kitchen table math, the sequel: CDC writing lesson & expert raters using holistic rubrics

Tuesday, October 28, 2014

CDC writing lesson & expert raters using holistic rubrics

I've been meaning to correct the flu-shot post & am doing so now because Linda Seebach has given me a nudge.

When the CDC writes:
"Over a period of 30 years, between 1976 and 2006, estimates of flu-associated deaths in the United States range from a low of about 3,000 to a high of about 49,000 people."
...What they mean is:
Over a period of 30 years, between 1976 and 2006, estimates of annual flu-associated deaths in the United States range from a low of about 3,000 to a high of about 49,000 people.
It's funny. I think of writing as 'fuzzy' compared to math, but lately I've realized there's a precision to writing, too.

The first sentence means 3,000 over 30 years.

The second sentence, the one the CDC should have written, means 3,000 per year, in some but not all years.

I still question universal flu shots (although I certainly don't object to people getting flu shots if they like), especially given the fact that flu shots aren't particularly effective. Flu vaccine isn't like measles or mumps vaccine; the composition of each year's shot is an informed guess as to which viruses will be making the rounds this year.

As I think I mentioned in the earlier post, I recall a time when health authorities recommended flu shots only for vulnerable populations, not for the entire country. Vulnerable populations included the elderly (for whom vaccine isn't as effective), small children, and, I think, people with respiratory ailments (hence: flu-associated deaths).

Reading between the lines on the CDC page, I wonder whether the universal flu-shot campaign is intended to reduce exposure of vulnerable populations to the flu by reducing the incidence of influenza in non-vulnerable populations.

The elderly don't respond particularly well to flu vaccine.

Their best protection comes from other people not getting the flu, and thus not exposing them.

Thinking about flu shots, I'm reminded of my friend in LA who refused to get flu shots because the only time she ever lost any weight was when she got the flu.

Speaking of precision

Speaking of precision, I'm reading a fascinating study of college writing quality as defined by "expert raters ... using a holistic rubric."

(And, yes, that construction does give me pause. Expert raters using a holistic rubric are the people who decided that my SAT essay on freedom and macroeconomics was worse than my son's SAT essay citing Martin Luther King.)

A major finding: good writing, to an expert rater using a holistic rubric, isn't any more cohesive than bad writing.

Cohesion, pretty much the sine qua non of good writing outside of college, has no bearing one way or another on an expert rater.

Instead, what matters are:
  1. Using lots of words before getting to the main verb (more is better)
  2. Using lots of different words
  3. Using low-frequency words (aka big words
If you've spent any time in academia, you already know this, and it's been a conundrum for me, speaking as an instructor of freshman composition.

Good writing as defined by composition textbooks and style guides isn't good writing as defined by the people grading papers.


gasstationwithoutpumps said...

I agree that CDC's omission of "annual" from their sentence changed the meaning. Of course, the writers who popularize the scientific findings are often not very adept at understanding what they write about, so they may not have understood the original data properly.

Although flu vaccines are less effective in the elderly, whose immune systems are less responsive, there are high-dose flu vaccines available for those 65 and older, which compensate somewhat for the lower immune response.

3000 to 49000 deaths a year in the US is significant. The massive bicycle helmet campaigns (which are more expensive per person than flu shots) are targeted at something that has a death rate of about 800 a year, and bike helmets are less effective than flu vaccines.

Disclaimer: I get a flu shot and I wear a bicycle helmet when I bicycle. Both provide enough additional protection to be worth the small amount of inconvenience and cast for me.

gasstationwithoutpumps said...

I grade a lot of writing in my college classes, but I don't use a rubric (I've never found one that worked—they never seem to capture what is important in any particular piece of writing).

I generally ding students for inflated diction, for using many different words for the same concept, or for burying the main point deep in the sentences. But I'm grading technical papers, not the sort of free-form "essays" that seem to be the only form of writing the students have ever previously done.

I generally end up having to stress focus and flow as the two main weaknesses of student writing. I usually suggest that students rely on two heuristics: topic sentences at the start of all paragraphs (with the whole paragraph supporting that topic sentence) and old information➔new information structure for sentences.

Auntie Ann said...

It often is about non-vulnerable people getting vaccinated. Japan did a study a while back where they made a major push to immunize kids for the flu. As a result, the number of elderly ending up in hospitals because of flu-related illness went way down.

Small children are vectors.

Anonymous said...

"It often is about non-vulnerable people getting vaccinated."

Giving the HPV vaccine to boys would be a good example of this.

-Mark R.

allison said...

The imprecision in writing is often because the writer has no idea what the math they have stated means, so they have no sense of its ludicrousness.

A blog on my local paper (the Mpls Star-Tribune) reported that here in MN, "odds of a deer related crash are 1 in 89, up from 1 in 81." It went on to state that it was very likely to havw a deer accident.

Readers corrected that 1 in 89 is down from 1 in 81.

But no one noticed the obvious problem: 1 in 89 whats?

It is absurd to suggest that 1 in 89 "drives" results in a deer related car accident. 1 in 89 miles is even more absurd.

So maybe what was meant was 1 in 89 car crashes is a deer related crashes.

Less than 2% of car crashes are deer related. Not exactly a large fraction. In fact, given that most of us don't have a car crash a year, odds are you won't have a deer related crash at all.

The precise language depended on understanding the precise math.

Anonymous said...

"The imprecision in writing is often because the writer has no idea what the math they have stated means, so they have no sense of its ludicrousness."


Which I find is often a failure to be specific about *UNITS*.

Your example of the odds of a deer crash moving from 1 in 81 to 1 in 89 is an excellent example of this. 1 in 81 per driver per year? Or 1 in 81 crashes involve a deer? Or what?

A perennial favorite of mine is when the press compares the total wealth of a person (in, say, dollars) with the GDP of a small country (which would be dollars-per-year). YOU CAN'T COMPARE NUMBERS WITH DIFFERENT UNITS MEANINGFULLY!!!!! But people do ...

-Mark Roulo