I accept that changes are easier to estimate than absolute values, but even here there is a problem.

]]>“When I [Cliff Zukin] first started doing telephone surveys in New Jersey in the late 1970s, we considered an 80 percent response rate acceptable, and even then we worried if the 20 percent we missed were different in attitudes and behaviors than the 80 percent we got. Enter answering machines and other technologies. By 1997, Pew’s response rate was 36 percent, and the decline has accelerated. By 2014 the response rate had fallen to 8 percent.”

The sophisticated contemporary methods pollsters use now make more detailed assumptions because extrapolation to the population requires information that isn’t present in the sample, so they are forced to use their best guesses just to get anywhere at all.

I agree that we can’t expect reliable poll findings that are predictions, but I think we can get reliable information about *changes* in public opinion. That is, I think the bias in poll results will be relatively stable and will cancel out when calculating differences (contra Andrew Gelman, who thinks differential nonresponse explained some big shifts in the polling in the last US election).

]]>Regards and seasons greetings, Dave

]]>Thank you for your views on your talk. Your response I mostly agree with. In future, I will state that mathematical modelling isn’t a field exclusive for mathematics, but it is a good field for them. In my professional experience, my job is to come in when mathematical models have gone wrong or the maths has been to complex for the model to be built. However, I must acknowledge the help of the SME who I questioned and interview, before I turn their ideas into mathematics. As for who is the right candidate to be a mathematical modeller? Well, this is a person who embraced a wide field of mathematics. In modelling – it a free for all, any method along as it is robust, logical and efficient works. A Mathematical modeller must have the confidence to challenge models that do not reflect reality and the ability to produce a model beyond a theoretical abstract world.

Best regards

Nira ]]>

Thanks again.

]]>