Inference Formulas & Conditions Test
Conditions
One-Sample z-test (for a population mean)
1.) known 2.) SRS 3.) population is normal, or large sample size (n ≥ 30) 4.) Independence 1.) 1 and 2 are known 2.) SRS 3.) population is normal, or large sample size ( n1 , n2 ≥ 30) 4.) Samples are Independent 1.) SRS 2.) population is normal, or large sample size (n ≥ 30) 3.) Independence 1.) SRS 2.) population is normal, or large sample size ( n1 , n2 ≥ 30) 3.) Samples are independent
Two-Sample z-test (for a difference in 2 population means)
One-Sample t-test (for a population mean) Two-Sample t-test (for a difference in 2 population means)
Matched Pairs t-test
1.) samples are not independent (they are matched) 2.) SRS 3.) differences are normal or large sample size (n ≥ 30)
Null Hypothesis
Ho: µ = µo
Ho: µ1 = µ2
Test Statistic
Z
normalcdf (lower limit, upper limit)
x o
Confidence Interval
x z*
n
n
z
x1 x2
12 n1
22
x o s Ho: µ = µo n with n – 1 df x –x H 0 : 1 2 t 12 2 2 s1 s2 or H 0 : 1 2 0 n1 n2 With smaller of n1 1 or n2 1 df x o s n with n – 1 df t
normalcdf (lower limit, upper limit)
x1 x2 z
*
12 n1
22 n2
n2
t
H 0 : d 0
p-value
tcdf((lower limit, upper limit, degrees of freedom tcdf((lower limit, upper limit, degrees of freedom
tcdf((lower limit, upper limit, degrees of freedom
* x tn1
s n
(x1 – x 2 ) t *
* x tn1
s n
s12 s22 n1 n2
Test One Proportion ztest
Conditions 1.) SRS 2. ) Independence: pop > 10 times sample 3. ) Normality np 10 n(1 p) 10
Null Hypothesis
H 0 : p p0
Test Statistic pˆ p0 z p0 (1 p0 ) n
p-value Confidence Interval normalcdf (lower pˆ (1 pˆ ) * limit, upper pˆ z n limit) Note: the condition for CI is
npˆ 10 and n(1 pˆ ) 10
Two Proportion ztest
Chi Square Goodness of Fit Test Chi Square test for independence t-test for the slope of a linear regression equation
1.) SRS 2. ) Independence: pop > 10 times sample 3. ) Normality n1 pˆ c 5 n1 (1 pˆ c ) 5 n2 pˆ c 5 n2 (1 pˆ c ) 5 1.) SRS or data is from an entire population. 2. ) all expected counts are ≥5 1.) all expected counts ≥5
1.) y responses are independent. 2.) y responses are normal. 3.) standard deviation about the regression line remains constant. 4.) variables are linearly related
H 0 : p1 p2 or H 0 : p1 p2 0
pˆ1 pˆ 2
z
1 1 pˆ c (1 pˆ c ) n1 n2 x x Note: pˆ c 1 2 n1 n2
normalcdf (lower limit, upper limit)
H 0 : pa p1, pb p2 ,...
(obsv. exp .) 2 exp . df=# of categories - 1
χ²cdf(χ² statistic, 100000,df)
H 0 : there is no association between the variables (independent)
(obsv. exp .) 2 exp . df= (# rows – 1)(# columns – 1)
χ²cdf(χ² statistic, 100000,df)
2
2
t H0 : 0
b SEb
Degrees of freedom = n2
pˆ1 pˆ 2 z*
tcdf((lower limit, upper limit, degrees of freedom
pˆ1 (1 pˆ1 ) pˆ 2 (1 pˆ 2 ) n1 n2
b t *SE b`