More on representativeness

J Clive Matthews
16 October 2007
Professors Fishkin and Luskin in the European Parliament

Below follows hastily-recorded partial transcripts / paraphrases (so don't take this as their exact words) of some of the comments made by Professors Fishkin and Luskin on the representativeness of the Tomorrow's Europe deliberative poll, both at the opening press conference and in conversation later - but all from before the demographics of the attendees were made available. So, do the attendance demographics justify their claims?

Fishkin: We take a scientific sample, the 3500 of all of Europe and then we randomly select some to be invited. We randomly invite in proportion to the representation in the European Parliament... We can compare all the people who come and all the people who don't come... This is something none of the other methods of consultation have...

The 3,500 were a good microcosm of Europe. The several hundred who come will also be a good microcosm and we will be able to judge it by that benchmark. The next question is who is in the room? How did they get there? What did they do?

Our aim with the deliberative poll is to see how ordinary citizens can come to good information... Many people have found other methods of citizen consultation where they have to come to a conclusion to lead to inequalities and lead to extremes... We have worked out a way to consult the public that does not distort public opinion...

Luskin: What you have seen in the audience is a true random sample of all of Europe. This means we have people of all walks of life... people from every country, people speaking every language.... It's expensive, which is why it's not done more often on this scale. This is the first.

No matter what polling method you use for a given project you only have one sample, so you have to consider what would happen if you did repeated sampling - what are the probabilities? Take any demographic, take a random sample from the population as a whole and repeat, gradually the sample averages will come to resemble the population perfectly.

If you set up quotas for your sample, what criteria should you select? That will create a selection bias just by the criteria you choose. Within a population as large and complex as Europe's how can you select appropriate criteria for quotas? Some groups will always miss out, and the quotas will by their very nature heighten the unrepresentativeness of those taking part.

Had enough of ‘alternative facts’? openDemocracy is different Join the conversation: get our weekly email


We encourage anyone to comment, please consult the oD commenting guidelines if you have any questions.
Audio available Bookmark Check Language Close Comments Download Facebook Link Email Newsletter Newsletter Play Print Share Twitter Youtube Search Instagram