Google Trends and reality: Do the proportions match? Appraising the informational value of online search behavior: Evidence from Swiss tourism regions
Google Trends and reality: Do the proportions match? Appraising the informational value of online search behavior: Evidence from Swiss tourism regions
复制标题
DOI：
10.1016/j.jebo.2017.10.011
复制
发表时间：
2018-01-01
影响因子：
2.2
通讯作者：
Wochner, Daniel S.
中科院分区：
文献类型：
--
作者：
Siliverstovs, Boriss;Wochner, Daniel S.
本研究考察了现实世界的经济活动在多大程度上反映在谷歌搜索的总体在线搜索行为。相对于以前的研究，受到潜在的测量误差问题时，检查搜索查询沿着其纵向尺寸，我们采用了另一种调查方法，利用横截面，而不是纵向的信息内容体现在谷歌的数据。此外，虽然以前的研究通常只研究一个单一的Google趋势系列，但我们的分析是基于60多个不同的系列，这使我们能够评估数据的信息量，不仅在每个系列内，而且在系列之间。最后，我们的Google趋势指数基于最近推出的Google知识图谱技术，可以非常准确地测量相关搜索查询量。我们评估的信息价值的数据为强，半强，或弱的无偏性和效率的考虑因素的基础上，在明瑟Zarnowitz型回归模型。在这里，（瑞士）旅游需求的背景证明是特别有用的，我们发现，基于搜索的旅游需求预测，平均而言，高度准确的近似现实。这表明，基于搜索的指标可以作为指导经济政策的宝贵的实时补充。(C)2017 Elsevier B.V.版权所有。
This study examines the extent to which real-world economic activity is reflected in aggregate online search behavior on Google Search. As opposed to previous studies, being subject to potential mismeasurement problems when examining search queries along their longitudinal dimension, we apply an alternative investigative approach that exploits the cross-sectional instead of the longitudinal informational content embodied in Googles data. Moreover, while previous studies most often examine a single Google Trends series, our analyses are based on over 60 distinct series, which allow us to assess how informative the data are, not only within each series, but also between series. Finally, our Google Trends indices are based on the recently launched Google Knowledge Graph technology, allowing for a remarkably accurate measurement of relevant search query volumes. We assess the informational value of the data as strong, semi-strong, or weak based on unbiasedness and efficiency considerations in a Mincer-Zarnowitz-type regression model. Here, the context of (Swiss) tourism demand proves particularly useful, and we find that search-based tourism demand predictions are, on average, highly accurate approximations of reality. This indicates that search-based indicators may serve as valuable real-time complements for the guidance of economic policy. (C) 2017 Elsevier B.V. All rights reserved.