Can data with different sampling frequency be added to calculate a simple average? Supposed that data is measured at every 5 minute interval for first 6 months (X1) and every hour interval for next six months (X2), can you add all the data and divide them by total count to find an average? Or do you have to adjust for the sampling frequency to get a common basis for all data?
My colleage says that data collected at every 5 minutes should be first scaled to hourly data (by taking average of all 12 samples) and then added with the average of hourly data (X2) to calculate the average. He says that the data with 5 minutes inteval will skew the average if not adjusted.
Just based on my hunch, since we are dividing the sum by total number of data, sampling frequency shouldn't matter, but I am a novice when it comes to Statistics.
Good question that deserves an answer by an expert !
Peace
-----------------------------------------------------------------
I am a Gents and not a Ladies.
The answer would depend on how sensitive the process is and if that sensitivity is critical to your needs when determining the average.
In other words, process stability is a must before capability is determined. If the process fluctuates heavily in the 5 minute time interval, that lack of stability or fluctuation will not be captured when collecting data at 1 hour intervals and hence your average will not accurately reflect the reality.
-----------------------------------------------------------------
Dimple2001
Seems like time of recording has lot of importance in the given data.
One of the way will be to take reading of only 12th interval into consideration (taken after every hour), ignoring readings from other11 intervals, from first six months data, thus matching the sampling pattern with next six months and then calculate the yearly average. In a way I am suggesting to adjust the frequency of sampling.
Btw , Difference Between Average of first six month and Average of next six month will tell the whole story of accuracy and method to be chosen. It will also clarify on few points raised by Dimple.
Advertise Contact Us Privacy Policy and Terms of Usage FAQ Canadian Desi © 2001 Marg eSolutions Site designed, developed and maintained by Marg eSolutions Inc. |