Bert Bates wrote:Got an off the wall question for you: When considering college freshman, Stats 101 kinds of classes, what are the handful (2-4?), topics or concpets that students have the hardest time grokking?
Well, in my case (although I only took stats to 'A' Level; and
that back in Eocene times) it was almost always understanding the
wording of the question in statistical terms - specifically as regards (a) the
independence of each action involved, and (b) important information that you may be missing.
Back when I was working at the Census Office, we came up with this remarkable statement:
If you get cancer of the lip, chances are you will live 5 years
longer than the average person.
And this wasn't just based on sample data; it was a blanket statement covering ALL deaths in England & Wales for which lip cancer was a primary or secondary cause.
And of course, while the statement as a statistic is absolutely true (at least it was then; maybe not now), it hides a pile of information that puts it in context:
1. Cancer of the lip is quite rare.
2. It is (or was) most common among pipe and cigar smokers, who tend to be older.
3. It develops slowly.
So, the likelihood is that by the time it becomes a problem, you're already around the average age to die anyway; and once it is, your life expectancy is actually lower than someone of a similar age with other types of cancer - like leukemia, which tends to be much more lethal to the young than the old.
So my advice: read the questions
well, and don't trust government statistics.
Winston