Jeff Pitman's S40: Winners at War recaps
Survivor and the problem with the first vote
By Jeff Pitman | Published: July 31, 2020
Survivor multi-season analysis

Survivor and the problem with the first vote


Survivor has made strides in recent seasons in broadening its roster of contestants, adding larger numbers of people from different cultural and ethnic backgrounds. Yet at the same time, as we'll see below, those contestants face tougher odds than their White counterparts when it comes time for their tribe to vote for the first time. The show also remains largely White and male behind the camera, in its production crew and editing teams, and while not specifically connected to this problem (in theory, but maybe?), that also is a situation that's past due for change.


While the show appears sincere in its willingness to cast more Black, Indigenous and People of Color (BIPOC) contestants, having their stories told by the same set of mostly White men that excised Julia Carter almost completely from Edge of Extinction, even with the best of intentions, could well lead to more oversights, elisions, and erasure that could otherwise be avoided by bringing a greater diversity of voices to the story-crafting team.


In that light, Survivor's pandemic-induced hiatus is an excellent opportunity for the show to hire additional editors, producers, and other crew that more closely reflect the diversity of the people in front of the camera. So please, read and consider signing the Petition for Anti-Racism Action by Survivor Entertainment Group.


And who knows, maybe having a more diverse set of people filming them will help future contestants avoid the main problem that we're here to talk about:


The early boot/ early vote problem

The early boot/ early vote problem


Fans and contestants have noticed for a while that when it comes time for a tribe to vote the first time, BIPOC contestants almost always seem to end up at the top of the target list. This is not necessarily due to overt racism, but it may be a symptom of unconscious bias, especially if the tribe has mostly White contestants, and just one or a few contestants of color. An all-newbie season's tribes are groups of strangers, and thus early Survivor votes are mostly based on first impressions, and/or implicit biases/lack of familiarity people bring into the game with them.


It's a Survivor truism that at the first few votes a tribe makes, you never want to stand out: You don't want to be the leader, you don't want to be the laziest, the bossiest, the weakest, or the most annoying. But if you're the only person of color on a tribe of otherwise all White people, you'll definitely stand out, and for a reason you can't do anything about: something people can see just by looking at you, something you can't change just by being more (or less) laid back, working harder, or putting in extra effort to get along.


So when that first vote comes around, who's getting targeted? People who stand out. Does that group almost always include BIPOC contestants? (Spoiler: Yup. But please keep reading anyway.)


One way to look at this is to just go through the early boots of recent seasons: In Millennials vs. Gen X, which was an otherwise dynamic and entertaining season, all but one of the pre-merge boots (including the first one out of each starting tribe) were women of color: Rachel Ako, Mari Takahashi, Lucy Huang, CeCe Taylor, Figgy Figueroa, and Michaela Bradshaw.


"But wait!" you may object. Since that season had a lot of BIPOC contestants overall, aren't they just more likely to be early boots, because there were more of them? That's a fair point, and so to look at how often BIPOC contestants are targeted, we need a systematic way to measure their being voted against, independent of their level of representation:


Enter EVAB, or Excess Votes Against BIPOC contestants. How do we measure that? For any particular Tribal Council, we count the total number of votes cast, the % of those votes that were against Black, Indigenous, of People of color contestants, and the overall % of BIPOC contestants casting votes. If votes were completely random, you would expect that eventually, over a large sample size/ many Tribal Councils, the % of votes cast against BIPOC contestants and the % of votes cast by BIPOC contestants would be roughly even. If they face the same hurdles as any other player, those two percentages should be the same.


Here's an example. Let's say there are three Black contestants on a starting tribe of 10. They make up 30% of the tribe, and you would expect that if this group voted say, 100 times, they should pick up around 30% of the votes, right? So if over those 100 votes, they actually receive 40% of the votes cast, that's worrisome. In contrast, something other than racial bias is probably going on if they only receive 20% of the votes cast. In the case of the 40% of votes received, that's 10% higher than expected (40% - 30%), so the EVAB score would be +10%. In the 20% case, that's 10% lower than expected (20% - 30%), so the EVAB score would be -10% (negative scores are, paradoxically, a positive thing).


(Note: Obviously, this doesn't work at the extremes. An all-White tribe like Blue Collar in Worlds Apart has no ability to vote against a Black player. Conversely, an all-Black tribe like Hiki in Cook Islands has no choice but to vote out a Black contestant. So this analysis only contains tribes where there was a choice to vote for at least one BIPOC contestant and at least one White contestant.)


So anyway, let's take a look at this, breaking Survivor's 40-season run into groups of 10 seasons at a time.


In the beginning...

In the beginning...


We start at the very beginning, Borneo (season 1) and extend through Palau (season 10). It was the best of times (as you'll see), it was the worst of times. Worst because in the early days, overall diversity was pretty low, and the rare Black contestants who were present often received crappy edits. In The Australian Outback, for example, Nick Brown was an Army officer entering Harvard Law. But you didn't hear about that. Instead you saw him sitting around, relaxing, or making furniture in which to sit around and relax. Despite clearly being an accomplished, high-achieving young man, he was "lazy." The next season (Africa), Clarence Black was presented as "selfish" and "untrustworthy" for taking too many beans. (At least according to his redneck tribemate Big Tom Buchanan, who had three times as many confessionals as Clarence during that episode.) The edit choices didn't turn out much better for JoAnna Ward in The Amazon, or Osten Taylor in Pearl Islands, or Rory Freeman in Vanuatu. The Black contestants seemed to almost exclusively be shown as creating conflict. Furthermore, overall BIPOC representation remained seemingly hard-capped at no more than three contestants per season (one Black man, one Black woman, then maybe one more slot for an Asian-American or Latinx contestant). Even as overall cast size swole from the original 16 to 18 to 20.


But it wasn't all bad. (See Figure 1.)


Seasons 1-10, excess votes against BIPOC contestants

Figure 1: Excess votes against BIPOC contestants (EVAB), seaons 1 through 10. Percentages show difference between percent of votes cast against contestants who Black, Indigenous, or People of color (at each tribe's first Tribal Council) minus the percent of votes by those same contestants.


Despite the show's chosen narrative that they were mostly pot-stirring troublemakers, over this first span of 10 seasons, BIPOC contestants on average received fewer votes against them (at the first vote) than would be expected by chance. This was particularly true at the first Tribals in Seasons 4 and 7 — Marquesas and Pearl Islands, which had a Black winner in Vecepia Towery and a Latina winner in Sandra Diaz-Twine, respectively. That's good! And in fact, despite the targeting of Kel Gleason in The Australian Outback (Survivor's first Native contestant) and of Jolanda Jones in Palau, as a whole, contestants of color more or less broke even on these first votes (see the "1->10" bar on the right-hand side). So, editing aside, it was a relatively equitable time to be a contestant of color (at least at the first vote). Out of ten seasons, only three have EVAB scores above zero.


Seasons 11-20: Change is afoot, but still (mostly) smooth sailing


The next 10 seasons brought on an era of big structural changes: Maya temples as camps and Tribal Council sets! Hidden immunity idols! Four starting tribes! Four tribes divided by ethnicity! The same format, but switched back to two tribes at the last minute because a White woman dropped out! Mixed returnees and newbies! Thankfully, despite all the big swings production was taking, for BIPOC contestants facing their tribe's first vote, it still was relatively non-terrible, with one glaring exception (Figure 2).


Seasons 11-20, excess votes against BIPOC contestants

Figure 2: Excess votes against BIPOC contestants (EVAB), seasons 11 through 20. Same methodology as Fig. 1, except for seasons 12 and 13 (Panama-Exile Island and Cook Islands), post-swap tribes were included.


So yeah, season 18, Tocantins — again, otherwise a solid season with some huge, memorable characters — wasn't so pleasant for its Black and Latina contestants in the early going. (Note the much larger y-axis scale in Fig. 2 relative to Fig. 1.) By episode 4, three votes had been held, and all but one of the non-White contestants were gone. (Although to be clear, only the first votes made by Jalapão and Timbira — booting Carolina Eastwood and Candace Smith — are reflected in the graph above.) Still, six of these 10 seasons were even or below zero for EVAB, so with that major Tocantins exception, things weren't terrible. Also, note that season 13, Cook Islands, despite the most diverse (to that point) cast in Survivor history, still ended up suggesting some level of bias.


Brief methodology aside: We should also note that Panama (season 12) and Cook Islands (season 13) both started off with four tiny tribes, then swapped early to form two large tribes. Because of that, this analysis also included those post-swap tribes, since the original tribes were so brief as to be meaningless. (Where possible, at least: post-swap La Mina was all-White, so it was skipped. Also, the Ep1 Hiki and Ep2 Aitu votes in CI were omitted, since the only choices to vote against were people of color.) For Cook Islands, all but one of the original Raro (White tribe) members made the jury, and the other was the final pre-jury boot. Meanwhile, every other original tribe lost at least two members pre-merge, and the Latinx Aitus lost four. It was much less of an issue the next season when the even more diverse Fiji cast started off as two self-selected tribes, and weren't initially pitched as "the battle of the races" or whatever.


Cumulatively, only six out of the first 20 seasons had EVAB scores > 0, and the net score for seasons 1-20 (and 11-20) was only barely above zero. Not a bad start! Then everything went to hell.


Seasons 21-30: Everything goes to hell

The early boot/ early vote problem

Seasons 21-30, excess votes against BIPOC contestants

Figure 3: Excess votes against BIPOC contestants (EVAB), seasons 21 through 30. For details on methodology, see Figure 1.


If you're looking for a silver lining here, maybe it's that Nicaragua (season 21) and One World (season 24), two of the least-beloved seasons in Survivor history, at least weren't horrible in this particular regard? Also, the scale of the positive EVAB scores in seasons 27-30 isn't *that* bad. Picking up 10% more votes than expected on a 10-person tribe is just one extra vote. Still not great, but not the end of the world. Seasons 22-26, though (One World excepted, again) ... whew. That's a bad look. (No, it does not help that a good portion of those unsightly stats is the double-first-booting of Francesca Hogi, who did nothing to deserve it.)


It's most likely a coincidence (Right? Because it also marked the fading away of noted Trump enabler and chief "Black men are lazy" editing trope enthusiast Mark Burnett?) but this coincides with when Jeff Probst took over as showrunner. We're not trying to pin this on Probst, anyway. Regardless, here we see the switchover from the vast majority of the seasons in the first half of the show's run having a negative EVAB score to the opposite: eight out these 10 clocking in at greater than zero. That results in a 13% cumulative EVAB for this span, which is not great. But hold on, because it gets worse.


Seasons 31-40: It gets worse. But maybe slightly better, eventually?

Seasons 31-40, excess votes against BIPOC contestants

Figure 4: Excess votes against BIPOC contestants (EVAB), seasons 31 through 40. For details on methodology, see Figure 1. Note: Post-swap Angkor's and Bayon's first votes were included for Cambodia, because of (1) early swap, and (2) an all-returnee season probably has less "first impression" bias anyway.


From two out of 10 at or below zero EVAB to ... one. Honestly, this project really is not an attempt to rehabilitate the reputations of all your most-hated seasons, but still, welcome to the tiny group of the past decade's seasons that are not racially problematic right out of the gate, Game Changers (season 34)! Conversely, remember how the entire set of pre-merge boots in Millennials vs. Gen X raised some uncomfortable questions? Well, unsurprisingly, the Millennials and Gen X tribes' first votes were a part of that. Ghost Island probably looks worse than it should here, but the Malolo tribe's first vote was the only one counted, because there was a 2-to-2 tribe swap at 18, and the original tribe lines were clearly the most important thing after that, so original Naviti never had a first vote. That makes Malolo's near-unanimous boot of Stephanie Gonzalez bear all the weight of the season. At least it had a Black winner, though?


This observation is not intended to assign blame to anyone, rather, it's to document a problem, and hopefully suggest solutions. An obvious one is suggested in the "Taking steps forward" section below. Until then, here's a positive spin on the numbers: The last two all-newbie seasons weren't that bad, EVAB-wise, at +10% for Edge of Extinction (season 38) and +5% for Island of the Idols (season 39). That's more clear when you look at all 40 seasons on a single graph:


Seasons 1-40, excess votes against BIPOC contestants

Figure 5: Excess votes against BIPOC contestants (EVAB), all seasons. For details on methodology, see Figure 1.


In summary, BIPOC contestants have clearly faced tougher obstacles than their white counterparts at all but one of the last 10 seasons' first Tribal Councils, and all but two of the last 19 seasons' first Tribals. On average in the S31-40 era, a BIPOC contestant would have picked up two extra votes (for a 10-person tribe) more than a White tribemate at that first Tribal. And this phenomenon has gradually gotten worse over time, going from a relatively even-handed start (-1% EVAB in S1-10) to not great (6% EVAB in S11-20) to its worrisome current state (13% EVAB for S21-30, 20% EVAB for S31-40).


Does starting tribe size matter?

Does starting tribe size matter?


Short answer: Maybe, but there's not a linear relationship between initial tribe size and targeting of BIPOC contestants. What is clear is that smaller tribes (three tribes of six, for example) are indeed the worst situation, as Brice Izyah experienced in Cagayan. Perhaps, as Rob Cesternino has often observed, that's because there's nowhere to hide on a six-person tribe. If you're a BIPOC contestant, you're (on average) starting off with at least 1 vote by default, so it doesn't take many more to reach a majority.


Here are the stats for tribe size. Note: For this, we included the first two votes each tribe made, to bump the numbers for each group up to a more meaningful sample size. (This where 7-person tribes come from.) We also tracked what percentage of the votes resulted in BIPOC contestants being booted (Boot%) and compared that to overall representation (BIPOC%) to get the EVAB-like Boot % - BIPOC%, which tracks increased risk of being voted out, as a second measure of being targeted.


# of people EVAB # of Tribals BIPOC % Boot % Boot % - BIPOC%
8 4.4% 20 27.5% 30.0% 2.5%
10 10.5% 19 25.3% 31.6% 6.3%
7 12.2% 21 25.9% 38.1% 12.2%
9 14.8% 18 32.7% 44.4% 11.7%
6 15.4% 13 32.1% 53.8% 21.8%


The results: Eight-person starting tribes are pretty rare nowadays, so while the best tribe size appears to be eight people (4% EVAB, and just a 2.5% increased risk of being booted), it's impossible to know whether that's because most 8-person tribes were on the early seasons when BIPOC contestants were not frequent early targets (Fig.1), vs. there actually being something special about there being exactly eight people. Both are possible interpretations.


Six-person tribes are clearly the worst situation for BIPOC players, though, with a sizeable 15.4% EVAB, and a 21.8% increased risk of being booted. Happily, Survivor might finally be tiring of having three starting tribes (of six), so maybe that's less of a problem going forward. Although if someone could advocate for that position in production, that would help.


Taking steps forward

Taking steps forward


Having more BIPOC contestants per season doesn't necessarily remove unconscious bias when it comes to first votes, as Cook Islands and Island of the Idols demonstrated. But it's still the best way to address the problem. Why? Easy: Because if you cast more contestants of color, you can't vote them all out at the first Tribal Council. Duh. Alternatively (back in the old model of just two Black contestants per season): Having 12 episodes in a 14-episode season where everyone is 100% White is unlikely to broaden the show's appeal much beyond its existing, mostly-White fanbase. And Survivor does want to expand its audience. (What show doesn't?)


The Island of the Idols casting model was a solid step forward in that regard. Two women of color ended up in the final five. Two White men still split all the jury votes, which is a different recent problem that also needs to be addressed, but is outside the scope of this analysis. (It obviously had another, much larger problem on its hands with the Dan Spilo situation, but racial diversity was clearly not at fault there.) But Island of the Idols' diversity was overall a good thing.


That's because, for a show soon to enter its 41st season, that diversity offers it one huge keep-the-audience-interested advantage that doesn't require convoluted twists or production machinations to bring about: Casting people that don't look like every other player, people that have different backgrounds than their tribemates and/or previous contestants, allows Survivor to tell new stories.


Island of the Idols featured Survivor's first contestant from the Hmong community in Vince Moua, and its first Indian-American contestant in Karishma Patel. Both were given the chance to talk about how they have had to balance traditional expectations from their communities and their personal interests, how those sometimes conflict, and how that's preparation for the game of Survivor. They both talked about struggling to fit in. They both talked about expectations and achievements. These are universal stories, accessible to everyone, and yet they're interesting to the broader audience because they offer a new perspective. More of this, please.


That's how the show can adapt and prosper over 40 more seasons: by casting a wider variety of Americans, and simply letting them tell their stories. We've seen the story of the retired White male athlete many, many times. Do we *really* need to keep seeing that almost every season? It should be noted that Island of the Idols also featured great, previously unheard stories from some of its White contestants, particularly Janet Carbin and Elaine Stott. The important thing is: keep telling fresh stories.


Casting a wider net to bring in new and different kinds of contestants is only part of the battle to ensure this approach continues, though. As many of the Black former contestants mentioned on the RHAP panel, having almost exclusively White (and usually male) producers and other crew members recording and editing those stories isn't the most welcoming, conducive environment in the first place, and can lead to mistrust and/or miscommunication between contestants and production.


Despite what it looks like on camera, confessionals are a group activity, not the result of a contestant talking to a camera on a tripod, isolated in the jungle somewhere. There's a producer there asking questions, and a camera crew recording it. As Kellee Kim's situation in Island of the Idols showed, it's not always easy for a contestant to say what they want to say, or what needs to be said, when the person asking the question is a White man. Sometimes it helps to talk to someone who at least has something in common with you. As the Black contestants on the RHAP panel mentioned, sometimes the producers ask leading questions, trying to get the contestant to support a particular narrative in their answers. When the later editing and shaping of those stories is also carried out almost exclusively by White men, that's another potential area for unintentional bias to take over, and for disparate voices to be silenced. (See, again, Julia in Edge of Extinction.)


Having a greater array of contestant stories to tell, as in Island of the Idols, is a great first step. Having a broader variety of behind-the-camera personnel, a spread that more closely reflects the diversity in front of the camera, is also essential for ensuring those stories are told accurately and fairly. (For additonal stories about why this is so important, please watch/listen to RHAP's panel on LGBTQ intersectionality. That and this article are not intended as an attack of Survivor, just a plea for change.) Survivor has hinted it wants to take steps towards this end, but it may need a nudge here or there to bring that to fruition.


So to help that along, please read and consider signing the Petition for Anti-Racism Action by Survivor Entertainment Group.


Jeff Pitman's recapsJeff Pitman is the founder of the True Dork Times, and probably should find better things to write about than Survivor. So far he hasn't, though. He's also responsible for the Survivometer, calendar, boxscores, and contestant pages, so if you want to complain about those, do so in the comments, or on twitter: @truedorktimes

Other Survivor diversity-related writing, podcasts, and resources

  • J'Tia Hart's petition: "Petition for Anti-Racism Action by Survivor Entertainment Group"