Let me count the ways

In today's column, Polly Toynbee writes:

The Audit Commission estimates that £42,000 on effective early interventions in children's lives from birth to adolescence spares £153,000 in incarceration.

There are a number of problems with this sentence, and here is a guide to them:

The Audit Commission estimates(1) that £42,000 on effective early interventions in children's(2) lives from birth to adolescence spares(3) £153,000(4) in incarceration(5).

(1) It should actually be "estimated", not "estimates".  When Polly first wrote about this back in March 2004, she actually had the good grace to use the past tense of "estimated".  Now, over two years later, writing about exactly the same thing, we've suddenly switched to the present tense.  Eh?

Oh, and if you doubt that it is the same thing, here is a sentence from her 2004 article [empahsis added]:

The commission estimated that spending £42,000 on early interventions from birth through adolescence would spare £153,000 on subsequent incarceration.

It is eerily familiar, isn't it?  Apart from the altered tense...

(2) "Children's lives".  Guess the sample size on which this broad generalisation isbased.  No, go on, guess.  It is one.  One!  And hardly a representative one either.  If you read the original research and their description of James (name has been changed), you'll see that:

• James is 15.
• He lives in his mother’s house; however, she is rarely there.
• His older step-sister is also living there. She is a known drug user with previous convictions.
• His father occasionally visits and is violent and disruptive when he does.
• James has been excluded from special school.
• He is not receiving any alternative educational provision.
• He was ten when he received his first caution.
• He is currently serving his second custodial sentence.

I would hesitate to draw broad conclusions about "children's lives" based on one example, and on this one in particular.  And the Audit Commission did not try to.  In fact, they speculated about the benefits of spending on early intervention in a particular child's life.

(3) The word "spares" here suggests that spending the £42,243 would automatically save the £153,687.  Of course that is not guaranteed on the basis of a counter-facutal involving one person.  Which is why in their research the Audit Commission uses phrases like:

"assuming crime route is avoided" [emhasis added]

"It is not possible to accurately estimate the costs of a young person’s involvement in crime and the impact on the lives of themselves and others."

"Had the services been available, James might have benefited from family support, pre-school education, anger management, learning support and mentoring." [emhasis added]

This is not to say that early intervention programmes don't work, but rather that the claim that they can save so much is pure conjecture.

(4) The actual figure is £153,687, so even with rounding this should be £154,000.

(5) No.  Look at the original research.  The £153,687 does not refer to incarceration alone.  It includes a special education needs assessment by the LEA, compiling an education 'package', social services undertaking a family assessment, and so on.