Number of Samples Outoome Variable
One sample Continuous
'IWo independent samples Continuous
Two dependent, matched samples Continuous
Opgsamp e _W_,*-Mmous.m
Two independent samples Dichotomous
estimate and builds in what is called a margin of er
Conﬁdence Intervals for Two Independdli =3.-
In the formulas in Table 6—8, X and 22 are the means of the
outcome in the independent samples, 2 or t are values from the
z or t distributions reﬂecting the desired conﬁdence level,
1 1 _
and SI, — + — IS the
ibis 95% CI for the d
based on the . use
mum Wider than the one 1e size produces a very impre
s ' 0d pressures.
' fleot: the diﬂ‘erence in mean syStOllC blo
esama C .
6 5 CONFIDENCE INTERVALSMIEOR MATC
SAMPLES, commuous ourco
. ' d in t
mace, we often do not know the value of the' -
wiﬁdﬂd deviation (0'). If the sample size is large ( r11) (:33,
aw, the sam P19 standard deviation (5) can be used to estimate
th _ ulaﬁon standard devratlon. Note that the prior derivation
die . , d on the C
generate 95% CI estimates for
tion and in the populatlons o
preceding formula, we need to satisfyr
criterion—speciﬁcally, we need at leas
‘ 1e, .
fail in each sample. In this exam? 7 ore th
CVEIZSnd a failure is freedom from CVD. Here W3 m
mph 6.6. In Example 6.2, we recapped the data pre-
senﬁad'm chapter 4 on the subsample of n = 10 participants
Who attended the seventh examination of the offspring in the
Framinghm Heart Study. Table 6—1 1 contains descriptive sta-
tisthS on the
. 100d Pressure
CI for SYStOhC b ,
crate a 95%amingh3m offsprlng Study.
we use the following fonnula.
We wiSh to gen .
using data collected in the PI
Because the sample size 15 large,
‘ '- = .96. Substituting the
The 2 value for 95% conﬁdence i
Conﬁdence Intervals for-0n!
ute descriptive statistics on the sample data using the
, ,ues described in Chapter 4. When the outcome of
Fe st i5 dichotomous, we record on each member of the sam-
lntere ether they have the characteristic of interest or not.
LEARNING OBJ ECTIVES
By the end of this chapter, the reader will be able to estimate the own parameter. The S,
. Deﬁne point estimate, standard error, conﬁdence level, and resentatlve 0fthe POPUlatiOIl;Wit-l§1 Pa
margin of error
. Compare and contrast s