Statistics - How to calculate a realistic average

[John K]

For the sake of doing it, I am keeping a log of my blood pressures and then graphing them – I can’t help it, I like graphs and spreadsheets. I'm a nerd...

The problem is there are definite atypical spikes in the Systolic caused by running up the stairs, watching old episode of Baywatch, talking to Linda that are definitely skewing the average.

So I want to drop the ‘outlying fliers’ off before getting the average, in my mind I can sort of see a Bell Curve with the average taken either side of the bell. So get the average, then only sample the figures within +/- 40% of the average.

Problem is, I’m not familiar enough with all the fancy statistical terms in Excel.

Any suggestions? Apart from "get a life"...

[Wobblyprop]

Taking the median is a well used approach to reducing the effect of outliers (used for 'average' income etc).  Easy to calculate, maybe XL has such a function?  PS I have a life too ;-)  PPS see Wikipedia.

[Kevin Wood]

It really depends on the data. Sometimes the outliers are the important bit, other times they are erroneous points you want to eliminate. Without knowing the medical significance of the data we're all guessing.

Having had my BP under observation for the last 15 years or so, the thing that sticks with me is my GP telling me to keep an eye on it "from time to time" and let him know if the situation changes, but not to get obsessed about keeping records, etc.

[Dave Eastwood (Gadgetman) - Club Chairman]

Having had one or two issues over the years, - really quite bad in the mid 2000's; I got a right rollicking g from the doctor for becoming obsessed with the figures, to the extent it gave me, what he referred to as "white coat syndrome".

I'd literally built it into such a big thing in my head, that the minute a doctor or nurse strapped the cuff on me it would rocket through the ceiling. In the end they had to be a 24 hour monitor on me, as the only way they could get reliable readings.

It took ten years or so before I could get my blood pressure measured without it spiking. So while it's interesting to study, if you're any sort of obsessive/worrier like me, just be careful not to focus too much on it!

[John K]

Absolutely agree on all the comments about not obsessing about it. And really I am not.

I just like collecting data and looking for trends.

Actually my BP and pulse is pretty good for an unfit middle aged dude at 125/75/65

I really am just trying to improve my trend analysis and the BP was a convenient data set.

 

[Rory's Dad]

But what if shows you are not average?? If you are average... well you know the answer to that!

[Andrew O Byrne White - Ireland AO]

A series of "if" statements and/or a pivot graph in Excel is probably your best bet. Otherwise, minitab is a really good program for plotting statistical data.  You can get a 30 day free trial of it but I think you have to pay more for anything beyond that.

[jim_l]

Thing is, it is relatively meaningless with all those outliers in, a collection of random measurements not telling you much at all, try to measure something that should be reasonably stable, eg your BP soon after you wake up,  or your BP after 5 minutes sitting at your desk in the afternoon, this will allow you to get a meaningful insight, is it in control, is it going up over time?  This, for example, is likely to show that on days after you have had a few beers your readings will be higher.

Do it the way your GP would ,  Sit down , after a few minutes doing nothing take a reading,  don't even talk! replicate the position , sitting feet flat, with the cuff level with your heart.

Point is, don't measure a whole bunch of different things (Your BP at rest, Your BP after lunch, your BP after running up the stairs, etc) and expect to glean anything useful from it.

Jim

[Kevin Wood]

1 hour ago, jim_l said:

Thing is, it is relatively meaningless with all those outliers in, a collection of random measurements not telling you much at all, try to measure something that should be reasonably stable, eg your BP soon after you wake up,  or your BP after 5 minutes sitting at your desk in the afternoon, this will allow you to get a meaningful insight, is it in control, is it going up over time?  This, for example, is likely to show that on days after you have had a few beers your readings will be higher.

Do it the way your GP would ,  Sit down , after a few minutes doing nothing take a reading,  don't even talk! replicate the position , sitting feet flat, with the cuff level with your heart.

Point is, don't measure a whole bunch of different things (Your BP at rest, Your BP after lunch, your BP after running up the stairs, etc) and expect to glean anything useful from it.

Jim

Indeed GPs always refer to is as resting BP for good reason. It's entirely normal for it to be higher if your body is not resting, and that might not always be obvious.

[SootySport]

As Dave says stop being obsessed with your blood pressure figures,  have a read and contribute to this instead  http://forum.wscc.co.uk/forum/topic/126232-which-beer-have-you-been-drinking-today/