.67% is a “significant correlation between customer satisfaction and insurer profitability”??? That’s ridiculous. What I see are some insurance companies blatantly misrepresenting underwriting factors in commercial advertisements. Those that pretend like they are charities are doomed to poor combined ratios. Business is business.
I looked at the press release and it does not say “percent,” just that there is the correlation. It was probably referring to a correlation coefficient, and presumably the author added the percent part. This would be wrong because .67 would obviously then be 67%, not .67%, and it’s not what a correlation coefficient even means. It could also be referring to the adjusted r-squared if they used regression analysis, but that’s unlikely based on their wording, and as worded here would be a little inaccurate.
A .67 correlation coefficient would be a strong correlaton. Generally anything over .5 is considered strong, with 0 being no correlation, and the scale being -1 to 1. Of course correlation does not prove causation, but it seems likely here.
This author generally writes decent articles but as I have noticed and pointed out before does not have a good grasp on statistics. Sadly this is more and more prevalent in the insurance industry.
.67% is a “significant correlation between customer satisfaction and insurer profitability”??? That’s ridiculous. What I see are some insurance companies blatantly misrepresenting underwriting factors in commercial advertisements. Those that pretend like they are charities are doomed to poor combined ratios. Business is business.
I looked at the press release and it does not say “percent,” just that there is the correlation. It was probably referring to a correlation coefficient, and presumably the author added the percent part. This would be wrong because .67 would obviously then be 67%, not .67%, and it’s not what a correlation coefficient even means. It could also be referring to the adjusted r-squared if they used regression analysis, but that’s unlikely based on their wording, and as worded here would be a little inaccurate.
A .67 correlation coefficient would be a strong correlaton. Generally anything over .5 is considered strong, with 0 being no correlation, and the scale being -1 to 1. Of course correlation does not prove causation, but it seems likely here.
This author generally writes decent articles but as I have noticed and pointed out before does not have a good grasp on statistics. Sadly this is more and more prevalent in the insurance industry.
Gee, I wonder why AIG was not ranked in the higher performing companies.
AIG is like a used car salesman, they sell you something and don’t recognize you later when you need something.
Nice Article.