Measurement Uncertainty Forum

Confidence interval of a mean

Picture of Roman V.
Confidence interval of a mean
by Roman V. - Tuesday, January 2, 2018, 8:50 AM


I have following problem:

There is a process in which I would like to determine exposure of sample irradiation (in Lux*hours). I measure illuminance every 10 min over 24 hours, getting 144 data points. There is some fluctuation in light source, but due to integral properties of a mean, one can calculate average from these 144 measurements, multiply by 24 hours and get exposure (or estimate of the surface under illumination curve over time).

The 95% confidence interval of the mean in this case will be +/- 1.96*SEM. (SEM -- standard error of a mean).

BUT, there is the uncertainty of an illumination measurement at every point. How should I use this information for calculating the CI of a mean?