Effect of Systematic Errors on the Calculated Error Estimate

Effect of Systematic Errors on the Calculated Error Estimates. The systematic errors (see Table 2 and p. 39) will in some cases affect the standard deviations 6 and o. is, as mentioned, usually calculated from equation (95). Systematic errors will leave the elements (B )i essentially unchanged. However, at least if the estimate (96) is used for o, some systematic errors will change Oo and thus a since the agreement between experimental and theoretical intensities may become poorer. The effect is often small, since the parameters will refine to slightly wrong values which compensate for the error and reduce the increase in the residuals. 

The scale errors (see p. 39), will leave the standard deviations and tr unaltered if the error is the same for ail the plates. This is often nearly true if only plates from one experiment are used. However, if the plates have been recorded separately and with difiFerent nozzle-to-plate distances, the standard deviations (especially ct ) will partly reflect these errors. 

The other systematic errors given in Table 2 may increase because the agreement between experiment and theory becomes somewhat poorer. An especially large value for uo [equation (96)] and thus for is found if the background is wrong. The effect on varies. Thus an error in the blackness correction will, of course, not be reflected in if all the plates used have nearly the same density. Other systematic errors will only increase o if results from different nozzle-to-plate distances are compared. 

---