# Social media self-branding and success: Quantitative evidence from a model competition

**Authors:** Fabian Braesemann, Fabian Stephany  
**arXiv:** [2007.04113](https://arxiv.org/abs/2007.04113)  
**Date:** 2020-07-08  
**Summary page:** https://www.summarizepaper.com/en/paper/2007.04113v1/

## Key points

- Study explores impact of social media activity on chances of models being elected for Playboy Magazine's Playmate of the Year award
- Researchers hypothesize that active management of Instagram accounts can increase likelihood of winning award
- Findings indicate that social media activity can predict outcome of the award
- Candidates who manage their social media accounts more actively have higher probability of becoming Playmate of the Year
- Researchers conducted cross-validated out-of-sample rank predictions for successful candidates
- Model ranked eventual successful candidate among top-three in seven out of eight cases, and correctly predicted three competition outcomes
- Results demonstrate that model surpasses baseline distribution and indicates predictive power of social media data for forecasting contest outcomes

## Summary

The study presented in this research paper explores the impact of social media activity on the chances of models being elected for Playboy Magazine's Playmate of the Year award. The researchers hypothesize that candidates who actively manage their Instagram accounts can increase their likelihood of winning the award by using social media to gain more followers, who may then vote for them in the award polls. The findings indicate that social media activity has predictive capacity to estimate the outcome of the award. In fact, candidates who manage their social media accounts more actively than others have a higher probability of becoming Playmate of the Year. To further support their findings, the researchers conducted cross-validated out-of-sample rank predictions for successful candidates. They trained their model with data from seven competitions and predicted the outcome of the remaining contest. In seven out of eight cases, the model ranked the eventual successful candidate among the top-three (in six cases among the top-two) and correctly predicted three competition outcomes. These results demonstrate that the model surpasses a baseline distribution of a simple random draw, indicating that social media data indeed has predictive power to forecast contest outcomes.