Experimental Design
1. We randomize all participants into treatment groups with varying conditions for managerial attention. Managerial attention contains the information that the principal (primary investigator) recognizes the work effort and results of the agents. Therefore, we build expectations about granting recognition, which are then systematically met or not met.
2. Depending on the treatment, subjects receive an announcement that there will be "attention" from the primary investigator multiple times over the next six weeks and also which condition must be fulfilled for ongoing managerial attention.
3. Subjects, with the exception of the control group, receive management attention e.g., in the for short videos, messages, or voice notes, depending on the treatment. Some groups receive attention regardless of their performance, while other groups must meet performance-based conditions.
Participants are randomly distributed into different treatment groups (n ≈ 150 participants/group). We designed the following 4 groups, based on a between-subject design.
[Group 1] Control group, will receive no messages or managerial attention at all.
[Group 2] Participants in this group receive a short announcement that occurrence of recognition is independent of performance (i.e., happens with good/bad performance or declining/increasing/staying the same). Over the six weeks, we let the random number generator decide the probability whether there is attention or not multiple times per week. Therefore, the range of managerial attention messages lies between 0 (min.) and 3 (max.) messages per week.
[Group 3] Participants in this group receive a short announcement that occurrence of recognition is dependent on performance and only occurs as performance increases. If participants increase their individual performance, they receive managerial attention, if they hold their performance or its decreasing, they won’t. The range of managerial attention messages lies between 0 (min.) if the participants never managed to increase the performance to 3 (max.) messages per week if the participants constantly increase the performance.
[Group 4] Participants in this group receive a short announcement that occurrence of recognition is dependent on performance and occurs not only when performance increases, but also when performance remains constant. The range of managerial attention messages lies between 0 (min.) if the participants never decrease the performance multiple times to 3 (max.) messages per week if the participant performance is constant or increases multiple times.
We expect the following trends in performance:
H1: With decreasing expectations that the principal will perceive their performance, subjects lower their performance.
H2a: If subjects have an expectation that the probability of managerial attention increases with performance, they increase their performance.
H2b: If the probability of managerial attention occurring depends on performance, then poor (measured in absolute terms) /decreased (measured in relative terms) performance leads directly to even poorer/further decreased performance.
H3a: Over time, we find that subjects who have not received managerial attention for a longer period of time/who have repeatedly failed to receive expected recognition lower their performance.
H3b: If during a downward trend, triggered by the withdrawal of expected managerial attention, the principal takes corrective action through managerial attention, he will no longer be able to stop the drop in performance, since the principal's reputation is permanently damaged.
H3c: The downward trend does not occur when managerial attention is independent of effort or also relates to poor performance.