Problem- Q1) If the background is constant for your sample, why bother to subtract it from each reading prior to regression analysis? Under what circumstances would this effect ne less/more noticeable?
Q2) Suppose the half life of your sample is 3 minutes. You have 2 independent variables to assign - firstly, the intervals at which you collect the data - secondly, the interval of time that you use to collect data. Obviously, the second independent variable cannot easily be larger than the interval chosen for initiating the count. Suppose that you chose to make measurements at 10 minute intervals, although still retaining the 40 seconds collecting period. Would this be less/more accurate than the method used (one minute intervals and 40 second counting periods)?
Q3) Now suppose that you choose 5 minutes as your measurement interval, but elect to use 4 minutes as your counting period. During this 4 minutes used to accumulate counts the sample counting rate will change markedly due to decay. Will this affect your final answer for the half life? What would happen if you elect to measure the activity at 2 seconds intervals, using 1 second to accumulate counting data? Would this method provide you with a better representation of the countingrate, and thus a more detailed histogram for calculation purposes?
Q4) Suppose now that your inherently erratic behavior intrudes, and you measure activities at non-equal intervals. Does this matter? Why? What happens if you collect counting data at carefully assigned elapsed times, but you use different integrating periods to determine the count rate? Are the measurement intervals and the counting period's truly independent variable?
As much detail as probable is really appreciated