Survivor’s Race Problem: How the Odds are Stacked Against BIPOC Players

This post was written by Amanda Rabinowitz and me (Sean Falconer).

---

Murmurs that the game of Survivor is more difficult for non-White contestants have been with the show since its early seasons. Survivor’s first Black winner, S4’s Vecepia Towrey, discussed issues of race with fellow contestant Sean Rector in a scene that was groundbreaking television in 2002 and remains resonant today. Vecepia and Sean share concerns that their White castmates are suspicious of the friendship developing between the only two Black members of the cast.

Sean says in confessional,

Me and her are playing a whole other mental game that they don’t even know. That when you’re a person of color, the only one, you have to play. That’s something that they don’t have to worry about. Everybody can just be themselves. We have to be ourselves, but then hold back a little bit.

Photo: CBS

Sean’s insight suggests that Black players face an uphill battle in a social game in which they nearly always find themselves in the minority. His thesis is supported in the ensuing seasons. Vecepia has been joined by only 3 other Black winners, two Asian American winners, and only 1 Latinx winner (Sandra Diaz-Twine, who has won twice). Some among the fanbase have noted that the proportion of non-White winners is relatively small, and that Black, Indigenous, people of color (BIPOC) players are often targeted as early boots. RHAP’s Lita Brillman wrote a piece on the impact of race and gender in Survivor’s first 30 seasons, including player placement, which you can read here.

The show’s handling of race was brought to the fore in it’s 13th season, Survivor: Cook Islands, which was unfortunately themed as a “race war”, pitting four equal tribes of Black, White, Asian, and Latinx players against one another. This concept struck many viewers as racially-insensitive. But despite the questionable premise, S13 warranted positive attention for featuring the most diverse cast to that date, and along with its successor, S14 Survivor: Fiji, still represents the two most diverse Survivor casts by far (see Cast Diversity by Season image below).

Cast Diversity by Season

However, despite a low din of concern from small corners of the fanbase, the full extent to which the game Survivor is biased against BIPOC players has received relatively little attention.

Enter Spring 2020, amidst the early days of COVID quarantine and the immediate afterglow of the S40 WAW finale, the murder of George Floyd put police brutality against Black and Brown communities in the national spotlight. This tragic, and tragically common, incident spurred increased interest in the Black Lives Matter movement, national protests, and social justice campaigns. Among these is the Soul Survivor Diversity Campaign, a group of Black Survivor Alumni who have organized demanding that producers take action to improve diversity and inclusivity (sign their petition here). This campaign has garnered well-deserved national attention in major media outlets. Black Survivor alumni, podcasts, and other content platforms have participated in a number of one-on-one interviews and moderated panel discussions over the recent months focusing on the experiences of Black contestants on Survivor.

These discussions have amplified the voices of Black former contestants that were previously underrepresented in the Survivor commentariat. A theme that has emerged from these discussions is the same challenge, of “playing a whole other mental game,” voiced by Sean Rector in S4. The game within a game that Sean and others refer to references the notion that in addition to the challenges of navigating the social aspects of Survivor, Black players have an additional burden of navigating racial biases, prejudice, and racism that pervade the broader culture.

We decided to evaluate this issue empirically by examining the relationship between BIPOC status and performance in the game of Survivor across its 40 seasons. In this analysis, we operationalize performance in three ways:

1. overall placement (winner finishes is 1st place, runner up in 2nd, and so on),
2. likelihood of making the merge,
3. likelihood of making the final 5 (F5)

We also explore how gender may interact with BIPOC status to influence performance, and examine the role of overall cast diversity.

Recently Jeff Pittman over at True Dork Times conducted an excellent analysis of how race and ethnicity bear on likelihood of receiving votes for first boot, and we recommend you read more about that here.

BIPOC status and player performance

Average placement

One way to look at players’ performance is by comparing average overall placement for individuals from different groups. By this metric, BIPOC players appear to be at a slight disadvantage relative to White players. White players have competed a total of 536 times across all 40 seasons, compared to 194 times for BIPOC players.

Across these outings, the average White player has finished at final 9, whereas the average BIPOC player has been voted out one spot earlier at final 10. This one place difference may not seem like a big deal, but it is statistically significant (t=-2.30; p = 0.02), suggesting that it is unlikely that the discrepancy in performance is due to chance. In fact, in a model using demographic characteristics (age, gender, and BIPOC status) to predict overall placement, BIPOC status was the only characteristic with a significant relationship to game performance.

Not only do BIPOC players on average underperform relative to White players, but this pattern appears consistent across seasons. In 30 of 40 seasons, or 75%, the White players on the cast had a better average finish than the BIPOC players. If race were independent of placement, we would expect this to be much closer to 50%. Perhaps even more troubling is that 8 of the 10 seasons in which BIPOC players finished better than their White counterparts took place in the first half of show’s run, implying that the situation for BIPOC players is getting worse, not better.

Because distinct racial and ethnic groups have different cultural and social experiences which may impact Survivor performance in different ways, we also examined these groups separately. Although Black, Asian, and Latinx players all placed more poorly than Whites, there was not much difference in average player placement across the different categories. All BIPOC groups had average placements between 10th and 11th place (Black=10.3, Asian=10.7, Latinx=10.4).

Role of gender

Interestingly, the racial discrepancy in overall placement appears to be driven predominately by a disadvantage for BIPOC women. We compared performance of White and BIPOC players within each gender separately. White men have competed a total of 282 times across 40 seasons, compared to 82 times for BIPOC men; White women have competed a total of 254 times across 40 seasons, compared to 111 BIPOC women.

Across these outings, BIPOC men on average are eliminated a half a place earlier (poorer placement) than their White male counterparts, a difference which is not statistically significant (t=-0.83; p=0.41). However, White women on average finished at least 1 place better than their BIPOC counterparts, which represents a statistically significant advantage for White female players (t=-2.02; p=0.04).

Average Placement by Race and Gender

Making the merge

To explore performance differences further, we also looked at the likelihood of making the merge for White versus BIPOC players. Assuming all things are equal, the diversity of the cast should be the same both pre and post merge. Using this assumption, we calculated the actual versus expected diversity of the cast at merge for each season by taking the difference between the actual and expected values. A negative number means the actual was worse than the statistically expected value.

There were 15 seasons where the actual level of diversity was lower post merge than the expected value, 19 seasons where the actual was better than the expected, leaving 6 seasons where everything worked out as expected (see image below).

Good news; this finding suggests that BIPOC players on average are doing as well or better than we would expect in terms of making the merge. However, digging deeper into these seasons, we see that the average percentage of BIPOC cast in the 15 seasons where BIPOC players underperform is much smaller than the 19 seasons where the actual is better than the expected value, 22.8% versus 29.5% respectively. This implies that a more diverse cast helps level the playing field for BIPOC players.

BIPOC Merge Differential (Actual – Expected)

Making the F5

We applied a similar analysis to making it to the F5 of a season. Unlike the merge results, we find a disadvantage for BIPOC players in making the F5 for the majority of seasons— in 24 seasons (or 60%) the actual number of BIPOC players making F5 is worse than the expected value, compared to 16 seasons for which the actual is better than expected (see image below).

Similar to the merge results, in seasons where BIPOC players outperform the statistical expectation, the cast is more diverse. In the 24 seasons where the performance is worse than expectations, on average 24.4% of the cast consists of BIPOC players, while in the 16 seasons with better performance, the cast is made up of 28.7% BIPOC players.

BIPOC Top 5 Differential (Actual – Expected)

The results are most troubling for BIPOC women. With regards to making the F5, only 5 times in the show’s history have the BIPOC women outperformed their expected value (see image below). Interestingly, 2 of these 5 seasons represent the performance of one extraordinary woman, Cirie Fields, who is considered by many to be one of the very best to ever play.

This result is striking. If race and gender had no bearing on making F5, we would expect BIPOC women to meet or exceed the expected value in close to 20 of the show’s 40 seasons.

BIPOC Women Top 5 Differential (Actual – Expected)

Photo: CBS

Putting it all together

We examined the question of whether CBS’s Survivor, as a game, is more difficult for BIPOC players. While there are a variety of ways to tackle this complex question, as an initial foray into the topic we chose to evaluate the relationship between BIPOC status and player performance.

We found that BIPOC players performed more poorly than White players according to average placement and likelihood of making the F5. The deficit in BIPOC performance appears to be largely driven by a striking disadvantage for BIPOC women, who exhibit the poorest average placement, and are underrepresented in the end game in over 87% (!!) of seasons. The bias against BIPOC players cannot be dismissed as a relic of the show’s early years. On the contrary, our results suggest that the situation for BIPOC players has gotten worse in recent seasons, not better. One encouraging conclusion that can be inferred from these findings, however, is that greater diversity of the cast appears to reduce, or even eliminate, some of the disadvantages faced by BIPOC players. When BIPOC representation in a given cast approached 30%, BIPOC players tended to make the merge and F5 at or exceeding their expected numbers.

It is likely that multiple factors contribute to poorer performance for BIPOC players. Survivor is a social game, and decades of research on social behavior have demonstrated that categorizing individuals by race and ethnicity influences our behavior towards them. The forces that drive elimination of BIPOC players from the game probably run the gamut from overt racism on the part of some players (which may be relatively rare, but not absent) to more subtle forms of bias (which is undoubtedly more common, if not ubiquitous). It has been noted that as soon as the game begins, players start to form bonds based on similarity. Those bonds become alliances that carry a player further in the game. Contestants who find themselves on a starting tribe as the lone representative or their race, or with one other member of their group, face an uphill battle when it comes to making those early connections.

Another challenge facing BIPOC players is the pressure to serve as good representatives/ambassadors of their demographic, both to their fellow castmates and also the much larger television audience. One can imagine that this concern would make a player more socially inhibited as they get to know their fellow tribemates and learn how accepting or unaccepting the group may be of people who look like them. This is consistent with Sean Rector’s observations about the distinct challenges he and Vecepia faced in Marqueses, “Everybody can just be themselves. We have to be ourselves, but then hold back a little bit.”

The pressure of serving as a lone representative of a group may even hamper the sense of freedom and playfulness required to explore the more strategic and devious aspects of gameplay— styles of play that are applauded in beloved players such as Tony Vlachos and Boston Rob. What this amounts to is a game with a greater level of difficulty for BIPOC players. We have previously written about challenges facing female players, and unsurprisingly, our results here reveal a compound disadvantage for BIPOC women— who face intersecting challenges related to both gender and race.

The conclusion that racial underrepresentation within a cast contributes to individual players’ performance is supported by our finding that BIPOC players were more likely to maintain their proportion through the merge and F5 when the cast was more diverse. Interestingly, the average proportion of BIPOC players for seasons on which BIPOC players were most likely to last through merge and F5, was 29.5% and 28.7% respectively. These numbers are very close to the 30% threshold for BIPOC casting that is being called for by the Soul Survivor Diversity Campaign’s petition.

Some have defended historical casting practices by noting that only 13% of Americans identify as Black, and Survivor should only be expected to cast BIPOC players according to their proportions within the US population. This amounts to between 2 and 3 Black members of every cast, ensuring that Black contestants can expect to be either the lone representative of their race on their starting tribe, or paired with exactly one other fellow Black castmate.

The truth is that Survivor casting has the leeway to explore more interesting and socially responsible casting approaches. Imagine if Survivor casting was beholden to matching US demographics on attributes such as age, profession, or body mass index. Of course casting doesn’t restrict their decisions based on those criteria, nor should they. The casting department’s primary function is to assemble a group of players who will produce a good TV product. We posit that CBS’s notion of what makes “good TV” is overwhelmingly influenced by a White perspective. We agree with the Soul Survivor group that Survivor casting should play an active role in expanding the audience’s familiarity and comfort level with characters and stories that represent diverse identities and experiences. This starts with casting more BIPOC players. Encouragingly, our findings suggest that this small step could result in more BIPOC players persisting into the late stages of the game, giving production an opportunity to tell their stories with more nuance and in richer detail.

We’d also like to note that increasing BIPOC casting alone may not be enough to level the playing field for BIPOC players. It has been reported by former Survivor players that the representation of BIPOC “behind the camera” is also poor. Production staff make decisions that shape players’ experiences in innumerable subtle ways that may influence game performance. Even well-meaning members of the production team who don’t share the same experiences and background as BIPOC members can act in ways that unwittingly disadvantage or burden these players. For example, choices ranging from the interview questions asked in confessional to the types of beauty and personal care products that are provided on a spa reward may constitute microaggressions that place additional stress on BIPOC players.

It is also possible that the mechanics of the game itself could result in a disadvantage for certain players. We explored the relationship between idol finding and gender in a previous article. Others have examined how jury size and final TC format may also disadvantage certain players. Exploring these game mechanics in relationship to BIPOC status is an interesting topic for future investigation.

Here we’ve presented empirical support for the notion that BIPOC players are disadvantaged in the CBS version of the game of Survivor. In light of the evidence of bias against BIPOC players, we would like to amplify the calls from the Soul Survivor campaign to increase casting of BIPOC players to 30% or greater, and to increase the representation of BIPOC staff in key production roles. Importantly, the recognition of bias does not necessarily indicate conscious malice on the part of any one individual or group of individuals. It is likely that the discrepancies we observe largely arise from unconscious biases within individuals and systems that unknowingly cause and maintain these disadvantages.

As a final note, although we examined BIPOC players as a group encompassing Black, Latinx, and Asian American contestants, most of the focus of our discussion has centered on the experiences of Black contestants— in part because they make up the largest proportion of BIPOC Survivor alumni, but also because Black experiences are of particular interest and importance right now in light of current events. With that said, we want to highlight the need to further explore, with greater granularity, how identity impacts Survivor performance for contestants from all different backgrounds.

Do you have thoughts on how Survivor could improve diversity and inclusivity? Please let us know in the comments or on Twitter @arabinow and @seanfalconer.