Laboratory Evidence on the Effects of Sponsorship on the Competitive Preferences of Men and Women

Formalized workplace sponsorship programs benefit men far more than women, but harnessing the positive impact of sponsors’ confidence in female protégés’ abilities could help close gender gaps.


For the past twenty years, women have made up roughly half of entry cohorts at public accounting professions, yet less than 20% of firm partners are female. Previous research has found that long hours, client demands, the lack of work/life balance, limited access to positive role models, and pressure to compete and win business  – factors that disproportionately affect women – lead employees to veer off the partner track. Although firms have initiated flexible work schedules, professional development and mentoring programs, there has been little change in the advancement of female employees in recent decades (Moore, 2013).

In this laboratory experiment, researchers test whether women’s advancement can be improved using a sponsorship program. Distinct from a mentor, a sponsor – often an influential member of the organization – acts as an advocate for a protégé in what is ideally a transactional relationship. Sponsors are expected to use their knowledge, expertise and connections to advance the career of protégés, which they in turn benefit from via reputational and monetary ties to the protégé’s performance. Prior research suggests that sponsorship increases confidence, compensation, and willingness to compete within one’s cohort (Hewlett et al, 2010).

This paper examines the effectiveness of sponsorship programs in a lab setting in increasing women’s advancement by evaluating the effects of two potential features of a sponsorship program: 1) a positive, credible signal of a protégé’s ability and 2) tying the sponsor’s compensation to the protégé’s performance.


In a laboratory setting, sponsorship programs failed to close the gender gap in competitiveness, but instead increased competitiveness and earnings among male protégés, particularly among low-ability, overconfident men.

  • Men were more willing to compete than women: Overall, at every wage offered in the tournament, men chose to compete more than women. The largest gender gap in willingness to compete was for a tournament wage of $1.50 per correct answer where half of men and one-third of women chose to compete.
  • Sponsorship increased men’s willingness to compete by approximately $0.32 (i.e. were willing to accept $0.32 less per correct answer to participate in the tournament) compared to the control group, but sponsorship had no significant impact on women’s willingness to compete.
  • Men made significantly more money than women: More than a third of the earnings gap was driven by differential entry into competition. When given the option to compete, men earned an average of $11.25, while women earned $7.94 on average. Controlling for the number of problems solved, women still earned an average of $1.03 less than men.
  • Sponsorship benefitted low-performing men the most, who increased their willingness to compete by an average of $0.57 when sponsored in any form. By comparison, high-performing men who were sponsored only increased their willingness to compete by $0.14, though the latter finding is not statistically significant.
  • Men were more confident in their ability to perform than were women. Nearly 55% of men believed that their performance would place them in the top 25% of performers, while only 34% of women made this prediction. Sponsorship did not affect women’s estimation of their ability to perform, though it did increase overconfident men’s willingness to compete.
  • Women responded more positively to positive feedback: Women were more willing to compete, though still at lower levels then men, when their sponsor signaled they were doing well. Their competition did not increase when it was tied to their sponsor’s compensation. Men, on the other hand, responded strongly to both the signals about their ability and information that their sponsors’ compensation was tied to their performance.

Overall, sponsorship neither reduced the gender gap in willingness to compete nor the gender gap in earnings. But while women saw no significant increase in willingness to compete or in their earnings, they did gain some benefit from their sponsors’ confidence in their ability.


The experiment consisted of four rounds where participants were given addition problems to solve, summing up five two-digit numbers.  In each round, a list of 30 problems appeared on the screen and participants had four minutes in which to complete as many problems as possible. The incentive schemes varied by round.

In the first round, participants earned $0.50 for each problem solved. In the second round, payment varies by relative performance. Participants who placed in the top 25% earned $2 for each problem solved, while those in the bottom 75% earned nothing. In Round 3, participants were given the choice between being paid according to the payment scheme in Round 1 or the payment scheme in Round 2. However, an additional component was added to the Round 2-style payment scheme to gauge willingness to compete. This time, an individual would be paid nothing if they finished in the bottom 75% of performers, but if they finished in the top 25%, they would receive a reward ranging from $1 to $3 per problem completed. Participants were asked which payment scheme they preferred, as the authors varied the potential reward for the Round 2 scheme by $0.25 at a time. This allowed the researchers to measure how willingness to compete varied with the size of the reward from competition. The authors used the lowest tournament wage at which the participant preferred the tournament to the piece rate payment scheme as the participant’s cutoff to measure willingness to compete.

The sponsorship intervention occurs after Round 3 and consists of two features: a signal of ability and linking sponsor compensation to protégé performance. Participants are informed that Round 4 is a repetition of Round 3. They are then randomly assigned into one of five treatment groups: 1) Sponsor: act as sponsor instead of participant, 2) Control: continue as participant without a sponsor, 3) Belief Signal: participants receive a credible, positive signal of their ability from a sponsor, 4) Payment-Tying:  a sponsor’s compensation is tied to participants’ choices and performance are tied, and 5) Belief Signal and Payment-Tying: participants experience both features of sponsorship.

Within each lab session, three individuals were randomly selected as sponsors. Sponsors earned $0.25 per correctly solved problem for each protégé who both (a) ranks among the top 25% of performers and (b) chooses Payment Scheme B (the tournament option). For the Payment-Tying treatment, sponsors were provided with a list of randomly selected participants to sponsor.

In the Belief Signal treatment and Belief Signal and Payment-Tying treatment, sponsors could make decisions about whom they want to sponsor. In making their decisions, they were provided with the Round 1 and 2 scores of the participants and made a series of pair-wise decisions between potential protégés. After choosing one protégé per pair-wise choice, a coin was flipped to determine how sponsors would be paid. If the coin was heads, sponsors earned payment based upon the performance and choices of their protégés. If the coin was tails, sponsors instead earned a fixed payment for Round 4. This generates the “Belief Signal” treatment, since chosen protégés will know they have been chosen as a protégé, but their performance does not affect the remuneration of their sponsor. Participants other than the control group received a handout explaining that sponsors were chosen randomly from the participant pool. They also received additional information about whether they had been randomly assigned to a sponsor or chosen by a sponsor. Participants then completed Round 4, which was structured exactly like Round 3. At the end of Round 4, participants are asked demographic questions.

The experiment comprised 12 sessions with 354 participants, 176 males and 178 females.

MLA: Baldiga, Nancy R., and Katherine B. Coffman, “Lab Evidence on the Effects of Sponsorship on Preferences for Competition”. Working Paper.

APA: Baldiga, Nancy R., and Katherine B. Coffman, “Lab Evidence on the Effects of Sponsorship on Preferences for Competition. Working Paper.

Chicago: Baldiga, Nancy R., and Katherine B. Coffman, “Lab Evidence on the Effects of Sponsorship on Preferences for Competition. Working Paper.