Measurement Error in Dependent Variables in Accounting: Illustrations Using Google Ticker Search and Simulations

deHaan, Ed; Lawrence, Alastair; Litjens, Robin

doi:10.2139/ssrn.3398287

Cited by 13 publications

(6 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Depending on how researchers use these classified variables in subsequent statistical models, measurement shortcomings from independent classification could lead to biased and/or unreliable inferences. Indeed, to the extent that the latter classification approach leads to higher measurement error in one's post-classification independent (dependent) variable(s), inconsistency (inefficiency) in one's regression estimates-and a higher corresponding risk of bias and incorrect inferences when said measurement error arises in either one's independent or variable(s)-can arise (Wansbeek and Meijer 2000;deHaan, Lawrence, and Litjens 2019).…”

Section: Multi-label Frameworkmentioning

confidence: 99%

Multi-Label Prediction for Political Text-as-Data

Erlich

Dantas

Bagozzi³

et al. 2021

Polit. Anal.

View full text Add to dashboard Cite

Political scientists increasingly use supervised machine learning to code multiple relevant labels from a single set of texts. The current “best practice” of individually applying supervised machine learning to each label ignores information on inter-label association(s), and is likely to under-perform as a result. We introduce multi-label prediction as a solution to this problem. After reviewing the multi-label prediction framework, we apply it to code multiple features of (i) access to information requests made to the Mexican government and (ii) country-year human rights reports. We find that multi-label prediction outperforms standard supervised learning approaches, even in instances where the correlations among one’s multiple labels are low.

show abstract

Section: Multi-label Frameworkmentioning

confidence: 99%

Multi-Label Prediction for Political Text-as-Data

Erlich

Dantas

Bagozzi³

et al. 2021

Polit. Anal.

View full text Add to dashboard Cite

show abstract

“…Announcement period Google search volume ( AISVI_EA ) is the average abnormal Google search volume during the announcement period. My result is robust to using log‐transformed abnormal Google search volume (the same measure as AISVI2 in deHaan et al (2019)). This data set is available from 2010.…”

mentioning

confidence: 80%

“…14 14 Google search volume is a commonly used proxy for retail investors' information acquisition (Da et al 2011; H. Stice 2022). I thank the authors of deHaan et al (2019) for sharing the data of the daily Google's Investor Search Volume Index (ISVI), which captures searches for finance and investing topics more closely. Following prior literature (Drake et al 2012; deHaan et al 2019), I construct daily abnormal Google search volume, calculated by subtracting the average ISVI on the same weekday over the prior 10 weeks, scaled by the average ISVI on the same weekday over the prior 10 weeks (the same measure as AISVI in deHaan et al (2019)).…”

mentioning

confidence: 99%

“…I thank the authors of deHaan et al (2019) for sharing the data of the daily Google's Investor Search Volume Index (ISVI), which captures searches for finance and investing topics more closely. Following prior literature (Drake et al 2012; deHaan et al 2019), I construct daily abnormal Google search volume, calculated by subtracting the average ISVI on the same weekday over the prior 10 weeks, scaled by the average ISVI on the same weekday over the prior 10 weeks (the same measure as AISVI in deHaan et al (2019)). Preannouncement Google search volume ( AISVI_PRE ) is calculated as the average abnormal Google search volume during the preannouncement period.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Information Coordination: Does Preannouncement Media Coverage Improve Earnings Expectations?*

Xu¹

2022

Contemporary Accting Res

View full text Add to dashboard Cite

This study examines whether and how media coverage improves price discovery before earnings announcements. Given the broad reach and timeliness of media coverage, understanding its role in forming earnings expectations is important. I posit that the media specialize in searching for and screening news, while institutional investors and financial analysts specialize in interpreting and integrating information. I expect these coordinated efforts to improve earnings expectations and preempt upcoming earnings announcements. Consistent with this prediction, I find (i) preannouncement media coverage is negatively associated with price revaluations during earnings announcements, (ii) preannouncement media coverage is positively associated with institutional investors' information acquisition and financial analysts' forecast frequency, and (iii) institutional information acquisition and analyst forecast activity are two mechanisms by which media coverage preempts upcoming earnings. My main results are robust to short‐window tests and identification concerns. Overall, my findings suggest that media coverage coordinates information efforts and accelerates the incorporation of information in prices, facilitating preannouncement price discovery.

show abstract

“…Lawrence, and Litjens (2019), retail attention is measured only for stocks whose tickers have three or more letters to reduce measurement errors. FollowingBen-Rephael et al (2017), we measure abnormal institutional attention (AIA i ) as a dummy variable, equal to 1 if INST_ATTN i equals 3 or 4, and 0 otherwise.…”

mentioning

confidence: 99%

Retail Attention, Institutional Attention

Liu¹,

Peng²,

Tang³

2022

J. Financ. Quant. Anal.

View full text Add to dashboard Cite

We document distinctly different clientele effects on investor attention and return responses to information. Macro news crowds out retail investor attention to firms’ earnings news by 49%. For stocks with high retail ownership, macro news dampens earnings announcement returns by 17% and substantially increases post-announcement drift, especially during high VIX periods. In contrast, macro news increases institutional investor attention to scheduled earnings announcements but not their attention to unscheduled analysts’ forecast revisions. The findings confirm the implications of rational inattention models and highlight the importance of considering clientele effects in understanding the effect of news on attention and asset prices.

show abstract

Measurement Error in Dependent Variables in Accounting: Illustrations Using Google Ticker Search and Simulations

Cited by 13 publications

References 18 publications

Multi-Label Prediction for Political Text-as-Data

Multi-Label Prediction for Political Text-as-Data

Information Coordination: Does Preannouncement Media Coverage Improve Earnings Expectations?*

Retail Attention, Institutional Attention

Contact Info

Product

Resources

About