It's an observational study, not an experimental one: "A cohort of human patients previously taking CBD had significantly lower SARS-CoV-2 infection incidence of up to an order of magnitude relative to matched pairs or the general population".
If the study is purely observational they could be measuring ‘COVID-19 mostly affects old people and old people don’t use these products’
that's similar to what was said about covid and tobacco smokers from mid-2020. I wonder if a correlation exists to this CBD population and a tendency toward confirmation bias.