Professional Documents
Culture Documents
Neil A. Weiss
FORMULAS
NOTATION The following notation is used on this card: • Rule of total probability:
n sample size σ population stdev
k
P (B) P (Aj ) · P (B | Aj )
x sample mean d paired difference
j 1
s sample stdev p̂ sample proportion
(A1 , A2 , . . . , Ak mutually exclusive and exhaustive)
Qj j th quartile p population proportion
N population size O observed frequency • Bayes’s rule:
µ population mean E expected frequency P (Ai ) · P (B | Ai )
P (Ai | B) k
j 1 P (Aj ) · P (B | Aj )
CHAPTER 3 Descriptive Measures (A1 , A2 , . . . , Ak mutually exclusive and exhaustive)
x
• Sample mean: x • Factorial: k! k(k − 1) · · · 2 · 1
n
m!
• Range: Range Max − Min • Permutations rule: m Pr
(m − r)!
• Sample standard deviation:
• Special permutations rule: m Pm m!
(x − x)2 x 2 − (x)2 /n m!
s or s • Combinations rule:
n−1 n−1 m Cr
r! (m − r)!
• Interquartile range: IQR Q3 − Q1 N!
• Number of possible samples: N Cn
• Lower limit Q1 − 1.5 · IQR, Upper limit Q3 + 1.5 · IQR n! (N − n)!
x
• Population mean (mean of a variable): µ CHAPTER 5 Discrete Random Variables
N
• Population standard deviation (standard deviation of a variable): • Mean of a discrete random variable X: µ xP (X x)
(x − µ)2 x 2 • Standard deviation of a discrete random variable X:
σ or σ − µ2
N N
σ (x − µ)2 P (X x) or σ x 2 P (X x) − µ2
x−µ
• Standardized variable: z
σ • Factorial: k! k(k − 1) · · · 2 · 1
CHAPTER 4 Probability Concepts n n!
• Binomial coefficient:
x x! (n − x)!
• Probability for equally likely outcomes:
• Binomial probability formula:
f
P (E) ,
N n x
P (X x) p (1 − p)n−x ,
where f denotes the number of ways event E can occur and x
N denotes the total number of outcomes possible.
where n denotes the number of trials and p denotes the success
• Special addition rule: probability.
FORMULAS
CHAPTER 8 Confidence Intervals for One Population Mean • Pooled t-interval for µ1 − µ2 (independent samples, normal
populations or large samples, and equal population standard
• Standardized version of the variable x:
deviations):
x−µ
z √
σ/ n (x 1 − x 2 ) ± tα/2 · sp (1/n1 ) + (1/n2 )
• z-interval for µ (σ known, normal population or large sample):
with df n1 + n2 − 2.
σ
x ± zα/2 · √
n
• Degrees of freedom for nonpooled-t procedures:
σ
• Margin of error for the estimate of µ: E zα/2 · √
2
n s12 /n1 + s22 /n2
2 2
2 ,
• Sample size for estimating µ: s12 /n1 s2 /n2
2 +
zα/2 · σ n1 − 1 n2 − 1
n ,
E
rounded down to the nearest integer.
rounded up to the nearest whole number.
• Studentized version of the variable x: • Nonpooled t-test statistic for H0 : µ1 µ2 (independent samples,
and normal populations or large samples):
x−µ
t √
s/ n x1 − x2
t
• t-interval for µ (σ unknown, normal population or large sample): (s12 /n1 ) + (s22 /n2 )
s
x ± tα/2 · √ with df .
n
with df n − 1. • Nonpooled t-interval for µ1 − µ2 (independent samples, and normal
populations or large samples):
CHAPTER 9 Hypothesis Tests for One Population Mean
• z-test statistic for H0 : µ µ0 (σ known, normal population or large (x 1 − x 2 ) ± tα/2 · (s12 /n1 ) + (s22 /n2 )
sample):
x − µ0 with df .
z √
σ/ n
• Mann–Whitney test statistic for H0 : µ1 µ2 (independent samples
• t-test statistic for H0 : µ µ0 (σ unknown, normal population or and same-shape populations):
large sample):
x − µ0 M sum of the ranks for sample data from Population 1
t √
s/ n
• Paired t-test statistic for H0 : µ1 µ2 (paired sample, and normal
with df n − 1.
differences or large sample):
• Wilcoxon signed-rank test statistic for H0 : µ µ0 (symmetric
population): d
t √
W sum of the positive ranks sd / n
with df n − 1.
CHAPTER 10 Inferences for Two Population Means
• Pooled sample standard deviation: • Paired t-interval for µ1 − µ2 (paired sample, and normal differences
or large sample):
(n1 − 1)s12 + (n2 − 1)s22
sp sd
d ± tα/2 · √
n1 + n2 − 2
n
• Pooled t-test statistic for H0 : µ1 µ2 (independent samples, normal
populations or large samples, and equal population standard with df n − 1.
deviations):
• Paired Wilcoxon signed-rank test statistic for H0 : µ1 µ2 (paired
x1 − x2
t √ sample and symmetric differences):
sp (1/n1 ) + (1/n2 )
with df n1 + n2 − 2. W sum of the positive ranks
INTRODUCTORY STATISTICS, 6/E
Neil A. Weiss
FORMULAS
CHAPTER 11 Inferences for Population Standard Deviations • Two-sample z-test statistic for H0 : p1 p2 :
• Margin of error for the estimate of p: • Test statistic for a chi-square independence test:
χ 2 (O − E)2 /E
E zα/2 · p̂(1 − p̂)/n
with df (r − 1)(c − 1), where r and c are the number of possible
• Sample size for estimating p: values for the two variables under consideration.
2
zα/2 2 zα/2
n 0.25 or n p̂g (1 − p̂g ) CHAPTER 14 Descriptive Methods in Regression and Correlation
E E
• Sxx , Sxy , and Syy :
rounded up to the nearest whole number (g “educated guess”)
FORMULAS
• Total sum of squares: SST (y − y)2 Syy CHAPTER 16 Analysis of Variance (ANOVA)
• Regression sum of squares: SSR (ŷ − y)2 Sxy
2
/Sxx • Notation in one-way ANOVA:
• Error sum of squares: SSE (y − ŷ) Syy −
2 2
Sxy /Sxx k number of populations
• Regression identity: SST SSR + SSE n total number of observations
SSR x mean of all n observations
• Coefficient of determination: r 2
SST nj size of sample from Population j
• Linear correlation coefficient: x j mean of sample from Population j
1
(x − x)(y − y) Sxy sj2 variance of sample from Population j
r n−1
or r
sx sy Sxx Syy Tj sum of sample data from Population j
• Confidence interval for the conditional mean of the response variable SSTR SSE
MSTR , MSE
corresponding to xp : k−1 n−k
• Test statistic for one-way ANOVA (independent samples, normal
1 (xp − x/n)2
ŷp ± tα/2 · se + populations, and equal population standard deviations):
n Sxx
MSTR
with df n − 2. F
MSE
• Prediction interval for an observed value of the response variable with df (k − 1, n − k).
corresponding to xp :
• Confidence interval for µi − µj in the Tukey multiple-comparison
1 (xp − x/n)2 method (independent samples, normal populations, and equal
ŷp ± tα/2 · se 1 + +
n Sxx population standard deviations):
with df n − 2. qα
(x i − x j ) ± √ · s (1/ni ) + (1/nj ),
• Test statistic for H0 : ρ 0: 2
√
t
r where s MSE and qα is obtained for a q-curve with parameters k
1 − r2 and n − k.
n−2 • Test statistic for a Kruskal–Wallis test (independent samples,
with df n − 2. same-shape populations, all sample sizes 5 or greater):