You may have noticed that we're not paying much attention here at the moment to political polling. The reason for that is that since the polling machines started winding back up, the results they've produced have been a bit odd. Not necesarrily wrong mind you, but just odd. We're seeing Satisfaction/Approval ratings and Better/Preferred PM changes not being particularly consistent with voting intention changes and some of the relationships not being particularly consistent with history.
one of these odd things with a difference popping up between respondent allocated preferences and 2010 election preferences for instance. Another couple of examples can sort of be done visually, where we can look at how the government's two party preferred and primary vote leads stack up against the PM net satisfaction and Better PM ratings for the first 3 polls after every government re-election since 1993 (so we knock out 1996 and 2007 as they were new governments, and we can't use 1998 because of the One Nation effect stuffing up the polling - as it always does with these things, which is why I always leave it out).
First, the table of the data being used (it's Newspoll data where the early years are my calculations of TPP from election preferences, as Newspoll didn't publish many TPP results back then)
Now the charts: