nobody even in the area voting for their interests, a fair system is not evident.

Sportswriters across the country have hinted that the poll is interesting and have done profiles of individual voters (or the "several voters have ties to Local U" story). But I haven't been able to find a comprehensive view of all the voters, and I think that when you look at the big picture, it's a much uglier picture than expected. I'd like to give a credit to the many local writers whose columns inspired me to take a stab at this.

The end result: The Harris Poll is problematic now and deserves criticism before it becomes a major problem when the inevitable BCS controversy rears its head in November.

Finally, I'd like to apologize if there are any errors in the above. This is a hobby for me, and while I've put a lot of time into it, I have other responsibilites too. If you see any errors, please email me or post in comments and I'll make clear corrections.