Considering there's no trend line indicating any downward movement and this is quite a sharp shift, my gut says outlier. The poll being summer polling strengthens that thought.
However, we'll really know (for sure) in the next couple of months.
53-51-48-43 isn't a downward trend? It wasn't downward before April, but that doesn't look like a healthy sign to me.
You can't just start at one point in the cycle and tell me a trend exists. Look further.
Consider that the five months before your numbers have hovered between 47%-49% and his disapproval ratings from August of last year have hovered between 41%-44% throughout (with the exception of January, which looks like an obvious outlier). That's pretty stable.
Thus, it appears like for quite a long while, Coleman's numbers hovered around 48% approve, 42%-43% disapprove.
Taken into context, the 53% and 51% would have originally seemed like possible data points that his numbers were going up. The 48% number would have maybe indicated a down month or that the 53% and 51% were merely indicating outlier/upper end of the MOE scale.
With the 43% number, we now have a couple of interpretations - 1) trend line downward (your theory); 2) outlier or edge of MOE (interpretation I'm leaning towards).
Unfortunately, it will take a little while to work the interpretation out, as all of this is quite fluid. For example, if next month is 49%, it would be likely that my interpretation is correct, but not necessarily so, as later months could establish that 49% is the new high of the MOE. Or, for example, he could get 45%, which would further establish your theory, but not necessarily so, as later months could establish that the 45%/43% numbers were merely lower ends of MOE (like the 53% and 51%, for example). Or, if next month's number were, say, 52%, it would lend further credence that this number is a true outlier and that the trend to his numbers is
actually going higher.
Add to this the problem of summer polling and we have some real fun at hand.
Remember, all of this could just be MOE movement, since the MOE is +/- 5%. Or there may be movement.
