Men Have Earned Nearly Double What Women Have? The Study Was Flawed!

Andy Saunders Andy Saunders

Published February 1, 2017 - 5:13 p.m. Modified February 1, 2017 - 5:14 p.m.

Current Champion

Nikolas Polsinelli, career statistics:
44 correct, 6 incorrect
2/2 on rebound attempts (on 5 rebound opportunities)
35.96% in first on buzzer (41/114)
4/5 on Daily Doubles (Net Earned: $10,600)
1/2 in Final Jeopardy
Average Coryat: $17,500

To win:
3 games: 64.111%
4: 41.102%
5: 26.351%
6: 16.894%
7: 10.831%
Avg. streak: 3.786 games.
(Click Here for Methodology )

Upcoming Encore Presentations

September 26-27: S42 E174 (May 7, 2026)
October 3-4: S42 E175 (May 8, 2026)
October 10-11: S42 E176 (May 11, 2026)
(Source: Bell Fibe TV program guide)

All Time Winnings

All Time Jeopardy! Winnings, Regular Play Only:

1. Ken Jennings $2,520,700
2. James Holzhauer $2,462,216
3. Matt Amodio $1,518,601
4. Amy Schneider $1,382,800
5. Jamie Ding $882,605
6. Cris Pannullo $748,286
7. Mattea Roach $560,983
8. Jason Zuffranieri $532,496
9. Scott Riccardi $455,000
10. David Madden $430,400
11. Julia Collins $428,100
12. Matt Jackson $411,612
13. Austin Rogers $411,000
14. Ray Lalonde $386,400
15. Harrison Whitaker $373,999
16. Adriana Harmeyer $349,600
17. Caleb Groen $346,968
18. Adam Remsen $306,415
19. Ryan Long $299,400
20. Arthur Chu $297,200

All Time Jeopardy! Winnings, Including Tournaments (and Consolation Prizes):

1. Brad Rutter $4,968,436
2. Ken Jennings $4,372,700
3. James Holzhauer $3,614,216
4. Matt Amodio $1,964,601
5. Amy Schneider $1,864,800
6. Yogesh Raut $1,598,403
7. Victoria Groce $1,023,801
8. Mattea Roach $897,983
9. Jamie Ding $885,605
10. David Madden $785,733
11. Cris Pannullo $754,286
12. Roger Craig $706,200
13. Andrew He $684,365
14. Larissa Kelly $671,930
15. Matt Jackson $623,612
16. Jason Zuffranieri $549,496
17. Paolo Pasco $548,717
18. Scott Riccardi $533,000
19. Jerome Vered $499,102
20. Julia Collins $495,767
21. Austin Rogers $493,000
22. Juveria Zaheer $491,000
23. Ben Ingram $449,201
24. Adriana Harmeyer $446,600
25. Buzzy Cohen $441,603

A recent study published at Maltese site Casino.org, reported on by the AV Club, made the outrageous claim that “men have earned nearly double what women have”.

It is the belief of The Jeopardy! Fan that these claims are inherently biased by a poor use of a dataset and completely flawed to the point that it is irresponsible to report on them.

The Casino.org methodology, while it acknowledges that the data is incomplete, appears to be working on the assumption that the incomplete data is representative.

As a Founding Archivist at J! Archive, I can categorically say that the data is far from representative, and not representative to the point that making any conclusions based on that data is improper and irresponsible.

One claim in the original study was that the 1991-92 season was “easiest”, because it had a ratio of “4.2 correct answers for every wrong answer”. That season, in J! Archive, has only 48 games out of 230 recorded. Of those 48 games, easily the featured set was Jerome Vered’s complete run of 5 games. It would therefore stand to reason that Jerome, who completely dominated and steamrollered his opponents in May of 1992, would boost that average significantly because his games take up so much of that season’s average. Its conclusion is thus flawed.

The other outrageous claim was comparing men with women. Again, the vast majority of the early games in J! Archive were put in there of players who later returned to the show in tournaments, either a Tournament of Champions, Million Dollar Masters, Super Jeopardy, the Ultimate Tournament of Champions, or the Battle of the Decades. Most of those “dominant” players were male. Thus, again, the dataset is inherently biased towards male players, and female players who might not have won enough (but still enough that it would have brought things back closer to its true average) are less included from earlier seasons (simply out of the fact that fewer of their games have appeared in private collections or video sharing sites over the years).

It is thus the opinion of The Jeopardy! Fan site that any story that reports on this study or dataset is inherently flawed and its conclusions are inherently invalid.

Looking to contact Andy Saunders for further information about this story? Check the About page!

—

I respectfully disagree with your prima facie rejection of the conclusions simply because we don’t want it to be true.
You criticize one non-provocative finding about the 1991 season because their data set was a mere 20% of the lot? And therefore, all other correlations and associations are null and void?

It doesn’t sound like they were irresponsible or wreckless, nor did the article I read seem judgmental or authoritative.

It is important to filter results. When Donald Trump says anything, you can start with it as false, and look for something factual. If any part is a lie, then anything from that source needs questioned.
Thank you for describing a potential softness in the data review. Until you provide the contradictory and improved data set, I will accept the data at face value.
We will not obtain equality and fairness by pretending we have more than we do.
kdmeares

Andy Saunders | February 5, 2017 at 12:18 am | Reply

Did you even look at the article? They listed clue values as 100, 200, 300, 400, 500, 600, 800, and 1000. This complete and utter lack of domain knowledge renders them, in my opinion, completely 100% unqualified to analyze this data.

Look at this statement: “While round one questions worth $1,000 had the lowest likelihood of being answered correctly (almost 1-to-1), questions valued at $500 proved to be the second most difficult. Additionally, categories worth $600 and $800 had a ratio of over 2-to-1 in both rounds, which just goes to show – you never know which question is going to be the one that stumps you.”

If you don’t know enough about the history of Jeopardy! to know that $500 and $1000 are the exact same difficulty level ($500 clues were at the bottom row of the Jeopardy! round until November 2001, before dollar values were doubled), you do not have the relevant expertise to make conclusions based upon the data.

In terms of an improved data set? Until Seasons 1-13 reach the same level of completeness as Seasons 14-33 do on J! Archive, a competing data set does not exist in order to do a proper analysis. Which, frankly, was the entire point in my calling it flawed in the first place.

I will continue to reject the conclusions made in the two articles wholesale. It was not a rejection out of not wanting it to be true; it was a rejection because, as a Founding Archivist of J! Archive, I am intimately aware of the limitations of the data set as it applies to early seasons and a knowledge of the inherent biases within those early seasons.