opinion polls, Michael martin questions the methadologies behind them

techman1 · 04-11-2023 11:43AM #1

Just heard an interview with Michael Martin on newstalk Anton Savage show. He was being asked about the latest opinion polls on FF but Michael Martin was questioning the whole methadologies behind opinion polls in Ireland. He stated that they have been way out in comparison to election results. He called into question why discussion of the latest opinion polls is so prominent in discussions yet those polls turn out to be completely wrong subsequently. In fairness he had Anton Savage in the back foot with that as Anton and most other people don't understand the methadologies either.

I think he was calling for european oversight or something into the methadologies and how those polls are carried out

04-11-2023 11:48AM

I don't think he would be so concerned with the methodology of polls if FF were topping them.

KaneToad · 04-11-2023 11:50AM

The methodologies for the opinion polls (carried out by the mainstream companies) are very clearly stated.

The real problem is that people don't understand the science beyond sampling & opinion polls. There are still many that believe that a sample of 1,500 can't be accurate for the whole population of Ireland. Most lay people think that there is a linear relationship between population size and sample size.

There isn't. Sampling theory is sound. All the opinion polls state their methodology, quotas achieved, confidence intervals/margin of errors etc. What can be misleading is the questions asked of the sample, leading questions can be misused...

"How worried would you be if SF were in govt?"

"How much do you hate Leo Varadkar?"

Exaggerated examples, but you get the point!

techman1 · 04-11-2023 12:56PM

There isn't. Sampling theory is sound. All the opinion polls state their methodology, quotas achieved, confidence intervals/margin of errors etc. What can be misleading is the questions asked of the sample, leading questions can be misused...

So thats where the issue is then, the leading questions leading the person in certain direction before they answer the actual question for the poll. I think "Yes Minister" briilantly exposed this 40 years ago. Clearly something is wrong when the polls are way out compared to actual election results

cute geoge · 04-11-2023 01:25PM

A fair share of them opinion polls are conducted online from the same panel of people over and over again.The majority doing these online oponion polls would naturally be from poorer working class , unemployed or students presume .So mr.martin has a point but I still dont believe ff will get any improvement who ever is surveyed

Danzy · 04-11-2023 01:28PM

https://www.boards.ie/discussion/comment/121324109#Comment_121324109

No one takes note of those polls though.

The ones that Micheal is terrified by are the ones that are carried out to the highest industry standard.

techman1 · 04-11-2023 01:37PM

https://www.boards.ie/discussion/comment/121324109#Comment_121324109

if thats the case well then the sampling is wrong, they are not taking a correct sample of the population but a skewed sample?

Podge_irl · 04-11-2023 01:59PM

All polling (pretty much) will standardise or weight the responses. How this is done can vary the results quite a lot.

Eg they may ask you your opinion and also how you voted last time. If they end up with 45% FG voters they weight the their opinions in the poll down to the 25% or so they actually got in the election. It's not an exact science.

techman1 · 04-11-2023 02:44PM

https://www.youtube.com/watch?v=6GSKwf4AIlI

great clip from Yes Minister about how leading questions from pollsters can lead people in a certain direction in a poll

Cookie_Monster · 04-11-2023 03:59PM

Heard most of it, interesting point made, rightly or wrongly.

I think he was right though that there is too much focus on the outcome of X working with Y etc rather than talk about actual policies and plans that each party has and critiquing them within the media to a larger degree. It dumbs down the discussions too much by focusing on the potential outcomes rather than the actual policies

opinion polls, Michael martin questions the methadologies behind them

Comments