Successful presentation at BigSurv23: New data sources and signals on presidential approval

Last October, our research associate Bastián González-Bustamante had the honour of presenting his paper entitled “New Data Sources and Signals on Presidential Approval: Daily Indicator Based on Google Trends” at the BigSurv23 conference, held on 26-29 October in Quito, Ecuador. This important international meeting brought together experts in data and social sciences to share ideas and experiences on producing socially relevant statistics.

The BigSurv conference is a unique opportunity for computational and data scientists, as well as experts in survey methodology and statistics, to come together to exchange ideas, collaborate, and learn from each other’s work. The main goal of this conference series is to improve the production of statistics on topics with broader impact and societal relevance.

The paper presented by Bastián González-Bustamante explores the use of innovative data sources to generate daily presidential approval indicators in 12 presidential democracies since the mid-2000s. The author uses a novel sampling approach to generate more stable daily Google search volumes (GSV), aligned and corrected with weekly and monthly series. This allows him to produce daily indicators of presidential approval that can be used to study that can be used to study public opinion in different contexts.

In order to validate his work, Bastián González-Bustamante provided several plausibility checks. First, he used an algorithm to return quarterly estimates of presidential approval using his indicator and a vector of confounding factors, including macroeconomic indicators and dummies, to control for the quadratic effects of presidential approval. Second, it used unpublished data on calls for ministerial resignations produced using optical recognition algorithms on archives and press reports, together with machine learning classifiers.

In addition, Bastián González-Bustamante discussed alternatives for integrating his indicator with survey data to improve measurement. This could be an important step towards creating more accurate and relevant social statistics.

The best thing is that our research associate not only presented a groundbreaking paper, but also received the award for the best early career researcher paper at the BigSurv23 conference. This is an important recognition of his dedication and effort in research.

* AI-generated text
Read more about how we generate our content