1 / 10

Anuj Kapoor and Catherine Tucker By Qiang (David) Gao Baruch College, City University of New York

Discussion of : How do Platform Participants respond to an Unfair Rating? An Analysis of a Ride-Sharing Platform using a Quasi-Experiment. Anuj Kapoor and Catherine Tucker By Qiang (David) Gao Baruch College, City University of New York. Summary.

jhanna
Télécharger la présentation

Anuj Kapoor and Catherine Tucker By Qiang (David) Gao Baruch College, City University of New York

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Discussion of:How do Platform Participants respond to an Unfair Rating? An Analysis of a Ride-Sharing Platform using a Quasi-Experiment Anuj Kapoor and Catherine Tucker By Qiang (David) Gao Baruch College, City University of New York

  2. Summary • Study how biased negative reviews from taxi riders affect taxi driver’s behaviors using six-month a taxi-sharing data. • Exploit the driver replacement • It may cause rider to give biased lower ratings • Examine the effect of biased lower ratings on driver’s response • Negative response to biased lower rating (i.e.,less driver efforts) leads to more lower ratings • Investigate the consequence of biased lower rating • Drivers are more likely to leave the platform

  3. From https://thepointsguy.com/2017/04/uber-changing-rating-system/ If you’re a driver, well, if your rating goes below a certain threshold, you won’t be driving for much longer. Uber is encouraging riders to be more fair when it comes to rating their UberPool driver. Say, for example, you request an UberPool and aren’t happy with the rider(s) you got paired with and gave the driver a lower rating because of it. Now, if you don’t give a five-star rating to your UberPool driver, you’ll be able to give more feedback on why you rated the way you did.

  4. from https://www.theverge.com/2016/10/3/13155578/amazon-incentivized-reviews-ban-vine-program-product-biasAmazon Amazon is cracking down on biased customer reviews Yet studies have shown that incentivized reviewers are less likely to give products negative feedback and review hundreds of products on average, potentially affecting the overall sales performance of otherwise mediocre items.

  5. Importance • Important question to study. • Most existing studies regarding peer reviews focus on • The reliability of review ratings • The impact of review • The impact of biased positive review • How about the impact of biased negative reviews?

  6. Comments • Comment #1: Biased review definition • Comment #2: Selection issue • Comment #3: Result interpretations and alternative models

  7. Comment 1: Biased review definition • Negative Rating Biased Rating • Current definition • Biased rating: Drivers with 5* ratings earlier but lower rating later as replacements • Original argument: “a bad review is often an absent review” • Using other drivers’ reviews as benchmark? • Using all rating levels? • Negative reviews • About drivers or other things? • Any text?

  8. Comment #2: Selection issue • Current data: • Total 550,000 trips • 3% of these trips reassigned • Many missing reviews after reassignment • Selection biases?

  9. Comment #3: Results interpretations and alternative models • Result interpretations • Some interpretations need to be clearer • e.g., Using IV will increase driver negative responses effects, why? • Some results need further explanations. • e.g., “drivers with a low rating are not more likely to leave the platform than the drivers with higher ratings”. • How large is this effect? • It may be better using standard deviation. • Alternative models • To verify LPM, conditional logit model with individual fixed effects.

  10. Thank You!

More Related