Delo ran a new poll yesterday, a month since pengovsky started following polls prior to autumn parliamentary elections. This poll is the first one done in a post-Clean Shovel enviroment and in this respect it is mildly surprising, to say the least.
Namely, there is no visible effect of the pre-dawn raids on construction powerplayers and supposed tax-evaders. What we are seeing, though, is a substaintial headway Zares seems to have made over LDS. How or why this happened is at this point a bit of a mistery – unless you are willing to subscribe to the theory that a battle for the legacy of the late Janez Drnovšek just ensewed and Zares won the first round.
In the duel between SD and SDS, things are interesting as well. As noted already, there was no spectacular gain on the part of SDS which we might be able to attribute to Operation Clean Shovel. So the rise in support for SDS can be considered “organic” in my view. The PM is on the news a lot lately (presiding the EU and all) and correspondingly the frequency of Borut Pahor seeing himself on the telly has fallen dramatically – there is only so much politics you can squeeze in the main news bulletin 🙂 On the other hand, Social Democrats are still bleeding support. Not at an alarming rate anyhow, but their strategists should not worry about their losses, but about gains SDS has made. Namely: if we took only Delo’s polls of yesterday and of a month ago, we see that SDS made substantial gains while SD kept on losing, even though it still kept the top spot.
Finally, we can now take a look at the entire month of February. As polls were made by different pollsters on different samples, the graph below is not exactly accurate, but it can give you a general idea of what is going on: SDS and SD continue breathing down each other’s necks, while the rest of the political gang fight for the leftovers – but there’s enough of that to go around as even those parties which are currently below the parliamentary treshold (the red dashed line) will – as things stand now – make it across to the safe territory. A notable exception to the rule is Zares, which pushed ahead of the “general populatation” and is loitering a coupe of percentage points above the rest of the parties, but is far far below the leading duo.
Pengovsky’s projection: I think we might still get to see a slight bump in SDS’s ratings as a result of the anti-corruption sweep, but it will not last long. But barring a major embarasment, the rise in support of the ruling party will continue, at least as long as the EU presidency lasts. Then the shit will hit the fan, and it the race will go down to the wire. Until then, however, the nine political parties (Lipa included, as it was formarlly established on Saturday) will want to secure the best possible starting positions.
Since the results go up and down from poll to poll, you can not see any trend but the fact that SD and SDS are about 20 %, that Zares has about 8 % share and the rest is around the 4 % border. The biggest party is still the »Don’t know« party.
The difference that any party can do is mostly only to get additional votes from the biggest party and not from voters from other parties. So Pahor and others have to think about how to get these votes. Pahor obviously by being the good fellow! But will this concept gain results? In good half year we will know!
WTF is Zares??? I have no clue… Is this comparable to Haiders BZÖ? An offspring of a liberal party with an orange background? Za novo politiko, nove vrednote ali oživitev pozabljenih is quite similar to wert- und zukunftsorientierte Tradition. A ne?
There’s something else I don’t understand: In Germany the Sonntagsfrage (see Wikipedia) is common. There they don’t itemize the dont-knowers, but try to predict the results of an election, if it would be the next sunday. Isn’t this concept more appropriate?
@Davor: I beg to differ… Firstly (as I noted in the post) the longitudal graph is a collection of different polls done on different samples and therefore not exactly accurate, but you can see the general direction in which parties’ ratings are headed. Especially in the case of SDS you can see that the party’s curve ends much higher that it started, whereas SD’s curve ends only marginally higher. Plus, the relative difference between the two parties is much smaller at the end of the graph than at the beginning. Conclusion: SDS is on an upward, while SD in on a downward trend.
Secondly: the undecided vote (don’t know) is exceptionally low in the latest poll, possibly indicating a high tendency to vote which might result in a high turnout (but it is waay to early to tell for sure). But as a certain percentage of people will definitely not vote, a mere 20+ percent of the undecided vote tells you that most people have made up their mind. Or – to put it more bluntly: a substantial number of the undecideds will not vote in any case, meaning that the results of parties will rise proportionally, putting everyone above the 4% treshold.
However, the trend is the only thing we can define for certain at this moment. Everything else is a more or less educated guess. In a month’s time we’ll have more polls by the same pollsters and we’ll be able to compare them more accurately.
@Robert: I think Gregor Golobič would be hugely insulted by the comparisson to Haider 😀 Perhaps this post will shed a bit more light on the origins of the party 🙂
As far as the Sonntagsfeage (which party would you vote for if the elections were held next Sunday) is concerned: To the best of my knowledge, this is the way it is done. But not knowing is not the same as not going to vote (although a substantial number of the undecideds will at the end stay home on election Sunday)
However, the percentage of the undecideds does show how much (if any) room for expansion parties have. According to the latest poll (21% undecided) there is almost no room for impovement for any party save stealing votes from other parties. So this is a rather vital piece of information, because it does suggest a future course of action.
Concerning the Sonntagsfrage: I was convinced that there’s a huge amount of calculation – or as some say: manipulation – between the raw data like yours above and results published like this one. But it’s difficult to find anything more concrete… There’s only this citation: Zwischen dem, was wir an Rohergebnissen erhalten und dem, was wir als Prognose publizieren, liegt manchmal eine Differenz von zehn oder elf Prozent. Elisabeth Noelle-Neumann – Between the raw data we obtain and prediction we publish there’s a difference of ten or eleven percent.
Admitedly, statistics is not my forte, but obviously a lot depends on the sample and the question – even variations of the same questions.
There fore you will get substantially different results if you ask a) which party would you vote for if the elections were held next Sunday or b) which of these parties would you vote for if the elections were held next Sunday (list only a few parties). Even reversing the order of parties in the question is likely to give you a different result.
And obviously, predicting the result is a tad more tricky, as the percentage detected does not exactly correlate with the number of seats in the parliament (the system prefferes larger parties) as well as the fact that a lot depends of the actual turnout, especially in a strongly polarised society like Slovenian.
I have read your post and I am aware that each sample is a different own story, but you know samples give results within a margin +- …%. Regarding all differences together ALL results in a number that is only an approximate number. The higher the number the lower is the +- margin.
When you will take your number in the MS Excel shit and draw a linear regression line you will get a result (I am sure) that will not show a trend that will exceed the +- margins. This was my point.
Secondly, the nonvoters consist of voters who in the polls said they are not going to vote and of people that said they are GOING TO VOTE. So the “don’t know” people are not the same as the nonvoters in elections.
And not everybody who says in the polls that he is going to vote a party than really sticks with this party.
I was looking for that function when I was doing above graphs, but couldn’t find it – must have been wee hours of the morning that were getting to me.
I’ll include the linear regression line when the next poll comes out and if you’re right, I’ll take that into account in futute posts.