Experimental Design in Two-Sided Platforms: An Analysis of Bias

Experimental Design in Two-Sided Platforms: An Analysis of Bias
复制标题

两侧平台的实验设计:偏差分析

DOI:
10.1145/3391403.3399507
复制
发表时间:
2020
期刊:
ACM Economics and Computation (EC
影响因子:
--
通讯作者:
Weintraub, Gabriel
Weintraub, Gabriel
中科院分区:
--
文献类型:
--
作者:
Johari, Ramesh;Li, Hannah;Weintraub, Gabriel

文献摘要

参考文献

被引文献

相似文献

我们开发了一个分析框架来研究双边市场中的实验设计。这些实验中有许多都涉及比特干扰,即对一个市场参与者的干预会影响另一个参与者的行为。这种干扰会导致对干预措施治疗效果的偏倚估计。我们开发了一个随机市场模型和相关的平均场限制,以捕捉动态在这样的实验中,并使用我们的模型来研究不同的设计和估计的性能是如何受到市场干扰效应。平台通常使用两种常见的实验设计:需求方“客户”随机化()和供应方“列表”随机化(),以及沿着相关的估计量。我们表明,良好的实验设计取决于市场平衡;在高度需求约束的市场,是无偏的,其中有偏;相反,在高度供应约束的市场,是无偏的,其中有偏。我们还介绍和研究了一种新的实验设计,基于双侧随机化(),其中客户和列表都被随机分配到治疗组和对照组。我们表明,适当的选择ofdesigns可以在市场平衡的两个极端是公正的,而产生相对较低的偏差,在中间制度的市场balance.This论文被接受的大卫Simchi-列维,收入管理和市场分析。
We develop an analytical framework to study experimental design in two-sided marketplaces. Many of these experiments exhibitinterference, where an intervention applied to one market participant influences the behavior of another participant. This interference leads to biased estimates of the treatment effect of the intervention. We develop a stochastic market model and associated mean field limit to capture dynamics in such experiments and use our model to investigate how the performance of different designs and estimators is affected by marketplace interference effects. Platforms typically use two common experimental designs: demand-side “customer” randomization () and supply-side “listing” randomization (), along with their associated estimators. We show that good experimental design depends on market balance; in highly demand-constrained markets,is unbiased, whereasis biased; conversely, in highly supply-constrained markets,is unbiased, whereasis biased. We also introduce and study a novel experimental design based ontwo-sided randomization() where both customers and listings are randomized to treatment and control. We show that appropriate choices ofdesigns can be unbiased in both extremes of market balance while yielding relatively low bias in intermediate regimes of market balance.This paper was accepted by David Simchi-Levi, revenue management and market analytics.
限制在线市场实验中测试控制干扰的偏差
DOI: --
发表时间: 2020
期刊:
影响因子: --
作者:
David Holtz;Sinan Aral
通讯作者: Sinan Aral
DOI: --
发表时间: 2019
期刊: Management Sciences
影响因子: --
作者:
Stefan Wager;Kuang Xu
通讯作者: Kuang Xu
DOI: 10.1111/insr.12170
发表时间: 2016-04
影响因子: 2
作者:
C. Blumberg
通讯作者: C. Blumberg
纯跳跃马尔可夫过程的流体极限:实用指南
DOI: --
发表时间: 2002
期刊:
影响因子: --
作者:
R. Darling
通讯作者: R. Darling
检测网络效应:对随机实验进行随机化
DOI: 10.1145/3097983.3098192
发表时间: 2017
期刊: Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
影响因子: --
作者:
Martin Saveski;Jean Pouget;Guillaume Saint;Weitao Duan;Souvik Ghosh;Ya Xu;E. Airoldi
通讯作者: E. Airoldi