Comments on Cyprus Elections & Politics: Our Prediction for Round 1

I think the assumption you make, that the age grou...

2008-02-15T13:33:00.000Z

I think the assumption you make, that the age group 18-25 within Cyprus will behave *similarly* to the same age group that will land from abroad is fundamentally wrong. It ignores an important dimension, education, which has been shown to divide the vote amongst the candidates.

Thus, for your assumption to hold, it must be the case that the vast majority of the 18-25 electorate that were polled within Cyprus are college students at the University of Cyprus (at least). Otherwise, I see no reason why college students from Greece or the UK or other countries will behave in similar ways. Whether we like it or not, education IS an important factor for these elections and you are completely assuming it away.

CyBC polls? Why not give us the link to the presid...

2008-02-12T13:11:00.000Z

CyBC polls? Why not give us the link to the presidential palace and cut out the middleman!

aceras

thanks for your comment. Like mr Lo...

2008-02-11T13:15:00.000Z

aceras

thanks for your comment. Like mr Lodros we are also a bit skeptical about the weakness of telephone surveys.

The point you raise and your suggested analysis are very interesting and useful. Unfortunately, the only polls that were based on a ballot than telephone interviews is the one by CyBC and in our sample we only have one such survey so any comparison between telephone and ballot box polls is impossible with the available data.

But, if it means anything the CyBC poll (when excluding the undecided votes) gave:

Tasso 34.03, Christofias 33.48 , Kasoulidis 30.19. i.e. the results from this poll for all 3 candidates fall within our proposed statistical margin of error.

You can find the result of the CyBC in the list we provide in the blog entry below this one.

Ο Λόρδος στον Πολίτην της Κυριακής εκφράζει κάποιε...

2008-02-11T13:01:00.000Z

Ο Λόρδος στον Πολίτην της Κυριακής εκφράζει κάποιες επιφυλάξεις για το αποτέλεσμαν των τηλεφωνικών δημοσκοπήσεων. Απο την ανάλυσην σας φκαίννει κάποια διαφορά μεταξύ τηλεφωνικών δημοσκοπήσεων τζαι δημοσκοπήσεων κάλπης;

Antonis

thanks for your comment. Indeed the...

2008-02-09T23:44:00.000Z

Antonis

thanks for your comment. Indeed the issues you point are likely to have happened. But our analysis of the poll results shows that there were only very small fluctuations from poll to poll and from week to week. For example we also averaged the first 12 and the last 11 polls and the differences were also quite minimal to what we predict.

Your question gives us an opportunity to present some more results from our meta analysis and we will do that soon by updating this post. Check back soon.

I would just like to make a comment, not from a st...

2008-02-09T23:40:00.000Z

I would just like to make a comment, not from a statistical point of view, but merely common sense, regarding the methodology used: you are unifying all 23 polls. However, in the whole of your sample, some people changed their minds since then. Additionally, others might have been asked twice. Unless you consider this a very small number of people, but then again, is this not a mere hypothesis?

Excellent work guys! I also believe that the resul...

2008-02-09T22:18:00.000Z

Excellent work guys! I also believe that the result will be very close to your prediction!