StATS: Always ask why (May 8, 2006)
I have a three year old boy at home and he's learned that one way to keep the conversation going with an adult is to simply ask the question, "Why?"
I'll say "We're going to church this morning" and he'll say "Why?" I'll say, "Because it's Sunday" and he'll say "Why?" At this point, I'm stumped. Why exactly is it Sunday today and not Tuesday.
Or in the morning I'll point out that it's raining outside and he'll say "Why?" And I have to struggle with an answer like, "When there is too much moisture in the air, it falls down to the ground in the form of precipitation."
At work when people ask me to do something, I need to emulate my little boy and and ask them why. Not in a hostile way, but to get them to talk some more so I can find out exactly what they want.
Yesterday, someone asked me to compute a standard deviation.
Would you be willing to look at this attached chart and give me your advice on how I can calculate the standard deviation of the amount of drug (dose) that a patient may receive due to standardization compared to what is requested, please? For example, if a prescriber wants to order 15 mg/kg of acetaminophen and the dose is standardized. What is the standard deviation from what was ordered on average? Does that make sense?
She sent me a table showing standardized dosing for acetaminophen. For children, the dosage is usually a certain amount per kilogram of body weight. But the medicine itself comes in a discrete form, such as drops or tablets. The typical dose is 10-15 mg/kg, so a child weighting between 12 and 22 kg might get 2ml drops where represents a target dose of 200 mg, but at the extremes of the weight range could represent a dose as small as 180 (12*15) or a dose as large as 220 (22*10). A child weighing 27.5 to 57.5 kg could get a tablet, and this represents a standardized dose of 500 mg, but it could be as small as 412.5 or as large as 575. There are 12 rows in the spreadsheet just like this.
Now I could probably figure out how to do something that would produce a degree of deviation from the norm or standard, but there's a very good chance that I will produce something that is not quite what this person wants. So I wrote back and asked politely why do you need a standard deviation? It's just a question to get someone to talk more about what their goals are.
I try to do this a lot. If they want to match subjects based on age and race, I'll ask them why they think matching is important. If they respond, I thought you had to match or the study wouldn't be valid, then we can talk about the various ways of ensuring validity of a research trial. Asking why always seems to start a productive dialog.
I have a great quote on one of my web pages which is worth repeating here because Dr. Cox explains this far better than I could:
The statistician who supposes that his main contribution to the planning of an experiment will involve statistical theory, finds repeatedly that he makes his most valuable contribution simply by persuading the investigator to explain why he wishes to do the experiment, by persuading him to justify the experimental treatments, and to explain why it is that the experiment, when completed, will assist him in his research. -- Gertrude M. Cox.
This web page was written by Steve Simon, edited by Steve Simon, and was last modified on 04/01/2010.
This page was written by Steve Simon while working at Children's Mercy Hospital. Although I do not hold the copyright for this material, I am reproducing it here as a service, as it is no longer available on the Children's Mercy Hospital website. Need more information? I have a page with general help resources. You can also browse for pages similar to this one at Category: Human side of statistics.