Database Reference
In-Depth Information
Figure 7-19. A new parameter to select desired confidence level
Next, he needed to create a number of calculated fields to generate the error bars:
▪
% in Chess Club
, p = [In Chess Club]/([In Chess Club]+[Not in Chess Club])
▪ Sample Size,
n
= [In Chess Club]+[Not in Chess Club]
▪
Standard Error
= SQRT(([% in Chess Club]*(1-[% in Chess Club]))/[n])
▪
z upper
= CASE [Confidence level]
— WHEN 0.99 THEN 2.575829
— WHEN 0.95 THEN 1.959964
— WHEN 0.90 THEN 1.644854
▪
Margin of Error
= [Standard Error]*[z upper]
▪
Lower limit
= [% in Chess Club]-[Margin of Error]
