Power quality data is often reported using statistical confidence levels. This will exclude the most extreme data for a certain length of time depending on the interval over which the confidence level is applied. There is considerable conjecture as to the effect of applying statistical measures over different time intervals, e.g. several days, weeks or one year. If statistical confidence levels are applied over long intervals, the length of time not included in the statistical confidence interval is long. During such intervals disturbance levels may be continuously high and not be accounted for in the statistical parameter. This study investigates the effect different methods of aggregating data to a specific reporting period will have on the calculated index. Several data processing methods are trialled to evaluate the effect of using different aggregation intervals to produce an index to characterise disturbance levels for the whole year.