Exploring How Sentiment Measures Interacts with the Data Produced by the GCP

Abstract: This paper explores how the random numbers produced by the Global Consciousness Project (GCP) covary with various sentiment measures. Previous findings suggests that variations in such measures, conceptually, also should covary with the GCP data. The hypothesis, that the GCP data covaries with sentiment, is tested using time series statistics and by correlating several well-known quantitative daily sentiment measures with daily GCP data aggregates. Results supportive of such correlations and thus the hypothesis are found. The results motivate an even further exploration on how the GCP data interacts with, predicts, and affects various sentiment measures and its derivatives.  

Introduction: The random numbers produced by the Global Consciousness Project (GCP) have been shown to react strongly to events of perceived global importance (Nelson & Bancel, 2011; Nelson et al., 2002; Radin, 2002; Radin et al., 2006). Results suggesting that the GCP data significantly correlates with seemingly unrelated data sources such as stock market returns and internet search trends have also been found (Holmberg, 2020; Holmberg, 2021; Holmberg, 2022). What could cause the GCP data to react to events of perceived importance, such as stock market movements and internet search trends? Arguably, the observed effects are mostly unrelated, but it is acknowledged that they could come to be if they are caused by some unseen cofounder. One such unseen cofounder is global sentiment, and notably, the literature already implicitly suggests that such a link exists. It is well known that the economy at large, and equity prices in particular, are highly affected by sentiment (Howrey, 2001; Fraiberger, 2017; Shiller, 2017). Furthermore, internet search trends are strongly related with the perceived importance of the events searched for, and thus affected by sentiment directly. In this paper it is hypothesized that many of the correlations found originate from how sentiment covaries with the GCP data. It is also acknowledged that such a hypothesis can be tested for by correlating sentiment with variations in the random numbers produced by the GCP.