So these results should be treated with caution & may differ from estimates derived from a more sophisticated weighting procedure. But the fact that other surveys (including the state-run VTsIOM) show similar declines is telling and increases our confidence in the pattern. /END

We caution that although these data are weighted by sociodemographic characteristics (to make our online sample more closely approximate the Russian population in socio-demographic terms) there were some changes in sampling methods by our online provider over the period. 6/

Consistent with this explanation, we find that Putin's approval dropped more among those respondents who rely most on the internet as a source of news. 5/

Strong approval declines and weak approval increases noticeably. By September the president's approval ratings bounced back quite a bit, but are still much lower than they were at this time last year. 3/

The drop is barely noticeable when looking at aggregate approve/disapprove ratings from our survey (first fig.) but can be seen when breaking down the data into *degrees* of approval/ disapproval along the full Likert scale (second fig.) 2/