meaning

[chrysanthemum]

What policy makers ultimately decide to do may depend on how far the results of the recent study are judged to be an accurate reflection of the general situation. Cecilia Reynolds empasizes that gender differences are statistical, with significant numbers of individuals everywhere not following the general trend. I can't understand the meaning of the underlined part. Will you please explain it for me?

 

[Raymott]

What policy makers ultimately decide to do may depend on how far the results of the recent study are judged to be an accurate reflection of the general situation. Cecilia Reynolds empasizes that gender differences are statistical, with significant numbers of individuals everywhere not following the general trend. I can't understand the meaning of the underlined part. Will you please explain it for me?

She means that when we say women do A and men do B, this is only a statistical effect. It might mean that most women do A, or that women do A most of the time, etc.
She says that there are many individuals everywhere who do not what their gender predicts of them (Many women do B, many men do A).

This is a very restricted usage of the word 'statistical'. Even if the two normal curves for men and women overlapped by 90%, there would still be real differences. Reynolds seems to be implying that if there is any overlap of tendencies at all, the results are merely statistical (which can be a code-word for 'not real').

 

[waflob]

The mathematician in me wants to say that the standard deviation is too high to use the results meaningfully. Does that help, or shall I try again?

 

[TheParser]

What policy makers ultimately decide to do may depend on how far the results of the recent study are judged to be an accurate reflection of the general situation. Cecilia Reynolds empasizes that gender differences are statistical, with significant numbers of individuals everywhere not following the general trend. I can't understand the meaning of the underlined part. Will you please explain it for me?

NOT A TEACHER


(1) Your excellent question reminded me of a famous quotation:

There are three kinds of lies: lies, damned lies, and statistics.

(a) NOTES:

(i) When I was younger, "damned" was considered a naughty word. Today it is

allowed to be said and written, but it is still not a good idea to use it in polite society.

(ii) The quotation is attributed to Benjamin Disraeli, one of Queen Victoria's first

ministers.

(2) So I think that the gist (main point) of your sentence is that statistics are not

always accurate. In other words, you can always count things so that the numbers

support what you want to support.

(a) Let's say that you go to one ice cream store and notice that 25 ladies buy

strawberry ice cream, and 75 gentlemen buy vanilla ice cream. You might then

say that your statistics prove that most ladies do not like vanilla ice cream.

But you went to only one store. Maybe if you had gone to more stores, you would

have gotten different results.

(3) Here in the United States, we are often given numbers that are later proven to

be wrong. Certain groups find ways to make up false numbers so that they can then

get money from the government or favorable publicity from the media (newspapers, TV

stations, websites, etc.).

(4) As first minister Disraeli reminded us, statistics are often lies or less than the

whole truth. For example, let's say that country X says that only 1,000 soldiers were

killed in a fight. Maybe country X does not say that 3,000 more soldiers who were

wounded in the fight later died when they returned home. Therefore, the true number

was 4,000 dead, but the government told the "truth" when it reported that "only" 1,000

soldiers had actually died on the battlefield.

(5) Your words in bold, then, mean something like:

According to the collected statistics, there are differences between females and

males in certain areas, but in real life many females and males do NOT behave as the

statistics say they do. So policymakers must be careful not to pass laws based on

faulty statistics. In other words, always be suspicious of numbers. You should always

ask "Who collected those numbers?" "How did they collect those numbers?" "Why

did they collect those numbers?"

 

[SoothingDave]

The mathematician in me wants to say that the standard deviation is too high to use the results meaningfully. Does that help, or shall I try again?

I read it as "please don't abuse me for not being politically correct." The author is statistically showing that men and women are different, which is a sin against PC. So the author is clear to point out that the results only show trends and do not represent each individual.

 

[waflob]

(1) Your excellent question reminded me of a famous quotation:

There are three kinds of lies: lies, damned lies, and statistics.

The (not so famous) quote that came unbidden to my head is this:

"Don't believe any statistics that you didn't fake yourself"

 

[SoothingDave]

83 percent of statistics are made up on the spot.

 

[waflob]

Or the politician who was shocked to read that 50% of the population were under average intelligence and made a speech about improving this dreadful situation

(apologies for slight tangent)

 

[5jj]

83 percent of statistics are made up on the spot.

No, 91.43%

 

[Raymott]

Or the politician who was shocked to read that 50% of the population were under average intelligence and made a speech about improving this dreadful situation

(apologies for slight tangent)

That scares me every time I hear it, no matter how trivially true it is.

In any case, given that 100 is the most common IQ, and that IQs don't come in fractions, either a significant proportion of that 50% are indeed on 100, or the percentage on 99 or below is closer to 49%*. (Put simply, 50% cannot be above and 50% below 100 unless no one is on 100. A statistician can work out the exact percentage below 100).

So the politician wasn't necessarily missing the point entirely - though he was possibly 99% likely to have been.

* Assuming also that intelligence = IQ.

 

[SoothingDave]

I would be failing in my mission to be pedantic if I did not note that half the population is below the median, by definition. Depending on the distribution, the average could be higher or lower than the median.

 

[5jj]

I would be failing in my mission to be pedantic if I did not note that half the population is below the median, by definition. Depending on the distribution, the average could be higher or lower than the median.

See the preceding post from Raymott.

If 50% ( for example) of the population were exactly on the median, then one would assume that only 25% were below, and 25% above.