Well that can easily be switched just by the variables... for example: isolation factor (ie, lockdowns) vs depression rates. When isolation factor decreases (l

Well that can easily be switched just by the variables... for example: isolation factor (ie, lockdowns) vs depression rates. When isolation factor decreases (lockdown releases), depression rates increase. Negative correlation. Social contact (ie, lockdown releases) vs depression rates. When lockdowns release and social contact increases, depression rates increase. Positive correlation. If it was straightforward coordinate plane values like to describe the slope of a line, you can't really easily reverse the correlation just by redefining the variables.