Although Meg had slept a fair bit the day before, she had a good night's sleep and seemed cheerful in the morning with the carers. But Thursday is my 'shopping' day and although I raced around the store as fast as my legs would carry me, Meg suffered an anxiety attack shortly after I left the house. The relatively new carer did a good job in calming Meg down until I arrived when I took over, as it were, whilst putting the shopping away. I was not looking forward to today because it was one day of the year when I need to go have my eyes tested to ward off incipient diabetes and this involves putting drops that sting in both eyes to dilate the pupils before a high definition camera takes a photo of the back of each eye. This procedure is fairly short and evidently non invasive but the drops in one's eyes keeps the vision blurred for hours afterwards which I find disconcerting. Very kindly, our Irish friend from down the road gave me a lift both to the surgery where the investigation was to be conducted and then brought me home again for which I was immensely grateful. Fortunately, Meg had been quite calm during my second absence of the day and my son took some time off to sit with her. Then we discovered some excess curry which made for an instant meal.taken a little late, after which we watched the third episode of 'Tess' on BBC catch up.
The news media today has been full of the transition arrangements between Biden and Trump and Joe Biden is promising that there will be a smooth transition of power (which Trump did not offer to Biden by the way) Arousing a kind of fascinated horror is the role in the new government might be given to the maverick Robert F Kennedy Jr, who has a history of repeating debunked claims, including linking vaccines with autism in children. What precise role he will be given in government remains unclear but one of his pledges was to remove all fluoride from drinking water 'on Day 1'
The day after the election in the US, there is always a certain amount of soul searching for those of us who have followed these events closely and are looking for some answers. One of the principal questions that arises is to wonder how the pollsters managed to get the result so wrong and not for the first time. Although the polls regularly showed Kamala Harris to be 1.0-1.5 percentage points ahead of Trump, the actual result is that Trump won by a margin of 3.5 percentage points - in other words a 5% difference between what the polls had been saying and the actual result. Upon consulting some internet sources, I did ascertain that there were at least two factors that seemed to emerge. One of these I alluded to just the other day and that is that Democrat voters are more likely to divulge their opinions whereas conversely, Republicans are less likely to do so. Another factor seems to be that the opinions of non-educated white males seemed to be under-represented and this was just the group most likely vote for Trump. I found a political scientist who had really put his finger on the problem and he wrote as follows: 'The key thing is people made the same mistakes they did in 2016. They understated the Trump voter who is less likely to be engaged politically, and crucially, more likely to be busy, not spending 20 minutes talking to pollsters… people working a pretty common job or, as the case of many Hispanic voters, juggling two or three jobs at a time. Was it underestimating Trump support? Overestimating Democratic turnout? Maybe both of those explanations, at their root, have the same cause: polling firms are increasingly bad at contacting less politically engaged voters. And unlike the talking heads which make up the political media (myself included!), most voters are not highly engaged with politics.Whichever way those voters skew, politically, is far less likely to be picked up accurately by a pollster. Selzer, for example, famously refuses to contact voters via any method besides a live call — in an age of seemingly nonstop spam calls, it is a method that just feels inherently dubious.Whatever the case, it is clear that some pollsters and, more importantly, the political media needs to get a lot better at talking to low-propensity voters and people who are tuned out from the political news bubble.'
So here we have a clue to the eventual shortfall but the rest of the explanation lies in the method of conducting a poll. As I used to teach survey and research methods, I was well aware of the strengths and weaknesses of different methods of polling and all I had to remember to do was to remember some of the lessons that I used to teach. First we start off with a poll conducted in which those caught in the sample are selected by a random process using, for example, a random number generator. This will give results that are likely to be quite accurate and to which a body of statistical theory can be applied. But random samples are expensive and take a long time to conduct. It is axiomatic that only the person indicated by the random number should be part of the sample - choosing the person next door, for example, would destroy the principle of randomness. Because they are time consuming and expensive, despite being the 'gold standard' in polls random sampling methods are not used. What is used is a quota sample in which the known characteristics of the population (age, sex, ethnicity, socioeconomic status etc.) are applied and then interviewers are asked to select any individuals that fall within the quota. For example a quota sample of 1000 individuals might suggest you needed to select 100 female white collar workers aged 20-29. A quota sample, once completed, will by design be representative of all of the ages, genders, class etc.built into the sample design. How if you were to ask a survey assistant to collect the views of 100 young white collar females, they might put themselves in the city centre where there are likely to be a lot of o young office workers perhaps popping out for a sandwich. Similarly, to collect the views of older male manual workers, you might locate yourself in an industrial part of the town near to the factory zone. The trouble is with the whole of this methodology is that a sample of using female white collars will not reflect the views of those few white collar workers working in the suburbs nor the male manual workers working in the city centres.
© Mike Hart [2024]