# Confidence interval for absolute changes [duplicate]

I have used the STDRATE procedure to get the estimates and standard errors for 5 different timeperiods between 1980-2000 in 5 years interval (1980, 1985, 1990, 1995 and 2000).

This means that I have point estimate and 95% CI for $X_1, X_2, X_3, X_4$ and $X_5$ which spans between the years 1980-2000.

For example i have the estimates with 95%Cl for

X1 = 22 (18 to 25),

X2= 30 (24 to 33)

X3= 14 (12 to 16)

X4= 10 (8 to 11)

X5= 6 (2-7)

Now if I wanted to calculate the absolute change with point estimates and 95% CI between the start of the study (i.e for X1) and the end (i.e X5), how do I do this?

A similar study have done this and this is how it looks, their variables are a1, a2, a3, a4 and a5.

This is their data. a1= 4.2 (3.8 to 4.5),

a2= 3.7(3.3 to 4.0),

a3= 2.4(2.3 to 2.5)

a4= 1.9(1.8 to 2.0)

a5= 1.5(1.4 to 1.6).

Their absolute change was -2.7 (-3.0 to -2.4)

I would preciate any help that I can get!

• There is not enough information in here for an answer! You should really state your real problem, with the application context, detailed. Commented Jul 21, 2014 at 9:42
• Do you think that a simple t-test works to compare the rates for first and last 5-years interval, and if so is this a valid method? Commented Jul 21, 2014 at 10:49