Key points are not available for this paper at this time.
When a set of observations form a time series it is wrong to use the ordinary random sampling theory for calculating standard errors. Dr Jowett discusses the problem and shows what should be done.
G. H. Jowett (Tue,) studied this question.