Authors:
(1) Muhammad Zia Hydari, Katz Graduate School of Business, University of Pittsburgh and corresponding author;
(2) Idris Adjurid, Pampelin College of Business;
(3) AAARON D. STRIEGEL, Department of IT and Engineering, University of Notre Dame.
Ties
Summary and 1 Introduction
2. Context and 2.1. Classification
3. Effect of rankings on healthy physical activity and 3.1. Competition
3.2. Social influence
3.3. Moderation of the effects of previous activity levels and classification size
4. Data and model
4.1. Data
4.2. Model
5. Estimate and robustness of the main effects of rankings
5.2. Verification of robustness for the initiation of rankings
5.3. Fitbit compliance
5.4. Attrition of fitbit, classification dimensions and additional robustness checks
6. Heterogeneous effect of rankings
6.1. Heterogeneity by previous activity levels
6.2. Interaction of the size of the classification, row and previous activity levels
6.3. Summary of the results of the heterogeneous effects analysis
7. Conclusions and discussion, end notes and references
3.3. Moderation of the effects of previous activity levels and classification size
The contradictory effects of competition and social influence not only make the direction of the average effect uncertain, they also underline the presence of heterogeneity in the effects. To disentangle this heterogeneity, we consider factors that can have an impact on the propensity to observe the positive dynamics vs negative of classification on physical activity.
3.3.1. Classification size. First, we examine whether the size of the classification, that is to say the number of other participants active in the classification, is an important potential moderator of the classification impacts. Garcia et al. (2013) suggest that an important situation factor with an impact on comparison concerns and competitiveness is the number of competitors. On the one hand, the increase in the number of active participants is likely to increase the probability of the positive dynamics that the rankings introduce. Obviously, competition mechanisms, mutual responsibility and perceived capacity changes are nonexistent if there are no other active users in a classification. Furthermore, competitive reasons can be stronger for larger rankings, because the classification of larger rankings can be more motivational than dominating smaller rankings. That said, the effect of increasing the size of the classification is probably more nuanced. For example, it is likely that certain advantages of participants in an additional classification are decreasing on the margin. Too many participants can make the classification less effective because participants get lost in the crowd, weakening the positive impacts of competition or mutual responsibility (Garcia and Tor 2009, Garcia et al. 2013). The marginal benefit decreasing an additional classification member implies non-linearity in favor of more classification members and can even lead to harmful effects of the rankings if they become too important.
3.3.2. Previous activity levels. Second, we consider an individual's physical activity level before the adoption of rankings.
3.3.2.1. Competition. If we consider only the role of competition (vis-vis-at-vis anterior activity levels), the expectation in the literature according to which highly active individuals should benefit from the rankings are the most plausible (Patel et al.2015, Wu et al.2015, Shameli et al.2017). Individuals with high activity levels before the adoption of rankings gain a high utility of healthy activity and are therefore likely to perform well in the rankings. This positive performance on rankings can be motivating for them and encouraging the increase in future physical activity. The impact of competitive dynamics on relatively more sedentary individuals can be more nebulous. On the one hand, these people can benefit from the most extrinsic motivations such as competition and rank against others. On the other hand, the value of the rankings for these individuals can be limited by their lower intrinsic ability and their motivation for physical activity. This makes them subject to the diso-motivating impacts of dull performance on rankings.
3.3.2.2. Points of responsibility and reference. If we also consider the theorized mechanisms linked to social influence, the EX ANTE wait is more uncertain. Since individuals at the bottom of the distribution of physical activity are more likely to have other participants in a classification which are more active than them, there is an increased potential for the classification of acting as a tool which maintains them responsible; Persons more active than the focal user can be more credible in their attempts to hold the responsible focal user. Furthermore, individuals at the lower end of the distribution of physical activity are more likely to meet other users who facilitate upward comparisons and have a positive impact on their reference point for exercise and their perceived capacity to engage in physical activity. In addition, individuals with low activity levels before the adoption of rankings can benefit most of the rankings because they have more room for improvement and a higher need for external motivation. The dynamics around social influence is somewhat reversed for those who are very active before the adoption of the rankings. After the same justification, people who are already very active may be less likely to join rankings where other users can keep them responsible (that is to say that few others in their classification can correspond to their levels of physical activity). In addition, these people are at high risk of rankings facilitating downward comparisons which have a negative impact on their reference points. These comparisons can induce slowness if they highlight the disproportionate level of activity of the focal user compared to the others. Finally, very active individuals can suffer from ceiling effects, that is to say that any extrinsic intervention is unlikely to increase their will or its ability to increase physical activity.
3.3.3. Classification mechanisms, previous activity levels and size of the rankings. The theorized effects of previous activity levels and the size of the rankings can also cross in order to have implications for the various mechanisms by which the rankings can have an impact on behavior. First, our theorized mechanisms indicate that very active individuals are most likely to be injured by small rankings. Garcia et al. (2013) suggest that competitiveness emerges when there is a potential for comparisons, high or down, which in a credible way the rank of the individual. With smaller rankings (for example, another individual), these high performance are less likely to interact with another person who can compete in a credible manner or keep them responsible (thus canceling two key mechanisms for the value of the rankings). At the same time, they are more likely to be presented with a salient individual who facilitates downward comparisons which have a negative impact on their point of exercise, induce slowness and decrease their levels of physical activity. As the size of the rankings increases, there is an increased potential value for the very active, because the probability increases by at least an individual junction which can provide a credible threat to their rank, their mutual responsibility and their positive impacts on their reference point for the exercise. In addition, it is plausible that very active people are led to work even better when they are part of a relatively large classification. This phenomenon would be similar to the idea in certain sports of a “great match player”, someone who takes place above his average on special occasions and in front of a large crowd.
On the other hand, our theorized mechanisms have different implications for the size of the classification when individuals were sedentary before adoption. Unlike very active individuals, these people can always benefit from the adoption of small rankings because they are still likely to meet other users who are at a comparable or higher level of physical activity. Thus, even small rankings can often provide these people with an additional degree of responsibility and the potential for positive impacts on their exercise reference points. The question of whether these individuals benefit from competition with small rankings are less certain, because they can always be dominated on small rankings, which leads to deactivational effects of competition. The increase in the size of the rankings for lower activity users can always provide some of the advantages described above, but it is likely that these advantages are decreasing more quickly for this group. Unlike very active individuals, the advantages of mutual responsibility can be reduced for these people as the size of the rankings increases (via the phenomenon “getting lost in the crowd” described previously). Even more, these users, who are at the lower end of the distribution of physical activity, may be more likely to be stuck downwards and this can be more salient with more participating users. Overall, we conjecture that sedentary individuals can benefit considerably even when the rankings are low. However, increasing the size of the classification may have a decrease in marginal advantages for them.
