VoteFair!

VoteFair  ranking

Voting ID:

VoteFair American Idol poll

Season 14 Commentary

What's goin' on?  A demographic shift, and gender bias

The fans of American Idol have aged.  The show wants to help shape pop-music culture, which means they want to attract fans who are teenagers and young adults, because they are the people who buy more pop & rock music, and attend more pop & rock music concerts.  But the teenagers and young adults who were big fans in the earlier years of American Idol — and who now dominate the official voting — are no longer in the age range that the pop-music industry focuses on.

When I ask young women if they are fans of American Idol, they typically say "No, but my mother is a big fan."  And I've heard that the American Idol concerts are heavily attended by women in their thirties and forties (who are stereotyped as “cougars”).

The result of this demographic shift is a strong gender bias in favor of “boys” rather than “girls.”  That is why some great female singers, namely Jax Cole, Tyanna Jones, Joey Cook, and Sarina-Joi Crowe, were eliminated earlier than they deserved based on musical ability.  (Another factor this season was the presence of two very experienced — and therefore polished — male singers, Nick Fradiani and Rayvon Owen.)

In past years, vote splitting has accounted for the early elimination of some top female singers.  This year, the bias in favor of males was so strong that it overrode the vote-splitting advantage that favored Jax Cole.

To help measure the extent of this gender bias, the new cross-season poll includes all these female singers.  Doing so requires splitting the ranking into two questions, so you will see this season's Top 3 singers in both ranking questions.

Results of Top 3 poll, and why no Top 2 poll

There was not a Top 2 poll.  Why?  I didn't see or hear last week's show (because I was running the sound system for a local singing group), so I didn't hear the announcement about this week's Tuesday performance.  I didn't learn about the elimination of Jax Cole until late last night.

So, the votes that were cast after the start of Tuesday's show (on the east coast) will be removed.  I apologize for the ignoring of those most recent ballots, but there is no fair way to use those ballots.  Here are the final results of the Top 3 poll.

Popularity Choice VoteFair
ranking
score
Traditional
vote count

(for comparison)
First-most popular Jax Cole  Ranking score is 54 for Jax Cole 97 voters marked Jax Cole as their first choice
Second-most popular Clark Beckham  Ranking score is 26 for Clark Beckham 105 voters marked Clark Beckham as their first choice
Third-most popular Nick Fradiani  Ranking score is 0 for Nick Fradiani 60 voters marked Nick Fradiani as their first choice

 

De-corrupted results, Jax ahead of Clark, 114 to 94

This week, none of the three is safe.  In particular, vote splitting between the two males does not overcome the tendency for the winner to be a male (when the singers are somewhat equally talented).

After removing invalid ballots, here are the results as of Sunday evening.

Popularity Choice VoteFair
ranking
score
Traditional
vote count

(for comparison)
First-most popular Jax Cole  Ranking score is 58 for Jax Cole 87 voters marked Jax Cole as their first choice
Second-most popular Clark Beckham  Ranking score is 26 for Clark Beckham 82 voters marked Clark Beckham as their first choice
Third-most popular Nick Fradiani  Ranking score is 0 for Nick Fradiani 41 voters marked Nick Fradiani as their first choice

 

Poll closed, Wednesday de-corrupted results, Jax still ahead

As of mid-day Wednesday, here are the results after removing just a few invalid ballots.  Jax Cole is slightly ahead of Clark Beckham (136 to 133).

Popularity Choice VoteFair
ranking
score
Traditional
vote count

(for comparison)
First-most popular Jax Cole  Ranking score is 56 for Jax Cole 76 voters marked Jax Cole as their first choice
Second-most popular Clark Beckham  Ranking score is 38 for Clark Beckham 103 voters marked Clark Beckham as their first choice
Third-most popular Nick Fradiani  Ranking score is 20 for Nick Fradiani 56 voters marked Nick Fradiani as their first choice
Fourth-most popular Rayvon Owen  Ranking score is 0 for Rayvon Owen 37 voters marked Rayvon Owen as their first choice

 

De-corrupted results, Jax ahead (104 to 91)

Below are results that were calculated mid-day Sunday, immediately after removing all invalid votes, especially the latest 100-plus ballots from the greedy voter in Toronto (in the North York suburb), who is now ranking Jax Cole in last place, apparently in a childish attempt to “help” the greedy voter's favorite choice, who is Clark Beckham.

This corruption in a for-fun poll nicely demonstrates the importance of fairness in real life.  Places like U.S. state capitols and Canadian provincial capitols — including Toronto — attract greedy people who fail to understand the economic suffering that they cause when they use money to change election results in ways that favor their greedy goals.  And then those biggest campaign contributors, and their puppet lawmakers, fail to comprehend that their lack of ethics is the root cause of the economic suffering of average hard-working citizens.  That suffering is what sparks political protest demonstrations, including the annual May first marches in the U.S.

The opposite of corruption is fairness.  And the way to get fairness in a democracy is to ban the use of single-mark ballots, and instead use ballots that ensure that the election winner is actually preferred by a majority of voters (compared to any other candidate in the race).

Popularity Choice VoteFair
ranking
score
Traditional
vote count

(for comparison)
First-most popular Jax Cole  Ranking score is 57 for Jax Cole 59 voters marked Jax Cole as their first choice
Second-most popular Clark Beckham  Ranking score is 37 for Clark Beckham 74 voters marked Clark Beckham as their first choice
Third-most popular Nick Fradiani  Ranking score is 19 for Nick Fradiani 42 voters marked Nick Fradiani as their first choice
Fourth-most popular Rayvon Owen  Ranking score is 0 for Rayvon Owen 24 voters marked Rayvon Owen as their first choice

 

Extremely(!) greedy Canadian in Toronto, and Jax now more popular than Clark

In the past few days a voter in Toronto Ontario (Canada) has attempted to cast more than 630 ballots!  He or she used three different computers to cast these votes.  That's a record!  And it's foolish because all(!) ballots from outside the United States are removed.

Currently, with 186 valid ballots, 97 voters prefer Jax Cole over Clark Beckham, and 83 voters have the opposite preference.  This puts Jax Cole into the most-popular position.  Apparently many viewers agree that Clark's most recent song choice was not a good one.

If you see Clark Beckham jumping way ahead in this poll, consider that the greedy Toronto “voter” might still be oblivious to the fact that his or her ballots (all of them) are being removed.

Swimming faster than the slowest

Why has Rayvon Owen been able to stay in the competition in spite of consistantly ending up in the bottom two?  In past seasons the reason was that either vote splitting or an unfair voting advantage typically accounted for delayed eliminations.  But in this case Rayvon Owen has been improving each week, while the other singers have not been improving as quickly or dramatically.

In past seasons there have been singers who were staying in the competition in spite of obviously being the worst singer.  Those cases revealed the unfairness of single-mark ballots and the sometimes unfair vote-counting methods — such as being vulnerable to automated voting.  Although there could be some unfair voting for Rayvon Owen, such as some of his fans getting lots of Facebook accounts and Google accounts, that alone is not accounting for his “luck”.

As explained in previous seasons, here is a useful way to understand single-mark ballots — which are the ones used in government elections and somewhat like the ones used on American Idol.  It's like a swimming contest — where the goal is to outswim a slow shark.  Being out front is not necessary.  All that's needed is to swim faster than the slowest swimmer.

Translated into this singing contest, Rayvon Owen does not have to be the best singer.  He only has to sing better than just one of the other singers.  And lately he has been doing that each week.

Very greedy Canadian in Toronto

This week there were lots and lots of stuffed ballots from a voter in Toronto Canada.  But since his or her favorite is Clark Beckham, and the voter didn't express secondary preferences, those ballots didn't change the ranking results.

As a reminder, all ballots from outside the United States are removed.

Are Twitter votes fair?

The Twitter votes have another potential unfairness — besides the geographic advantage of only the East coast voters getting to vote.  Twitter voting can be automated.  This potential unfairness is similar to the use of “modem” voting in the early years of this show, especially during Season Two when Ruben Studdard won.  Back then, lots of phone votes were generated by computers that were programmed to automatically do the calling.

Is this unfairness actually happening?  We don't know.

If the vote counters want to make sure that automated Twitter-based “ballot stuffing” voting isn't happening, they can do a statistical analysis of where — city-wise — the votes are coming from.  If a large portion of the votes for the “more popular” contestant are coming from a few cities or a small region, and the votes for the other contestant are coming from a wide diversity of locations, then an unfairness is likely.  Simply showing the vote densities on two maps would reveal whether this is happening.

Did this unfairness happen last week when Joey Cook was eliminated prematurely?  We don't know.

Phoenix (Arizona), Buffalo (New York), Saddle River (New Jersey), & Muskegon (Michigan)

These are the cities from which a few greedy voters attempted to cast the most stuffed ballots.

A different kind of gender bias

The pairwise contest between Joey Cook and Rayvon Owen reveals a clear gender bias in favor of the male.  This makes sense because females are more numerous in terms of being the most dedicated fans of American Idol, and they tend to prefer male singers.

This gender bias has not been as obvious in the official voting.  Why?  Vote splitting!  When a stereotypic female fan votes for their favorite male, they are shifting their votes to their favorite, which this season is Clark Beckham, and that leaves fewer votes for another male, such as Rayvon Owen.  But as long as their favorite (Clark) is not in the pairwise Twitter-based elimination contest, Clark's fans can give their Twitter votes to the (only) male in the pairwise contest.

Here in the VoteFair poll, the results indicate that Joey Cook is more popular than Rayvon Owen.  In fact, last week the VoteFair counting details reveal that 128 voters prefer Joey Cook over Rayvon Owen, and 53 voters have the opposite preference (and 15 voters do no have a preference between the two).  The week before, the numbers were 120 voters and 36 voters (with 14 having no preference).

So it might seem that the official votes are “wrong.”  They aren't.  Remember that the VoteFair poll limits each voter to a single ballot, whereas official voting facilitates submitting lots of votes from each voter, and this difference accounts for the bias in this situation.

As a clarification, if Joey Cook were from the West coast, then she would be at a huge disadvantage because fans on the West coast do not get to participate in the Twitter poll (unless they are told by an East-coast friend about the “need” to vote with a specific Twitter tag).  But both she and Rayvon are from East-coast time zones (Louisiana and Virginia), so neither is at a significant geographic disadvantage.  None of the remaining singers are from west of the Mississippi River, so this geographic bias will not apply in future weeks.  (This bias was a disadvantage for Daniel Seavey, but his talent limitations were a bigger issue.)

So why is this week's gender bias “different”?  Compare it to the arenas of politics and business.  In that world males have tended to prefer male politicians and male employees.  As this week's American Idol results show, female fans tend to prefer male singers.  Specifically, American Idol voters tend to prefer the opposite gender.  By contrast, in the world of politics, historically voters have tended to vote for the voter's own gender.  (Not that female voters have had any opportunity to vote for competent female politicians.)  Finally that bias is changing.  Not that the “new” bias is the opposite.  Rather, the bias is shifting in favor of talent, not gender.

Of course another bias that influenced this week's Twitter-based elimination results is looks.  Alas, it might take many generations before singers and politicians who are the most talented become more appreciated than the less-talented singers and politicians who are better-looking.

Down to 7

No surprises this week.  (At least not that I know of; I wasn't able to watch or listen this week.)

Phoenix Arizona and Charleston South Carolina

These are the cities from which two greedy voters cast the most stuffed ballots.  Fortunately VoteFair ranking results do not easily shift as a result of stuffed ballots, so the results were barely affected.  This resilience occurs because rankings contain so much more information than single-mark ballots, and because pairwise counting is so much more meaningful than just looking at some kind of “most” or “fewest” count, which is what instant-runoff voting and some other simplistic methods do.

Pairwise counting, finally!

The new “fan save” does a comparison between two contestants, and that's the way vote counting should be done: pairwise!

Of course the use of Twitter shifts the voter demographics.  This is significant because women in their thirties and forties are an especially large group of voters, and they might tend to favor Daniel Seavey rather than Rayvon Owen, and many of them might not have Twitter accounts.  As another point, the five-minute voting timespan prevents voters from using multiple Google accounts and multiple Facebook accounts, and eliminates the ability of fans to alert their non-watching friends that it's time to cast a vote.  So, in these ways, the “fan save” offers fairness advantages.

In this situation the result was fair because Daniel Seavey is clearly the weakest singer.

Now that the judges have used their only save, the new “fan save” provides a way to rescue a deserving singer from being the victim of vote splitting.

Did the show's producers encourage the judges to use their save early to pave the way for this “fan save”?  If so, the motivation was to increase the drama.  That is more important than increasing fairness.  Fairness is important in politics because many people suffer under unfair manipulations.  In the entertainment world, drama is more important than fairness.

From 11 to 9

With so many contestants and the need for two eliminations this week, vote splitting could have happened easily.  Fortunately it didn't happen.  The two eliminated singers — Maddie Walker and Adanna Duru — were among the least popular.

Likely winner

With the likely winner — Clark Beckham — being so obvious, this season is becoming predictable.  In previous seasons the winner was not as predictable.

Ooops

Last week I didn't have much time, and the ballots for the first week of the Top 11 were lost.  If readers here express an interest in seeing those results, I might be able to make the time to dig them up.

Judges' mistake this time?

The judges may later regret using their only "save" on Qaasim Middleton.  With so many singers remaining, there is likely to be lots of vote splitting for a few weeks, and this is likely to make the elimination sequence unfair.  With two singers getting eliminated next week, one of those eliminations is likely to be someone who does not yet deserve to be eliminated.  If Qaasim Middleton is one of the two singers eliminated next week (which is likely), the judges will have wasted their only chance to protect a possible winner from being eliminated.  As a clarification, the poll results here indicate that Qaasim Middleton is very unlikely to be the final winner — even though he's a great singer.

Judges' mistake?

Should the judges have used their "save" on Sarina-Joi Crowe?  No, because the next week the same kind of unfairness (or worse) would be likely.  As another point, Sarina-Joi Crowe might have gotten to the top 3 or 4, but she would not have won.  The judges need to reserve their one "save" in case a possible winner gets the fewest first-choice votes.

Top 11 elimination surprise

A big surprise on Thursday!  Singer Sarina-Joi Crowe was eliminated even though she has been the most-popular female singer for the last two weeks in these VoteFair polls.  What happened?  There are several factors that might have combined to produce this odd result:

  • Vote splitting:  The person with the fewest first-choice votes is not necessarily the least popular!  The only way to judge popularity of the one-person-one-vote variety is to ask each voter to rank all the choices.  Instead American-Idol voters must split their 20 official votes among their favorites.  With this unfairness, vote splitting is going to happen.
  • Vote-counting glitch:  Voting digitally is vulnerable to big mistakes.  The biggest example was the phone-system failure years ago that caused Constantine Maroulis to be eliminated very early; ironically that failure occurred because his phone lines were overloaded (or appeared to cause an overload and a telephone employee mistook it for a denial-of-service attack).  Are the vote counts at the American Idol website and the vote counts at Google likely to be 100% accurate?  That's not likely.
  • It's likely that fans of Sarina-Joi Crowe were also fans of one or more male singers, and those fans used their official votes on the male singers.  Here the VoteFair polls for males and females were separate, and ranking was used, so the overall popularity that is measured here was not affected by this divided allegience (and the resulting vote splitting).
  • It's possible that fans assumed that Sarina-Joi Crowe was getting plenty of votes and chose to give their official votes to other singers.  This is part of what happened in Season 3 when Jennifer Hudson was eliminated very, very early.  Perhaps, as happened in that case, time will reveal that Sarina-Joi Crowe did not lack popularity, and instead she was the victim of the show's unfair voting method: using only first-choice votes.

Keeping it at 12

For the coming week Sarina-Joi Crowe will continue to be in the VoteFair poll because the process of updating the poll is a lengthy process and I'm busy with another project.  Also, admittedly I was not expecting another elimination this week because I only watch (or just listen to) the Wednesday shows — because I no longer have a way to record TV broadcasts.  (On that topic, the only broadcast TV shows I really watch are The Big Bang Theory and Nova, and they are available as streaming downloads, unlike American Idol which is only downloadable to cable subscribers.)

Top 12 results

No surprises this Wednesday, especially because the judges were able to compensate for minor vote-splitting issues at the less-popular levels.

Trends

Some trends are finally emerging.  As of Monday morning, after removing stuffed ballots (from 3 people) and non-U.S. ballots, the following singers are emerging as favorites: Clark Beckham, Quentin Alexander, Nick Fradiani, Qaasim Middleton, Jax Cole, and Sarina-Joi Crowe.  In addition there are a significant number of people whose first choice is Daniel Seavey, but he is not popular with the fans of the other singers.

What is likely to happen?  When the males and females are grouped together for voting, females will be eliminated more often than males — because there are fewer popular females compared to popular males.  (This might appear to be gender bias, but it isn't.)  Then the most popular female singers (who currently are Jax and Sarina-Joi) will be relatively "safe" while the remaining weaker male singers get eliminated.  This year the top two is sure to include at least one male.

We don't yet know the balance of popularity between the top males and top females because they haven't yet been compared in this VoteFair poll.  The males and females will be combined here when the total number of singers is down to 12 singers.

As always, these predictions are based on popularity, as if each viewer could only vote once.  The actual outcome depends on which singer's fans are the most dedicated in voting repeatedly (such as by using multiple Facebook and Google accounts), and the most able to recruit other people to vote for their favorite.

Uncertainty

The VoteFair poll ranking for some contestants changed dramatically between the first week and the second week.  This reflects the fact that some performers had a good performance with a good song choice one week, and then had a weak performance with a poor song choice the next week, and vice versa.  This variation, combined with voters not being familiar with all the contestants during the first week of voting here, means that there were not any big surprises among the eight singers eliminated.

Poll closed until females eliminated

Each week the poll closes at (or slightly before) the broadcast in which contestants are eliminated.  Because there are eliminations on both Wednesday and Thursday, the poll is closed for about 30 hours.  (This time includes the delay to the Westcoast broadcast, and the time needed to update the poll.)  This still leaves you six days in which you can vote (once).  Thanks for your patience.

Possible surprises

The current poll results (as of Monday morning) indicate that Alexis Gomez and Riley Bria do not have lots of fans who rank them as their first choice, yet they are in the middle of the popularity ranking.  This means that it's possible they won't get enough votes to stay on the show for another week, even though Loren Lott and Shi Scott and Michael Simeon and Trevor Douglas are more "deserving" of being eliminated.  With so many contestants this week, an unfair surprise elimination — or two — is quite likely.

Second week for top-24 poll

A new poll has been started so that after tonight's and tomorrow-night's first unedited performances you can rank the top 24 singers.  If you vote before the females perform, you do NOT get to vote a second time (after the females perform on Thursday).

In last week's poll the results for the males reveal that the current favorite male singers are Clark Beckham, Michael Simeon, Daniel Seavey, Adam Ezegelian, and Nick Fradiani.

The results for the females reveal that the current favorite female singers are: Jax Cole, Tyanna Jones, and Shannon Berthiaume.

As usual, there were a couple of greedy people who tried to "stuff the ballot box," and there were a few votes from outside the U.S., and those votes were removed.

 


Like these American Idol polls?  Spread the word!  Also you are invited to voice your opinion — about VoteFair ranking (not about contestants) — on the Testimonials page.

 

Send feedback to Richard Fobes at Email address, non-machine-readable to reduce spam
or use the Testimonials page

© Copyright 2010, Richard Fobes at VoteFair.org

Top of Page Top of Page