March Madness Basketball Tournament:

There is More Parity on the Men’s Side than on the Women’s Side

By Dr. Paul Beckman

pbeckman@sfsu.edu

Department of Information Systems

College of Business

San Francisco State University

(Copyright 2023)

INTRODUCTION

I completed this research project for two reasons.  First, I wanted a mathematical answer to a question: in the setting of the annual March Madness college basketball tournament, is there more “parity” on the men’s side than on the women’s side?  Second, I wanted to show that some analytical methodologies are fairly simple to implement if you go about them in a sensible manner.  For this research project, “parity” is defined as the likelihood that a low-ranked team beats a higher-ranked team.  That is, more parity exists when more low-ranked teams beat higher-ranked teams.  Conversely, less parity exists when fewer low-ranked teams beat higher-ranked teams.

The reason I wanted to answer the question above is that I’ve read several sports analysts who say they believe there is more “parity” in the men’s March Madness tournament than in the women’s tournament.  By this, they generally mean that upsets, where a lower-ranked team beats a higher-ranked team, is more likely to happen on the men’s side of the tournament than on the women’s side.  While this is generally accepted as true, I hadn’t come across any truly analytical support for these statements regarding parity.

My approach to gathering data, analyzing it, and coming up with a result was to consider first how to measure “parity.”  It is certainly possible to collect an enormous amount of data about every college basketball game, including which teams played, where they played, when they played, who did not play because of an injury, and numerous other factors associated with winning a game.   However, I wanted a more general and basic process that would be easy to understand, easy to replicate, and yet would still yield insights about parity across many teams for both men’s and women’s college basketball.

In the March Madness tournament, the 64 teams that comprise a full tournament bracket (after four teams are eliminated through a set of four “play-in” games) are distributed across four regions comprised of 16 teams each.  Each region contains one “1” seed, one “2” seed, one “3” seed, etc., up to the “16” seed.  A team considered “better” it is assigned a lower number indicating it is a “higher” seed.  Seed numbers are assigned indicative of that team’s quality or ability, and a great deal of time and thought goes into determining seed numbers and region assignment.  Even with this significant time and thought, there are many cases where fans, coaches, and/or players feel that the seeding outcome is not a proper assessment of their team’s ability or quality.

From this simple description of the seeding process for the tournament, it became obvious that calculating the “Sum Of Seeds” value at a particular round of the tournament is one way to measure “parity.”  (Other researchers have used the “Sum Of Seeds” calculation to explain or identify some anomaly associated with the tournament but none that I found used this approach directly to examine or resolve the issue of “parity.”  See the References section below for similar research projects.)  For example, at the “Sweet Sixteen” round of the tournament, there are 16 teams left vying for the championship.  If one were to calculate the sum of the seeds of these remaining 16 teams (four teams coming out of each region), then one would have a basic understanding of the parity of that year’s set of Sweet Sixteen teams.  That is, if there were greater parity across teams, then the Sum Of Seeds value would be higher because more low-ranked teams (with higher seed numbers) would remain after beating higher-ranked teams (with lower seed numbers).  If there were less parity, then the Sum Of Seeds would be lower because more high-ranked teams beat lower-ranked teams.

DATA AND RESULTS

With this description of my analytical methodology, the Sum Of Seeds at the Sweet Sixteen level of the March Madness tournament for both men’s and women’s teams since the year 2000 is:

The result of calculating the difference in values between these two sets of data (using Microsoft Excel’s “t-Test: Two-Sample Assuming Unequal Variances”) yields:

CONCLUSIONS

The p-value results from the statistical calculation above shows that, since the year 2000, there is a statistically significant difference between the values of the Sum Of Seeds from the men’s versus the women’s tournament at the Sweet Sixteen level of the March Madness tournament.  Therefore, when using the Sum Of Seeds analytical methodology and including all years since 2000, there is a difference in parity between the men’s teams and the women’s teams.  That difference shows more parity on the men’s side of the tournament than on the women’s side.  Note that this statement about parity is true for the dataset that includes all years since 2000, not for each individual year since 2000.  I was looking for only a general answer to the question about parity, so I did not complete a time-series analysis using this dataset, which would indicate if parity was changing over time.

My goal with this little analysis was to first, succinctly and definitively answer the question about “parity” differences between men’s and women’s college basketball teams, and second, to show that meaningful analyses sometimes can be performed with fairly simple analytical methodologies.  The answer to the “parity” question comparing men’s and women’s is: using the Sum Of Seeds analytical methodology at the Sweet Sixteen round of the March Madness tournament, there is more parity on the men’s side of the tournament than there is on the women’s side.

As an aside, there are several proposed reasons why this difference in parity exists.  Some of those reasons relate to the overall number of male and female basketball players, the greater financial benefit for “one-and-done” players in men’s versus women’s college basketball, and less overall support for women’s college sports than for men’s college sports.

ASSUMPTIONS

I made a couple of assumptions when applying my analytical methodology.  One of those assumptions is that a seed number is a valid measure of team ability.  There is no perfect approach to assigning seed numbers to teams, but the current process is generally accepted as useful and appropriate.  Another assumption I made is that the final 64 teams are indeed the “best” 64 teams in the country.  The approach for selecting those teams is also under constant review and debate, but is also generally accepted as valid.

REFERENCES

A miracle or a trend? Saint Peter's a sign of parity in NCAA (March 26, 2022); URL: https://www.usatoday.com/story/sports/ncaab/2022/03/26/a-miracle-or-a-trend-saint-peters-a-sign-of-parity-in-ncaa/49988093/

First Round Shows There’s Plenty of Room for Madness in the Women’s Tournament (March 20, 2022); URL: https://www.si.com/college/2022/03/20/belmont-princeton-first-round-upsets-womens-bracket-march-madness

In the NCAA women’s tournament, Madness reigns, but upsets are the exception (March 17, 2022); URL: https://www.washingtonpost.com/sports/2022/03/17/march-madness-upsets-women-tournament/

March Madness “Anomalies”: Are They Real, and If So, Can They Be Explained? (February 21, 2020); URL: https://www.tandfonline.com/doi/abs/10.1080/00031305.2020.1720814

NCAA Basketball Tournament: Is There Parity in College Basketball? (March 24, 2010); URL: https://bleacherreport.com/articles/368375-parody-in-college-basketball

Parity: Why Women’s Basketball Is More Unpredictable (March 25, 2022); URL: https://www.2adays.com/blog/parity-why-womens-basketball-is-more-unpredictable/

The middle-seed anomaly: why does it occur in some sports tournaments but not others? (May 7, 2021); URL: https://www.degruyter.com/document/doi/10.1515/jqas-2020-0065/html?lang=en