Researchers increasingly use digital trace data from online platforms as an alternative or complement to survey data. To collect the data, researchers frequently rely on Application Programming Interfaces (APIs) provided by private companies. The APIs often return samples of the data based on undisclosed or intransparent sampling procedures and algorithms. Previous research identified issues with some APIs’ reliability over time and across API versions. For instance, downloading Google Trends data for identical parameters (i.e., search term, region, time range) but at different time points can give researchers different values on Google Trends’ search index. Users and tweets in the samples varied depending on the former Twitter API’s version used to draw the samples. In this paper, we extend the research on the reliability of digital trace data from APIs by examining the effect of the download location on inconsistencies across samples: Do we get different values from digital trace data APIs depending on where we download the data? We retrieve samples from Google Trends, YouTube, and the News API from four different countries on three continents (Austria, Germany, the U.S., and Australia) for the same query parameters (i.e., search term, region, and time range). We then compare the samples retrieved from each respective download country, keeping all parameters of the query constant. Our results point to inconsistencies across download locations, and thus another limitation regarding the reliability and replicability of findings from digital trace data. They serve as a cautionary tale for social science research relying on APIs that provide samples of digital trace data as the download location might impact the findings. Our results also help researchers working with digital trace data from APIs in making their research better replicable by drawing several samples if possible and reporting transparently where they retrieved the data from.
