JenniferMurphy
Well-known Member
- Joined
- Jul 23, 2011
- Messages
- 2,532
- Office Version
- 365
- Platform
- Windows
I am working on a worksheet to analyze blood sugar readings for diabetics. One of the columns contains the actual readings in mg/dL. As diabetic treatment progresses, the goal is to lower blood sugar levels, but also to get the readings to be more consistent (lower the variability). I would appreciate suggestions for the best measure of variability for the purposes of comparing readings for different time periods so as to be able to assess progress.
The table below has 4 sample readings (rows 5-8) over 4 different time periods (columns D-G) showing improvement in the readings. Rows 9-13 show the maximum, average, median, middle, and minimum of these values. Row 14 shows the range, which is one measure of variability. Rows 15 & 167 show the variance and std dev using built-in Excel functions, other measures of variability. I find the variance difficult to fathom. The std dev seems more comprehensible.
Rows 17-19 contain some measures I concocted. Each of them calculate the percentage that one half of one of the measures is of the average or middle values. For example, D17 shows that half of the range (183 / 2 = 91.5) is 27% of the average (338), plus or minus. D18 shows a similar result, but compared to the middle value. D19 does a similar calculation but using the std dev.
I would appreciate any comments or suggestions for better measures.
The table below has 4 sample readings (rows 5-8) over 4 different time periods (columns D-G) showing improvement in the readings. Rows 9-13 show the maximum, average, median, middle, and minimum of these values. Row 14 shows the range, which is one measure of variability. Rows 15 & 167 show the variance and std dev using built-in Excel functions, other measures of variability. I find the variance difficult to fathom. The std dev seems more comprehensible.
Rows 17-19 contain some measures I concocted. Each of them calculate the percentage that one half of one of the measures is of the average or middle values. For example, D17 shows that half of the range (183 / 2 = 91.5) is 27% of the average (338), plus or minus. D18 shows a similar result, but compared to the middle value. D19 does a similar calculation but using the std dev.
I would appreciate any comments or suggestions for better measures.