Interpreting the result of a Bayesian Structural Times Series model

A bit of context: I am reading a study called "Exploring the determinants of Bitcoin Price". In this research paper google search trends across different countries over time are used, in part, to determine price movements in Bitcoin. They use a Bayesian Structural Times Series model.

I am struggling to understand this description of their results:

"The marginal posterior mean and HDI for Colombia is 0.092 [0.075, 0.105], that is, 1 standard deviation change in the searches for “Bitcoin” in google from this country is associated with almost 0.1 standard deviations change in Bitcoin’s price."

Is this the same as saying, on average, a one unit increase in "Bitcoin" google searches, ceteris paribus, will result in a 0.1 increase in Bitcoin price?

Any help in clarifying this would be greatly appreciated!

I attached a cropped image of their results table if that helps...


Last edited:


Less is more. Stay pure. Stay poor.
Do you know how to interpret results from linear regression? If so this is the exact same. They standardized data (both the independent and dependent variables), so a unit equals a standard deviation. A unit increase in IV (a std) result s in a 0.1 change in DV (which is also is in standard deviation).
Last edited: