Formula Sheet BB113 Statistics and its Applications

A. Notation
𝑛 = sample size 𝜇 = population mean
𝑋̅ = sample mean 𝜎 = population standard deviation
𝑆 = sample standard deviation 𝑝̂ = sample proportion
𝑁 = population size 𝑝 = population proportion

B. Numerical Descriptive Measures

Sample mean
∑𝑋
𝑋̅ =
𝑛

Sample variance
2 (∑ 𝑋)2
2
∑(𝑋 − 𝑋̅)2 ∑ 𝑋 − 𝑛
𝑆 = =
𝑛−1 𝑛−1

Sample standard deviation

(∑ 𝑋)2
∑(𝑋 − 𝑋̅)2 √∑ 𝑋 2 − 𝑛
𝑆 = √𝑆 2 = √ =
𝑛−1 𝑛−1

Z score
𝑋 − 𝑋̅
𝑍=
𝑆

Covariance
∑𝑋∑𝑌
∑(𝑋 − 𝑋̅) ∑(𝑌 − 𝑌̅) ∑ 𝑋𝑌 − 𝑛
𝑐𝑜𝑣(𝑋, 𝑌) = =
𝑛−1 𝑛−1

Coefficient of correlation
𝑐𝑜𝑣(𝑋, 𝑌) ∑ 𝑋𝑌 − 𝑛𝑋̅𝑌̅
𝑟= =
𝑆𝑋 𝑆𝑌 √(∑ 𝑋 2 − 𝑛𝑋̅ 2 )(∑ 𝑌 2 − 𝑛𝑌̅ 2 )

C. Basic Probability
Probability of occurrence Independence
𝑋 𝑃(𝐴|𝐵) = 𝑃(𝐴)
Probability of occurrence =
𝑇

Marginal probability General multiplication rule

𝑃(𝐴) = 𝑃(𝐴 𝑎𝑛𝑑 𝐵1 ) + 𝑃(𝐴 𝑎𝑛𝑑 𝐵2 ) + ⋯ 𝑃(𝐴 𝑎𝑛𝑑 𝐵) = 𝑃(𝐴|𝐵)𝑃(𝐵)
+ 𝑃(𝐴 𝑎𝑛𝑑 𝐵𝑘 )

General addition rule Multiplication rule for independent events

𝑃(𝐴 𝑜𝑟 𝐵) = 𝑃(𝐴) + 𝑃(𝐵) − 𝑃(𝐴 𝑎𝑛𝑑 𝐵) 𝑃(𝐴 𝑎𝑛𝑑 𝐵) = 𝑃(𝐴)𝑃(𝐵)

Conditional probability Marginal probability using the general

𝑃(𝐴 𝑎𝑛𝑑 𝐵) multiplication rule
𝑃(𝐴|𝐵) = 𝑃(𝐴) = 𝑃(𝐴|𝐵1 )𝑃(𝐵1 ) + 𝑃(𝐴|𝐵2 )𝑃(𝐵2 ) + ⋯
𝑃(𝐵)
+ 𝑃(𝐴|𝐵𝑘 )𝑃(𝐵𝑘 )
𝑃(𝐴 𝑎𝑛𝑑 𝐵)
𝑃(𝐵|𝐴) =
𝑃(𝐴)

Page 1 of 5
 Formula Sheet BB113 Statistics and its Applications

D. Discrete Probability Distributions

Expected value, 𝝁, of a discrete variable
𝜇 = 𝐸(𝑋) = ∑ 𝑥𝑃(𝑋 = 𝑥)

Variance of a discrete variable

𝜎 2 = ∑(𝑥 − 𝜇)2 𝑃(𝑋 = 𝑥) = ∑ 𝑥 2 𝑃(𝑋 = 𝑥) − 𝜇 2

Standard deviation of a discrete variable

𝜎 = √𝜎 2 = √∑(𝑥 − 𝜇)2 𝑃(𝑋 = 𝑥) = √∑ 𝑥 2 𝑃(𝑋 = 𝑥) − 𝜇 2

Binomial distribution Poisson distribution

𝑛 𝑒 −𝜆 𝜆𝑥
𝑃(𝑋 = 𝑥|𝑛, 𝑝) = ( ) 𝑝 𝑥 (1 − 𝑝)𝑛−𝑥 𝑃(𝑋 = 𝑥|𝜆) =
𝑥 𝑥!

Mean of the binomial distribution Mean of the Poisson distribution

𝜇 = 𝐸(𝑋) = 𝑛𝑝 𝜇 = 𝐸(𝑋) = 𝜆

Standard deviation of the binomial distribution Standard deviation of the Poisson distribution
𝜎 = √𝜎 2 = √Var(𝑋) = √𝑛𝑝(1 − 𝑝) 𝜎 = √𝜆

E. Continuous Probability Distribution

Z transformation formula Finding an 𝑿 value associated with a known
𝑋−𝜇 probability
𝑍=
𝜎 𝑋 = 𝜇 + 𝑍𝜎

F. Sampling and Sampling Distributions + Confidence Interval Estimation

Population mean Finding 𝒁 for the sampling distribution of the
∑𝑋 proportion
𝜇= 𝑝̂ − 𝑝
𝑁 𝑍=
√𝑝(1 − 𝑝)
𝑛

Population standard deviation Confidence interval for the mean (𝝈 known)

𝜎
∑(𝑋 − 𝜇)2 1 (∑ 𝑋)2 𝑋̅ ± 𝑍𝛼⁄2
𝜎=√ = √ [∑ 𝑋 2 − ] √𝑛
𝑁 𝑁 𝑁

Standard deviation of the mean Confidence interval for the mean (𝝈 unknown)
𝜎 𝑆
𝜎𝑋̅ = 𝑋̅ ± 𝑡𝛼⁄2 with df = 𝑛 − 1
√𝑛 √𝑛

Finding 𝒁 for the sampling distribution of the Confidence interval estimate for the proportion
mean
𝑝̂ (1 − 𝑝̂ )
𝑋̅ − 𝜇𝑋̅ 𝑋̅ − 𝜇𝑋̅ 𝑝̂ ± 𝑍𝛼⁄2 √
𝑍= = 𝑛
𝜎𝑋̅ 𝜎 ⁄√ 𝑛

Page 2 of 5
 Formula Sheet BB113 Statistics and its Applications

̅ for the sampling distribution of the

Finding 𝑿 Sample size determination for the mean
mean 𝑍𝛼2⁄2 𝜎 2
𝜎 𝑛=
𝑋̅ = 𝜇 + 𝑍 𝑒2
√𝑛

Sample proportion Sample size determination for the proportion

𝑋 𝑍𝛼2⁄2 𝑝(1 − 𝑝)
𝑝̂ = 𝑛=
𝑛 𝑒2

Standard error of the proportion

𝑝(1 − 𝑝)
𝜎𝑝̂ = √
𝑛

G. Fundamentals of Hypothesis Testing: One-Sample Tests

𝒁 test of hypothesis for the mean (𝝈 known) 𝒁 test of hypothesis for the proportion
𝑋̅ − 𝜇 𝑝̂ − 𝑝
𝑍𝑆𝑇𝐴𝑇 = 𝑍𝑆𝑇𝐴𝑇 =
𝜎 ⁄√ 𝑛 √𝑝(1 − 𝑝)
𝑛

𝒕 test of hypothesis for the mean (𝝈 unknown) 𝒁 test of hypothesis for the proportion in terms
𝑋̅ − 𝜇 of the number of events of interest
𝑡𝑆𝑇𝐴𝑇 = with df = 𝑛 − 1 𝑋 − 𝑛𝑝
𝑆 ⁄√ 𝑛 𝑍𝑆𝑇𝐴𝑇 =
√𝑛𝑝(1 − 𝑝)

H. Analysis of Variance
Total variation in one-way ANOVA Mean squared in one-way ANOVA
(computing formula)
𝑆𝑆𝐴
(∑ 𝑋)2 𝑀𝑆𝐴 =
𝑐−1
𝑆𝑆𝑇 = ∑ 𝑋 2 −
𝑛
𝑆𝑆𝑊
𝑀𝑆𝑊 =
Among-group variation in one-way ANOVA 𝑛−𝑐
(computing formula)
𝑆𝑆𝑇
𝑀𝑆𝑇 =
𝑇𝑗2 (∑ 𝑋) 2 𝑛−1
𝑆𝑆𝐴 = ∑ ( )−
𝑛𝑗 𝑛
where 𝑇𝑗 = sum of data from group 𝑗

Within-group variation in one-way ANOVA One-way ANOVA 𝑭𝑺𝑻𝑨𝑻 test statistic

𝑆𝑆𝑊 = 𝑆𝑆𝑇 − 𝑆𝑆𝐴 𝑀𝑆𝐴
𝐹𝑆𝑇𝐴𝑇 =
𝑀𝑆𝑊
with df1 = 𝑐 − 1, df2 = 𝑛 − 𝑐

Page 3 of 5
 Formula Sheet BB113 Statistics and its Applications

I. Simple Linear Regression

Simple linear regression equation: The prediction line
𝑌̂𝑖 = 𝑏0 + 𝑏1 𝑋𝑖

𝑺𝑺𝑿, 𝑺𝑺𝑿𝒀 and 𝑺𝑺𝒀

(∑ 𝑋)2
𝑆𝑆𝑋 = ∑(𝑋 − 𝑋̅)2 = ∑ 𝑋 2 −
𝑛
∑𝑋∑𝑌
𝑆𝑆𝑋𝑌 = ∑(𝑋 − 𝑋̅)(𝑌 − 𝑌̅ ) = ∑ 𝑋𝑌 −
𝑛

(∑ 𝑌)2
𝑆𝑆𝑌 = ∑(𝑌 − 𝑌̅)2 = ∑ 𝑌 2 −
𝑛

Computing formula for the slope, 𝒃𝟏

𝑆𝑆𝑋𝑌
𝑏1 =
𝑆𝑆𝑋

Computing formula for the 𝒀 intercept, 𝒃𝟎

𝑏0 = 𝑌̅ − 𝑏1 𝑋̅
where
∑𝑋 ∑𝑌
𝑋̅ = and 𝑌̅ =
𝑛 𝑛

Measures of variation in regression

𝑆𝑆𝑇 = 𝑆𝑆𝑅 + 𝑆𝑆𝐸

Total sum of squares (𝑺𝑺𝑻)

(∑ 𝑌𝑖 )2
𝑆𝑆𝑇 = ∑(𝑌 − 𝑌̅)2 = ∑ 𝑌 2 −
𝑛

Regression sum of squares (𝑺𝑺𝑹)

2 (∑ 𝑌)2
𝑆𝑆𝑅 = ∑(𝑌̂ − 𝑌̅) = 𝑏0 ∑ 𝑌 + 𝑏1 ∑ 𝑋𝑌 −
𝑛

Error sum of squares (𝑺𝑺𝑬)

2
𝑆𝑆𝐸 = ∑(𝑌 − 𝑌̂) = ∑ 𝑌 2 − 𝑏0 ∑ 𝑌 − 𝑏1 ∑ 𝑋𝑌

Coefficient of determination

𝑆𝑆𝑅
𝑟2 =
𝑆𝑆𝑇
or
𝑆𝑆𝑋𝑌 2
𝑟2 =
𝑆𝑆𝑋 × 𝑆𝑆𝑌

Standard error of the estimate

𝑆𝑆𝐸
𝑆𝑌𝑋 = √
𝑛−2

Standard error of the slope

𝑆𝑌𝑋
𝑆𝑏1 =
√𝑆𝑆𝑋

Page 4 of 5
 Formula Sheet BB113 Statistics and its Applications

Testing a hypothesis for a population slope, 𝜷𝟏 , using the 𝒕 test

𝑏1 − 𝛽1
𝑡𝑆𝑇𝐴𝑇 = with df = 𝑛 − 2
𝑆𝑏1

Testing a hypothesis for a population slope, 𝜷𝟏 , using the 𝑭 test

𝑀𝑆𝑅
𝐹𝑆𝑇𝐴𝑇 = with df1 = 1, df2 = 𝑛 − 2
𝑀𝑆𝐸
where
𝑆𝑆𝑅 𝑆𝑆𝐸
𝑀𝑆𝑅 = = 𝑆𝑆𝑅 and 𝑀𝑆𝐸 =
1 𝑛−2

Confidence interval estimate of the slope, 𝜷𝟏

𝑏1 ± 𝑡𝛼⁄2 𝑆𝑏1 with df = 𝑛 − 2

J. Chi-Square Tests
𝝌𝟐 test for the difference between two Computing the estimated overall proportion for
proportions 𝒄 groups
(𝑓𝑜 − 𝑓𝑒 )2 𝑋1 + 𝑋2 + ⋯ + 𝑋𝑐 𝑋
2
𝜒𝑆𝑇𝐴𝑇 = ∑ 𝑝̅ = =
𝑓𝑒 𝑛1 + 𝑛2 + ⋯ + 𝑛𝑐 𝑛
all cells

Computing the estimated overall proportion for Computing the expected frequency
two groups Row total × column total
𝑋1 + 𝑋2 𝑋 𝑓𝑒 =
𝑝̅ = = 𝑛
𝑛1 + 𝑛2 𝑛

END OF FORMULA SHEET

Page 5 of 5