Team Leadership, Experience, and Gender

Subjects are stratified by gender and randomly assigned to the treatment condition (leadership position) or to the control condition (team member position). The laboratory experiment investigates the gender difference in the effect of leadership experience on preferences for positions including team leadership.

2023-09-12

2023-12-31

The primary outcome variables are the bids submitted in the application stage in which subjects apply to become either a leader with high or low salary.
It will be tested whether (i) the experience of being a leader in Stage 1 leads to a difference in bids submitted by Stage 1 team members and leaders for both types of applications (low or high salary leadership position) and (ii) whether the effects differ by gender. The main specifications will regress submitted bids on treatment, gender, and the respective interaction term. Standard errors will be clustered at Stage-1-group level.

Secondary analyses study the role of experience that leaders have made in Stage 1 of the experiment (positive or negative). The induced random variation in performance feedback will be exploited to estimate the causal effects of experience on bidding behavior. In addition, risk and confidence measures are elicited throughout the study which are also included in further analyses.
Furthermore, stereotypes connected to the leader positions in Stages 1 and 3 are studied using i) subjects’ assessment on whether the leader tasks are rather suited for men or women and ii) the (incentivized) beliefs about male and female team leaders’ performance in Stage 1 and 3 of the experiment. To further explore underlying mechanisms which explain subjects’ behavior, chat protocols as well as answers to open-ended questions about the reasons for subjects’ bidding behavior are analyzed. Moreover, survey answers on the willingness to be a leader will provide additional evidence.

Stage 1: Subjects play a standard public good game (PGG). The PGG is adapted to the context of interest by adding a communication tool. The treated subjects – the leaders – have the communication power within their groups. Treatment assignment is stratified by gender (please see Randomization Method for more details).

Stage 2: Subjects apply for a position as leader in Stage 3 by using the strategy method: they apply to a high and a low salary leadership position by submitting bids in a second price sealed bid auction (Vickrey, 1961). One of their bids is randomly chosen to be (payoff) relevant.

Stage 3: Stage 3 is almost equivalent to Stage 1. The main difference to Stage 1 is that subjects are not randomly assigned to leadership positions, but the assignment will follow the outcomes of Stage 2, the application stage.

During the experiment subjects answer demographic questions, provide survey evidence on e.g., their willingness to lead (Giessner et al. 2022), and state their beliefs about other subjects’ behavior. Subjects’ risk preferences and confidence is measured as well.

The experiment will be conducted in the Cologne Laboratory of Economic Research (CLER).

The final sample will exclude subjects that:
1. reveal their identity or gender in the chat;
2. fail the attention check in the questionnaires at the end of the experiment;
3. independently quit the experiment before its end;
4. do not submit bids in the application stage (Stage 2) because they are forced to drop out earlier*
5. do not choose the option “male” or “female” (due to an expected low number of participants identifying as “diverse” and subsequent power limitations)

*Subjects who do not finish the experiment because (one of) their group members quit(s) the experiment will be included in the final sample if they have submitted their bids in the application stage. If they cannot proceed to the application stage but are forced to drop out earlier due to their group member(s), they will be excluded from the final sample.


The experiment is designed as follows:

Before the main experiment starts, subjects fill in a consent form. Thereafter, participants’ risk preferences are elicited (incentivized), they answer a questionnaire about their demographic information (gender, age, occupation, faculty of study) followed by a second, incentivized risk measure.

Stage 1:
Subjects play a standard public good game (PGG) which induces a social dilemma situation. The standard PGG is adapted to the context of interest by adding a communication tool, i.e., a free-form chat in which subjects can freely chat with each other (e.g., Eisenkopf (2014)). The chat can thus be used to persuade others to contribute to the public good. Before the game starts, subjects are asked to state their belief about their relative persuasion ability in comparison to other subjects. Then, they are randomly assigned to groups of three.
Importantly, subjects will have different positions which they will be randomly assigned to (see Randomization Method for further details): leaders and team members. In each group, one subject is treated and the remaining two belong to the control group. The treated subjects, the leaders, can chat with both team members simultaneously in two distinct chat rooms whereas the control subjects, the team members, can only chat with the team leader but not with the other team member. The chats close after a predetermined time and all subjects have the possibility to make their contribution decision. The contribution decision is followed by additional elicitations of subjects’ self-confidence in their persuasion ability. Afterwards, subjects receive a noisy feedback measure: the team members’ contribution is biased by a noise term in the range of [-2, -1, 0,1,2]. Participants are asked about their guess about the true group members’ contribution level. Chat protocols will be saved to use them for further analyses, but subjects are not allowed to reveal their identities when chatting.

Stage 2:
In this stage, Stage 1’s groups are randomly mixed so that the group composition of Stage 1 and 2 might differ. Participants learn that in Stage 3 the team leader will be rewarded with a low or a high fixed salary and that it will be randomly determined which of the two will be the relevant position for them. By using the strategy method, subjects apply to both positions in a second price sealed bid auction (Vickrey, 1961). For this purpose, subjects are endowed with an extra budget unrelated to the payoffs they have earned in Stage 1. Winning subjects cannot pay more than the entire extra budget to become a leader and loosing subjects receive the entire extra budget from this stage. After the bid submissions, subjects briefly explain in an open-text field why they have (not) applied.

Stage 3:
Stage 3 is almost equivalent to Stage 1. Subjects again play the PGG with chats, contributions, and noisy feedback in the same groups as in Stage 2. Subjects are not randomly assigned to leader positions, but the assignment will follow the outcomes of Stage 2, the application stage. Also, leaders receive either the high or the low fixed salary depending on which position and application was randomly determined to be relevant for them.

Subjects are then asked to answer survey questions followed by the elicitation of subjects’ assessment on whether the leader tasks are rather suited for men or women as well as incentivized beliefs about male and female team leaders’ performance in Stage 1 and 3.

It will be randomly determined whether subjects’ payment is based on Stages 1 and 2 or on Stages 2 and 3. Payoffs earned during the incentivized elicitations of risk and belief questions are paid out according to subjects’ answers. Subjects will receive full information about their payoffs at the end of the experiment.

The randomization will be done using the software otree (Chen et al., 2016). An equal number of men and women will be invited to the experiment. Subjects will indicate their current gender in the beginning of the experiment. This information is used to assign all female subjects to one stratum and all male subjects to a second stratum. Subjects who choose the third option (“diverse”) are assigned to the team-member position because the share of potential participants who chose “diverse” when registering for experiments of the CLER is lower than 0.01%. However, it cannot be guaranteed that every person invited will show up which might lead to an unbalanced number of subjects in each stratum. The randomization method takes this into account:
In each stratum, every third subject is assigned to the treatment condition. Thus, ideally, an equal number of male and female subjects is assigned to leadership positions. The leaders are then randomly assigned to two randomly chosen team members. If there is a mismatch in the number of leaders and team members, the subjects who could not be assigned to a group of three will be grouped together. Within this group, one person is randomly chosen to be the leader.

Individual subject

No

The number of clusters is the same as the number of observations (please see below).

The planned number of participants is 900.

One third of the subjects (i.e. 300) will be randomly assigned to the treatment condition (i.e. the role as leader) in Stage 1.