This README file was generated on 2026-02-02 by Erik Nelson.

GENERAL INFORMATION

Title of Dataset: Data from: Does Beauty Pay? An Analysis of College Athlete Name, Image, and Likeness (NIL) Pricing.

Author Information

Name: Erik Nelson
Institution: Bowdoin College
Address: Brunswick, ME, USA
Email: enelson2@bowdoin.edu

SHARING/ACCESS INFORMATION

Licenses/restrictions placed on the data: CC0 1.0 Universal (CC0 1.0) Public Domain.

Publications that use the data:

McCarter, Jane, and Erik Nelson. (2026). Does Beauty Pay? An Analysis of College Athlete Name, Image, and Likeness (NIL) Pricing.

Recommended citation for this dataset:

McCarter, Jane, and Erik Nelson. (2026). Data and code for "Does Beauty Pay? An Analysis of College Athlete Name, Image, and Likeness (NIL) Pricing."

DATA & FILE OVERVIEW

The data and code in this repository have been used in the paper "Does Beauty Pay? An Analysis of College Athlete Name, Image, and Likeness (NIL) Pricing." The paper includes 8 tables, 3 figures, 26 supplementary material tables, and 5 supplementary figures. All of these tables and figures were created using the datasets and code in this repository. The datasets are in .csv format, and the code was written in R or Stata.

There are five.csv datasets:
A) AthleteDataforDryad.csv
B) AthleteDatawoNAsforDryad.csv (same as AthleteDataforDryad.csv but NAs have been replaced with blanks).
C) RaterDataforDryad.csv
D) RaterDatawoNAsforDryad.csv (same as RaterDataforDryad.csv but NAs have been replaced with blanks).
E) RespondentInfoforDryad.csv.

CODE/SOFTWARE/DATA

The best way to understand how R script is used in our project is to list the scripts that are used to create each figure or table in our paper.

• Figure 1 uses Figure1.R and AthleteDataforDryad.csv. Figure 1 presents histograms of NIL product prices as of 9/9/24 (US dollars) by athlete class. The top panel is comprised of price histograms for Superstar athletes. The bottom panel is comprised of price histograms for Conventional athletes.
• Figure 2 uses Figure2SuperstarConventional.R, Figure2ConventionalGender.R, and Figure2ConventionalRace.R. Figure 2 visualizes changes in mean NIL prices (US dollars) over time. Athletes can change the price they charge for NIL products on their Opendorse profile. We were able to observe some athletes’ profiles on 7 unique dates, others on 6 unique dates, etc. At the extreme, there were a number of athletes where we only observed their profile once. Each time we observed an athlete’s profile we recorded their set of published NIL prices. Then for each group of athletes, grouped by the number of times we observed their profile, we calculated and plotted the mean of their first observed price (gray circles) and the mean of their last observed price (black circles) (there is only one circle, a blue circle, for mean price charged by athletes with only one profile observation).The color and direction of an arrow indicates whether there was an increase or decrease in mean price for a given group.
• Figure 3 uses Figure3AllAthletes.R, Figure3SuperstarAthletes.R, Figure3ConventionalAthletes.R, Figure3ConventionalMaleAthletes.R, Figure3ConventionalFemaleAthletes.R, Figure3ConventionalWhiteAthletes.R, Figure3ConventionalNonWhiteAthletes.R, and RaterDatawoNAsforDryad.csv. Figure 3 presents NIL prices (US dollars) versus attractiveness score bin scatter plots (an attractiveness score ranges from 0 to 10). The y-axis gives published NIL prices as of 9/9/24. Each row of plots except the top row pertains just to the given subset of athletes. The plots give the mean, median, 90th percentile, and 10th percentile NIL prices for a given NIL product at each attractiveness score bin. An athlete is represented multiple times in each plot as each athlete has multiple attractiveness scores. For example, assume the fictitious Conventional white athlete Jane Smith lists an autograph price of $50 and has four attractiveness scores of 9, eight scores of 7, and three scores of 5. Her $50 will contribute to the 9-10 bin descriptive statistics four times, the 7 bin descriptive statistics eight times, and the 5 bin descriptive statistics three times in the Autograph plots in the Conventional, Conventional female, and Conventional white rows.
• Table 1 uses Table1.R and AthleteDataforDryad.csv.
• Tables 2 and 3 uses Tables2and3.R and AthleteDataforDryad.csv.
• Tables 4 and 5 uses Tables4and5AttractivenessMeans.R, Table4AthleticMeans.R, Table4TrustworthinessMeans.R, Table4FamiliarityMeans.R, Table4FamiliarityMeansVer2.R and AthleteDataforDryad.csv.
• Table 6 uses Table6.do and AthleteDatawoNAsforDryad.csv.
• Table 7 uses Table7.do and AthleteDatawoNAsforDryad.csv.
• Table 8 uses Table8.do and RaterDatawoNAsforDryad.csv.
• SM Table 2 uses SMTable2.R and RespondentInfoforDryad.csv.
• SM Table 3 uses SMTable3PartOne.R, SMTable3PartTwo.R, RaterDataforDryad.csv, and RespondentInfoforDryad.csv.
• SM Table 5 uses SMTable5.R and AthleteDataforDryad.csv.
• SM Table 6 uses SMTable6.R and AthleteDataforDryad.csv.
• SM Table 7 uses SMTables7and8PartOne.R, SMTables7and8PartTwo.R, and AthleteDataforDryad.csv.
• SM Table 8 uses SMTables7and8PartOne.R, SMTables7and8PartTwo.R, and AthleteDataforDryad.csv.
• SM Figure 2 uses SMFigure2Attractiveness.R, SMFigure2Athleticism.R, SMFigure2Trustworthiness.R, SMFigure2Familiarity.R, and RaterDataforDryad.csv. SM Figure 2 uses box and whisker plots to show the distribution of all attractiveness, athletic, trustworthy, and familiarity scores (all on a 0 to 10 scale) by athlete status and athlete characteristics. Median is given by red bar, mean by black diamond. The top of the box is the 75th percentile, the bottom of the box is the 25th percentile. The whiskers indicate minimum and maximum values. The plots do not indicate outliers. All athletes have multiple scores for each attribute (recall each athlete is rated by multiple survey respondents).
• SM Table 9 and 10 uses SMTables9and10.R and RaterDatawoNAsforDryad.csv.
• SM Table 11 uses Figure3AllAthletes.R, Figure3SuperstarAthletes.R, Figure3ConventionalAthletes.R, Figure3ConventionalMaleAthletes.R, Figure3ConventionalFemaleAthletes.R, Figure3ConventionalWhiteAthletes.R, Figure3ConventionalNonWhiteAthletes.R, and RaterDatawoNAsforDryad.csv.
• SMFigure 3 and SM Table 12 uses SMFigure3AllAthletes.R, SMFigure3andSMTable12SuperstarAthletes.R, SMFigure3andSMTable12ConventionalAthletes.R, SMFigure3andSMTable12ConventionalMaleAthletes.R, SMFigure3andSMTable12ConventionalFemaleAthletes.R, SMFigure3andSMTable12ConventionalWhiteAthletes.R, SMFigure3andSMTable12ConventionalNonWhiteAthletes.R, SMFigure3andSMTable12AllAthletes.R, and RaterDatawoNAsforDryad.csv. SM Figure 3 presents NIL product prices (US dollars) versus athleticism scores bin scatter plots (all athleticism scores are on a 0 to 10 scale). The y-axis gives published NIL prices as of 9/9/24. Each row of plots except the top row pertains just to the given subset of athletes. The plots give the mean, median, 90th percentile, and 10th percentile NIL prices for a given NIL product at each athleticism score bin. An athlete is represented multiple times in each plot as each athlete has multiple athleticism scores.
• SMFigure 4 and SM Table 13 uses SMFigure4andSMTable13AllAthletes.R, SMFigure4andSMTable13SuperstarAthletes.R, SMFigure4andSMTable13ConventionalAthletes.R, SMFigure4andSMTable13ConventionalMaleAthletes.R, SMFigure4andSMTable13ConventionalFemaleAthletes.R, SMFigure4andSMTable13ConventionalWhiteAthletes.R, SMFigure4andSMTable13ConventionalNonWhiteAthletes.R, and RaterDatawoNAsforDryad.csv. SM Figure 4 presents NIL product prices (US dollars) versus trustworthiness scores bin scatter plots (all trustworthiness scores are on a 0 to 10 scale). The y-axis gives published NIL prices as of 9/9/24. Each row of plots except the top row pertains just to the given subset of athletes. The plots give the mean, median, 90th percentile, and 10th percentile NIL prices for a given NIL product at each trustworthiness score bin. An athlete is represented multiple times in each plot as each athlete has multiple trustworthiness scores.
• SMFigure 5 and SM Table 14 uses SMFigure5andSMTable14AllAthletes.R, SMFigure5andSMTable14SuperstarAthletes.R, SMFigure5andSMTable14ConventionalAthletes.R, SMFigure5andSMTable14ConventionalMaleAthletes.R, SMFigure5andSMTable14ConventionalFemaleAthletes.R, SMFigure5andSMTable14ConventionalWhiteAthletes.R, SMFigure5andSMTable14ConventionalNonWhiteAthletes.R, and RaterDatawoNAsforDryad.csv. SM Figure 5 presents NIL product prices (US dollars) versus familiarity scores bin scatter plots (all familiarity scores are on a 0 to 10 scale). The y-axis gives published NIL prices as of 9/9/24. Each row of plots except the top row pertains just to the given subset of athletes. The plots give the mean, median, 90th percentile, and 10th percentile NIL prices for a given NIL product at each familiarity score bin. An athlete is represented multiple times in each plot as each athlete has multiple familiarity scores.
• SM Table 15 uses Table6.do and AthleteDatawoNAsforDryad.csv.
• SM Table 17 uses Table7.do and AthleteDatawoNAsforDryad.csv.
• SM Table 20 uses Table8.do and RaterDatawoNAsforDryad.csv.
• SM Table 21 and SM Table 22 uses SMTables21and22.do and AthleteDatawoNAsforDryad.csv.
• SM Table 23 and SM Table 24 uses SMTables23and24.do and AthleteDatawoNAsforDryad.csv.
• SM Table 25 uses SMTable25SuperstarBowdoin.do, SMTable25SuperstarProlific.do, SMTables25and26ConventionalBowdoin.do, SMTables25and26ConventionalProlific.do, SMTables25and26ConventionalMale.do, SMTables25and26ConventionalFemale.do, SMTables25and26ConventionalWhite.do, SMTables25and26ConventionalNonWhite.do, and AthleteDatawoNAsforDryad.csv.
• SM Table 26 uses SMTables25and26ConventionalBowdoin.do, SMTables25and26ConventionalProlific.do, SMTables25and26ConventionalMale.do, SMTables25and26ConventionalFemale.do, SMTables25and26ConventionalWhite.do, SMTables25and26ConventionalNonWhite.do, and AthleteDatawoNAsforDryad.csv.

DATA-SPECIFIC INFORMATION FOR: AthleteDataforDryad.csv and AthleteDatawoNAsforDryad.csv

Sources: See section 3 of the paper's main text and SM Table 1 in the paper's Supplementary Materials (SM) for sources of the variables.
Number of variables: 88.
Number of athletes/rows: 565.

Note on human subjects: In the author's copy of these files we identify college athletes by name. All of these athletes are public figures and all of their data was pulled from publicly accessible websites. Therefore, the athletes in this dataset are not human subjects whose identities needs to be protected. However, to meet Dryad's independent privacy protection policies we have removed athlete names from the published file. We have also redacted other information that would allow a reader, with some research, to identify the athlete. Redacted variables are noted below with the term REDACTED after the variable description. The full, unredacted data files are available from the authors upon request.

Variable List:

• ID. Athlete ID.
• PhotoGroup. A survey respondent saw pictures of 10 athletes (randomly selected) with the same PhotoGroup number. There are 25 unique PhotoGroup numbers. There are 22 or 23 athletes in each group.
• TOP100. Equals 1 if the athlete was in the On3 Top 100 (a "Superstar"); equals 0 if the athlete was "Conventional."
• Headshot. Equals 1 if the athlete's picture in the survey was a headshot; equals 0 otherwise.
• Helmet. Equals 1 if the athlete had a helmet on their head in their picture in the survey; equals 0 otherwise.
• Smiling. Equals 1 if the athlete was smiling in their picture in the survey; equals 0 otherwise.
• Sharpness. The sharpness of the athlete's picture was measured with ImageMagick, a free, open-source software suite (https://imagemagick.org/). Sharpness refers to how well an image preserves fine details and edges.
• Noise. The noise of the athlete's picture was measured with ImageMagick, a free, open-source software suite (https://imagemagick.org/). Noise is the unwanted random variation in brightness or color that makes an image look grainy or speckled.
• Width. The width of the athlete's picture in pixels.
• Height. The height of the athlete's picture in pixels.
• Resolution. The width x height of the athlete's picture in pixels.
• first_name. First name of the athlete. REDACTED.
• last_name. Last name of the athlete. REDACTED.
• full_name. First and last name of the athlete. REDACTED.
• Prolific_Attractive. Athlete's mean attractiveness score (on a 0 to 10 scale) based on Prolific respondents.
• Prolific_Athletic. Athlete's mean athleticism score (on a 0 to 10 scale) based on Prolific respondents.
• Prolific_Trust. Athlete's mean trustworthiness score (on a 0 to 10 scale) based on Prolific respondents.
• Prolific_Familiar. Athlete's mean familiarity score (on a 0 to 10 scale) based on Prolific respondents.
• Bowdoin_Attractive. Athlete's mean attractiveness score (on a 0 to 10 scale) based on Bowdoin College respondents.
• Bowdoin_Athletic. Athlete's mean athleticism score (on a 0 to 10 scale) based on Bowdoin College respondents.
• Bowdoin_Trust. Athlete's mean trustworthiness score (on a 0 to 10 scale) based on Bowdoin College respondents.
• Bowdoin_Familiar. Athlete's mean familiarity score (on a 0 to 10 scale) based on Bowdoin College respondents.
• Average_Attractive. Mean of Prolific_Attractive and Bowdoin_Attractive.
• Average_Athletic. Mean of Prolific_Athletic and Bowdoin_Athletic.
• Average_Trust. Mean of Prolific_Trust and Bowdoin_Trust.
• Average_Familiar. Mean of Prolific_Familiar and Bowdoin_Familiar.
• White. Equals 1 if the athlete was judged to be white; equals 0 if the athlete was judged to be non-white.
• height_athlete. Athlete height in inches. Blank if unknown. REDACTED.
• current_teams. Name of the school the athlete played for as of the fall of 2024. REDACTED.
• Conference. Name of current_teams' conference as of the fall of 2024.
• Association. Indicates whether the school is part of the NCAA, NAIA, or National Junior College Athletic Association.
• NCAAdummy. Equals 1 if the athlete's current_teams is part of the NCAA; equals 0 otherwise.
• NCAADivision. Equals 1 if the athlete's current_teams is in NCAA's D1, equals 2 if the athlete's current_teams is in NCAA's D2, equals 3 if the athlete's current_teams is in NCAA's D3, and is blank if the athlete's current_teams is not in the NCAA.
• Power4dummy. Equals 1 if the athlete's current_teams is in a Power 4 conference; equals 0 otherwise.
• enrollment. Undergraduate enrollment at the athlete's current_teams as of the fall of 2024.
• Prestige. A system that ranks current_teams on a prestige metric with 1 being the highest and 5 being the lowest using data from Chetty et al. (2026)
• region. The number indicate the school’s location according to the U.S. Census Bureau’s regional divisions (ranges from 1 to 9)
• previous_team_1. Name of the school the athlete previously played for, before joining current_teams. REDACTED.
• previous_team_2. Name of the school the athlete previously played for, before joining previous_team_1. REDACTED.
• previous_team_3. Name of the school the athlete previously played for, before joining previous_team_2. REDACTED.
• previous_team_4. Name of the school the athlete previously played for, before joining previous_team_3. REDACTED.
• transfer_dummy. Equals 1 if the athlete transferred to current_teams; equals 0 otherwise.
• multitransfer. Equals 1 if the athlete has transferred schools multiple times; equals 0 otherwise.
• female. Equals 1 if the athlete plays for a women's team; equals 0 otherwise
• sport_1. The athlete's primary sport (most athlete's just play one sport). Sports include Football, Women's Gymnastics, Women's Beach Volleyball, Women's Tennis, Men's Gymnastics, Men's Wrestling, Men's Basketball, Cheerleading, Women's Lacrosse, Men's Soccer, Women's Soccer, Women's Field Hockey, Men's Para Swimming, Men's Swimming & Diving, Women's Swimming & Diving, Softball, Baseball, Women's Basketball, Women's Ice Hockey, Men's Track & Field, Women's Track & Field, Women's Volleyball, Men's Volleyball, Women's Rowing, Women's Wheelchair Basketball, Women's Acrobatics & Tumbling, Women's Rifle, Men's Golf, Men's Fencing, and Men's Bowling.
• sport2. The athlete's secondary sport, if applicable.
• football. Equals 1 if the athlete's primary sport is football; equals 0 otherwise.
• football_ability. A football player's raw recruitment ranking, on a 1 to 300 scale with 1 being the best. Rankings came from the relevant year’s SportsCenterNext ESPN 300 recruiting database. Some football players are unranked. They are given a ranking of 301.
• mbasketball. Equals 1 if the athlete's primary sport is men's basketball; equals 0 otherwise.
• mbball_ability. A men's basketball player's raw recruitment ranking, on a 1 to 100 scale with 1 being the best. Rankings came from the relevant year’s SportsCenterNext ESPN 100 recruiting database. Some basketball players are unranked. They are given a ranking of 101.
• wbasketball. Equals 1 if the athlete's primary sport is women's basketball; equals 0 otherwise.
• wbball_ability. A women's basketball player's raw recruitment ranking, on a 1 to 100 scale with 1 being the best. Rankings came from the relevant year’s SportsCenterNext ESPN 100 recruiting database. Some basketball players are unranked. They are given a ranking of 101.
• baseball. Equals 1 if the athlete's primary sport is baseball; equals 0 otherwise.
• baseball_ability. A baseball player's raw recruitment ranking, on a 1 to 500 scale with 1 being the best. Rankings came from the relevant year’s Perfect Game recruiting database. Some baseball players are unranked. They are given a ranking of 501.
• softball. Equals 1 if the athlete's primary sport is softball; equals 0 otherwise.
• softball_ability. Equals 1 if the softball athlete was recognized with league conference performance honors in their freshman year (i.e., conference 1st, 2nd, or 3rd team or all-rookie team); equals 0 otherwise.
• vball. Equals 1 if the athlete's primary sport is women's volleyball; equals 0 otherwise.
• vball_ability. Equals 1 if the women's volleyball athlete was recognized with league conference performance honors in their freshman year (i.e., conference 1st, 2nd, or 3rd team or all-rookie team); equals 0 otherwise.
• multisport_dummy. Equals 1 if the athlete plays two or more sports in college; equals 0 otherwise.
• position_1. The name of the position played by the athlete in their primary sport (if applicable).
• Prominent. Equals 1 if the athlete's primary sport position is prominent; equals 0 otherwise. Prominent positions include Quarterback (football), Libero (women's volleyball), and Pitcher (baseball and softball).
• position_2. The name of the position played by the athlete in their secondary sport (if applicable).
• locations. The place name and state name of the athlete's current_teams. REDACTED.
• instagram_url. The athlete's Instagram URL (if applicable). REDACTED.
• ig_followers. The athlete's Instagram followers according to their Opendorse profile as of 9/9/24.
• twitter_url. The athlete's X URL (if applicable). REDACTED.
• twitter_reach. The athlete's X followers according to their Opendorse profile as of 9/9/24.
• tiktok_url. The athlete's TikTok URL (if applicable). REDACTED.
• tiktok_reach. The athlete's TikTok followers according to their Opendorse profile as of 9/9/24.
• biography. The athlete has the ability to write a brief biography on their Opendorse profile page. This is their biography as of 9/9/24. Some athletes did not include a biography on their profile. REDACTED.
• hometown. The athlete's hometown according to their Opendorse profile. REDACTED.
• interests. The athlete's interests according to their Opendorse profile. REDACTED.
• Shoutouts. The price the athlete charged for a Shoutout as of 9/9/24 on their Opendorse profile (in USD).
• ShoutoutChange. The athlete's last observed price for a Shoutout less their first observed price for an Shoutout. Equals 0 if their price for a Shoutout was only observed once or if it was observed multiple times but the difference between last and first price was 0.
• Appearances. The price the athlete charged for an Appearance as of 9/9/24 on their Opendorse profile (in USD).
• AppearancesChange. The athlete's last observed price for an Appearance less their first observed price for an Appearance. Equals 0 if their price for an Appearance was only observed once or if it was observed multiple times but the difference between last and first price was 0.
• Posts. The price the athlete charged for a Post as of 9/9/24 on their Opendorse profile (in USD).
• PostsChange. The athlete's last observed price for a Post less their first observed price for an Post. Equals 0 if their price for a Post was only observed once or if it was observed multiple times but the difference between last and first price was 0.
• Autographs. The price the athlete charged for an Autograph as of 9/9/24 on their Opendorse profile (in USD).
• AutographsChange. The athlete's last observed price for an Autograph less their first observed price for an Autograph. Equals 0 if their price for an Autograph was only observed once or if it was observed multiple times but the difference between last and first price was 0.
• Other. The price the athlete charged for an Other product as of 9/9/24 on their Opendorse profile (in USD).
• OtherChange. The athlete's last observed price for an Other product less their first observed price for an Other product. Equals 0 if their price for an Other product was only observed once or if it was observed multiple times but the difference between last and first price was 0.
• ConvPriceChange. Equals 1 if the athlete is a Conventional athlete who was observed changing at least one of their product prices at least once.
• LOShout. The last observed price the athlete charged for a Shoutout on their Opendorse profile (in USD).
• LOApp. The last observed price the athlete charged for an Appearance on their Opendorse profile (in USD).
• LOPosts. The last observed price the athlete charged for a Post on their Opendorse profile (in USD).
• LOAuto. The last observed price the athlete charged for an Autograph on their Opendorse profile (in USD).
• LOOther. The last observed price the athlete charged for an Other product on their Opendorse profile (in USD).
• Missing data codes: NAs in AthleteData.xslx and blanks in AthleteDatawoNAs.xslx

DATA-SPECIFIC INFORMATION FOR: RaterDataforDryad.csv and RaterDatawoNAsforDryad.csv

Sources: See section 3 of the paper's main text and SM Text 3, SM Tables 2 - 4 in the paper's Supplementary Materials (SM) for sources of the variables and more information on the variables.
Number of variables: 58.
Number of rows: 8044.

Note on human subjects: In the author's copy of these files we identify college athletes by name. All of these athletes are public figures and all of their data was pulled from publicly accessible websites. Therefore, the athletes in this dataset are not human subjects whose identities needs to be protected. However, to meet Dryad's independent privacy protection policies we have removed athlete names from the published file. We have also redacted other information that would allow a reader, with some research, to identify the athlete. This file also includes data taken from a survey. This survey was approved by Bowdoin College's IRB. The survey takers are identified with a ID number that would not allow any reader of this dataset to identify the survey taker. Further, for each survey taker we gathered information on survey takers' race, gender, sexual orientation, age, and education level. However, to meet Dryad's independent privacy protection policies that limits datasets to three indirect identifiers we have removed data on the survey takers' age and education levels. Redacted variables are noted below with the term REDACTED after the variable description. The full, unredacted data files are available from the authors upon request.

Variable List:

• AthleteID. Athlete ID (indexed by i). Matches with ID variable in AtheleteData.csv.
• NumTimesRated. The number of times the athlete was rated by survey respondents
• PhotoGroup. The ID of the photo group an athlete belonged to. A survey respondent saw pictures of the 10 athletes (randomly selected) with the same PhotoGroup number. There are 25 unique PhotoGroup numbers. There are 22 or 23 athletes in each group.
• TOP100. Equals 1 if the athlete was in the On3 Top 100 (a "Superstar"); equals 0 if the athlete was "Conventional."
• Headshot. Equals 1 if the athlete's picture in the survey was a headshot; equals 0 otherwise.
• Helmet. Equals 1 if the athlete had a helmet on their head in their picture in the survey; equals 0 otherwise.
• smiling. Equals 1 if the athlete was smiling in their picture in the survey; equals 0 otherwise.
• first_name. First name of the athlete. REDACTED.
• last_name. Last name of the athlete. REDACTED.
• full_name. First and last name of the athlete. REDACTED.
• RaterID. Unique ID of a rater (indexed by j). A rater saw the pictures of 10 athletes and rated each picture. Therefore, each rater ID (usually) shows up 10 times in the database. We say usually as there are seven raters who skipped the rating of one athlete and thus are in the database nine times.
• Attractive. Rater j's attractiveness score for athlete i, on a 0 to 10 scale.
• Athletic. Rater j's athleticism score for athlete i, on a 0 to 10 scale.
• Trustworthy. Rater j's trustworthiness score for athlete i, on a 0 to 10 scale.
• Familiar. Rater j's familiarity score for athlete i, on a 0 to 10 scale.
• Sharpness. The sharpness of athlete j's picture was measured with ImageMagick, a free, open-source software suite (https://imagemagick.org/). Sharpness refers to how well an image preserves fine details and edges.
• Noise. The noise of athlete i's picture was measured with ImageMagick, a free, open-source software suite (https://imagemagick.org/). Noise is the unwanted random variation in brightness or color that makes an image look grainy or speckled.
• Width. The width of athlete i's picture in pixels
• HeightPhoto. The height of athlete i's picture in pixels.
• Resolution. The width x height of athlete i's picture in pixels.
• Shoutouts. The price the athlete charged for a Shoutout as of 9/9/24 on their Opendorse profile (in USD).
• Appearances. The price the athlete charged for an Appearance as of 9/9/24 on their Opendorse profile (in USD).
• Posts. The price the athlete charged for a Post as of 9/9/24 on their Opendorse profile (in USD).
• Autographs. The price the athlete charged for an Autograph as of 9/9/24 on their Opendorse profile (in USD).
• Other. The price the athlete charged for an Other product as of 9/9/24 on their Opendorse profile (in USD).
• athletefemale. Equals 1 if athlete i plays a women's sport; equals 0 otherwise.
• AthleteWhite. Equals 1 if athlete i is considered white; equals 0 otherwise.
• height. Athlete i's height in inches. REDACTED.
• sport_1. Athlete i's primary sport.
• ig_followers. The athlete's Instagram followers according to their Opendorse profile as of 9/9/24.
• football_ability. A football player's raw recruitment ranking, on a 1 to 300 scale with 1 being the best. Rankings came from the relevant year’s SportsCenterNext ESPN 300 recruiting database. Some football players are unranked. They are given a ranking of 301.
• mbball_ability. A men's basketball player's raw recruitment ranking, on a 1 to 100 scale with 1 being the best. Rankings came from the relevant year’s SportsCenterNext ESPN 100 recruiting database. Some basketball players are unranked. They are given a ranking of 101.
• wbball_ability. A women's basketball player's raw recruitment ranking, on a 1 to 100 scale with 1 being the best. Rankings came from the relevant year’s SportsCenterNext ESPN 100 recruiting database. Some basketball players are unranked. They are given a ranking of 101.
• baseball_ability. A baseball player's raw recruitment ranking, on a 1 to 500 scale with 1 being the best. Rankings came from the relevant year’s Perfect Game recruiting database. Some baseball players are unranked. They are given a ranking of 501.
• softball_ability. Equals 1 if the softball athlete was recognized with league conference performance honors in their freshman year (i.e., conference 1st, 2nd, or 3rd team or all-rookie team); equals 0 otherwise.
• vball_ability. Equals 1 if the women's volleyball athlete was recognized with league conference performance honors in their freshman year (i.e., conference 1st, 2nd, or 3rd team or all-rookie team); equals 0 otherwise.
• rater_race. The race of rater j.
• rater_white. Equals 1 if rater j's race is "White or Caucasian" or "White"; equals 0 otherwise.
• rater_age. Rater j's age. REDACTED.
• rater_gender. Rater j's gender. Categories include female, male, other, and "Prefer not to say."
• rater_female. Equals 1 if Rater_gender = "Female"; equals 0 if Rater_gender ="Male."
• rater_orientation. Rater j's sexual orientation. Categories include bisexual, straight woman, straight man, asexual, gay man, gay woman, other, and "prefer not to say."
• likes_men. Equals 1 if the rater j's "rater_orientation" = bisexual, straight woman, or gay man; equals 0 if rater j's "rater_orientation" = straight man, gay man, asexual, other, or "prefer not to say."
• likes_women. Equals 1 if the rater j's "rater_orientation" = bisexual, straight man, or gay woman; equals 0 if rater j's "rater_orientation" = straight woman, gay man, gay man, asexual, other, or "prefer not to say."
• samegender. Equals 1 if athlete i's "female" = 1 (from AthleteData.csv) and Rater_gender = "female" or if athlete i's "female" = 0 (from AthleteData.csv) and Rater_gender = "male" , otherwise equals 0.
• gaybi. Equals 1 if the rater j's "rater_orientation" = bisexual, gay man, or gay woman; equals 0 if rater j's "rater_orientation" = straight man, straight woman, or asexual.
• bisexual. Equals 1 if the rater j's "rater_orientation" = bisexual, equals 0 if rater j's "rater_orientation" = straight man, straight woman, gay man, or gay woman.
• gay. Equals 1 if the rater j's "rater_orientation" = gay man or gay woman, equals 0 if rater j's "rater_orientation" = straight man, straight woman, or bisexual.
• samerace. Equals 1 if athlete i's "AthleteWhite" = 1 and rater_white = 1 or if athlete i's "AthleteWhite" = 0 and rater_white = 0, otherwise equals 0.
• rater_edu. Rater j's educational attainment. Categories include Bowdoin student, Associate degree in college (2-year), Bachelor's degree in college (4-year), Doctoral degree, High school graduate (high school diploma or equivalent including GED), Less than high school degree, Master's degree, Professional degree (JD, MD), and Some college but no degree. REDACTED.
• rater_reason. Rater j's response to "Please complete the following statement (choose all that apply): Generally, I thought the athletes were more physically attractive when they were... Respondents could check one or more of the following reasons: A) Playing for a team I like; B) Smiling; C) The same gender as me; D) The same race as me; E) Wearing dress clothes; F) Wearing their sports uniform; G) Other (please explain below). If they checked “Other” they had an opportunity to explain their answer in a text box.
• reason_other. Rater j's explanation of rater _reason = "Other" (if applicable).
• biasawareness. Rater j's answer to the question: "Do you think more beautiful people, all else equal, make higher wages?" Rater j could check “Yes” or “No.”
• Fairness. If rater j answered “Yes” to the biasawareness question they were asked “If yes, do you think this is fair?” They could say “Yes or “No” to this follow-up question as well."
• Duration. Rater j's survey duration (in seconds).
• Prolific. Equals 1 if rater j was hired through Prolific; equals 0 if rater j was a Bowdoin student.
• Varsity. Rater j's answer to "Do you participate in College sponsored varsity athletics? " They could answer “Yes” or “No.” Only asked of Bowdon students. REDACTED.
• varsity_dummy. Equals 1 if rater j checked "Yes" to Varsity question and equals 0 if they checked "No" to Varsity question. REDACTED.
  Missing data codes: NAs in RaterDataforDryad.csv and blanks in RaterDatawoNAsforDryad.csv.

DATA-SPECIFIC INFORMATION FOR: RespondentInfoforDryad.csv.

Sources: Survey.
Number of variables: 13.
Number of rows: 805.

Note on human subjects: This file contains data taken from a survey. This survey was approved by Bowdoin College's IRB. The survey takers are identified with a ID number that would not allow any reader of this dataset to identify the survey taker. Further, for each survey taker we gathered information on survey takers' race, gender, sexual orientation, age, and education level. However, to meet Dryad's independent privacy protection policies that limits datasets to three indirect identifiers we have removed data on the survey takers' age and education levels. Redacted variables are noted below with the term REDACTED after the variable description. The full, unredacted data files are available from the authors upon request.

Variable List:

• ID. Unique ID of a rater. This ID numbering system dos not match that of the rater ID in RaterDataforDryad.csv. This better protects the anonymity of the survey takers.
• Orientation. Rater's sexual orientation. Categories include bisexual, straight woman, straight man, asexual, gay man, gay woman, other, and "prefer not to say."
• Education. Rater's educational attainment. Categories include Bowdoin student, Associate degree in college (2-year), Bachelor's degree in college (4-year), Doctoral degree, High school graduate (high school diploma or equivalent including GED), Less than high school degree, Master's degree, Professional degree (JD, MD), and Some college but no degree. REDACTED
• BeautyReason. Rater's response to "Please complete the following statement (choose all that apply): Generally, I thought the athletes were more physically attractive when they were..." Respondents could check one or more of the following reasons: A) Playing for a team I like; B) Smiling; C) The same gender as me; D) The same race as me; E) Wearing dress clothes; F) Wearing their sports uniform; G) Other (please explain below). If they checked “Other” they had an opportunity to explain their answer in a text box.
• BeautyReason_9_TEXT. Rater's explanation of rater _reason = "Other" (if applicable).
• BiasAwareness. Rater's answer to the question: "Do you think more beautiful people, all else equal, make higher wages?" Rater j could check “Yes” or “No.”
• Fairness. If the rater answered “Yes” to the biasawareness question they were asked “If yes, do you think this is fair?” They could say “Yes or “No” to this follow-up question as well."
• Sex. Rater's gender. Categories include female, male, other, and "Prefer not to say."
• Age. Rater's age. REDACTED.
• Race. The race of the rater.
• TimeTaken. Rater's survey duration (in seconds).
• Prolific. Equals 1 if rater j was hired through Prolific; equals 0 if rater j was a Bowdoin student.
• Varsity. Rater's answer to "Do you participate in College sponsored varsity athletics? " They could answer “Yes” or “No.” Only asked of Bowdon students. REDACTED.
  Missing data codes: NAs.

Human subjects data

In this research project we surveyed 305 Bowdoin College students and 500 Prolific users. Bowdoin College's IRB approved the survey for both Bowdoin College and Prolific respondents. We have included survey results in this repository along with some demographic information of each survey taker. However, the combination of demographic information is not enough to identify a particular person. Further, we have created an internal ID system that would not allow any particular survey taker to be identified.

Further, we have included data on named athletes. However, all the athlete data was collected from public sources on the internet. None of the athlete data is private.