STATISTICS FORMULAS &

PROBLEM-SOLVING GUIDE
Complete Reference with Pattern Recognition
FPT University Statistics Course

Contents
1 PROBABILITY BASICS 3
1.1 Independent vs Disjoint Events . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Conditional Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Bayes’ Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 DISCRETE PROBABILITY DISTRIBUTIONS 6

2.1 Binomial Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Negative Binomial Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Hypergeometric Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Poisson Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3 CONTINUOUS DISTRIBUTIONS 11
3.1 Uniform Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 Exponential Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.3 Normal Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4 Finding Percentiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4 SAMPLING DISTRIBUTIONS & CLT 15

4.1 Central Limit Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

5 CONFIDENCE INTERVALS 17
5.1 CI for Population Mean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.2 CI for Proportion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

6 HYPOTHESIS TESTING 20
6.1 Understanding Hypothesis Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.2 Test for One Mean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6.3 Test for One Proportion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6.4 Two-Sample Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

7 SAMPLE SIZE DETERMINATION 24

8 CORRELATION & REGRESSION 25

8.1 Correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
8.2 Linear Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

1
Statistics Formulas & Problem-Solving Guide 2

9 DESCRIPTIVE STATISTICS 27
9.1 Quartiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
9.2 Stem-and-Leaf Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

10 QUICK DECISION FLOWCHARTS 29

10.1 Choosing the Right Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
10.2 Hypothesis Test vs Confidence Interval . . . . . . . . . . . . . . . . . . . . . . . 29
10.3 One-Sample vs Two-Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

11 COMMON MISTAKES TO AVOID 30

11.1 Distribution Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
11.2 Probability Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
11.3 Hypothesis Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
11.4 Confidence Intervals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

12 EXAM STRATEGY 32
12.1 Step-by-Step Problem Solving . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
12.2 Time Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
12.3 Common Sense Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Statistics Formulas & Problem-Solving Guide 3

1 PROBABILITY BASICS
1.1 Independent vs Disjoint Events
How to Identify This Problem Type
Key phrases in problem:

ˆ ”Are A and B independent?”

ˆ ”Are A and B disjoint/mutually exclusive?”

ˆ Given P (A), P (B), and P (A ∩ B) or P (A and B)

What to look for:

ˆ Three probabilities are given

ˆ Question asks about relationship between events

Definitions & Formulas

Disjoint (Mutually Exclusive): P (A ∩ B) = 0
Independent: P (A ∩ B) = P (A) × P (B)
Decision Process:

1. Check if P (A ∩ B) = 0 → Disjoint

2. Check if P (A ∩ B) = P (A) × P (B) → Independent

Worked Example

Given: P (A) = 0.4, P (B) = 0.3, P (A ∩ B) = 0

Step 1 - Check Disjoint: P (A ∩ B) = 0 ✓
Step 2 - Check Independent: P (A) × P (B) = 0.4 × 0.3 = 0.12 ̸= 0 ×
Answer: Disjoint but NOT independent
Statistics Formulas & Problem-Solving Guide 4

1.2 Conditional Probability

How to Identify This Problem Type
Key phrases in problem:

ˆ ”Given that...”, ”if...”, ”when...”

ˆ ”probability that A occurs given B”

ˆ Symbol: P (A|B) present in problem

ˆ Multi-stage problems with different groups

Common scenarios:

ˆ Two-way tables (categories Ö outcomes)

ˆ Sequential events with conditions

ˆ Medical testing (disease given test result)

Formula
P (A ∩ B)
P (A|B) =
P (B)
Rearranged: P (A ∩ B) = P (A|B) × P (B)

Worked Example

Given: P (2 cars|income > 20k) = 0.9, P (income > 20k) = 0.6

Find: P (2 cars AND income > 20k)
Solution: P (both) = 0.9 × 0.6 = 0.54

1.3 Bayes’ Theorem

How to Identify This Problem Type

Key phrases (REVERSE probability):

ˆ ”If a randomly selected [outcome] has [property], what is the probability it came
from [group]?”

ˆ ”Given that [result occurred], find probability of [original cause]”

ˆ Working backwards from effect to cause

Structure:

ˆ Multiple groups with known sizes and success rates

ˆ Question asks: ”Which group did it come from?”

ˆ Example: ”If student passed, probability they’re from Group X?”

Statistics Formulas & Problem-Solving Guide 5

Formula
P (B|A) × P (A)
P (A|B) =
P (B)
P
Where: P (B) = P (B|Ai ) × P (Ai )

Worked Example: Texas College

Given:

ˆ Southern: 60%, pass rate 60%

ˆ Northern: 30%, pass rate 70%

ˆ Out-of-state: 10%, pass rate 90%

Find: P (Out-of-state|passed)
Step 1: Calculate total pass probability

P (passed) = 0.6(0.6) + 0.3(0.7) + 0.1(0.9) = 0.66

Step 2: Apply Bayes’ Theorem

0.9 × 0.1 0.09
P (Out|passed) = = = 0.136
0.66 0.66
Statistics Formulas & Problem-Solving Guide 6

2 DISCRETE PROBABILITY DISTRIBUTIONS

2.1 Binomial Distribution
How to Identify Binomial Problems
Key phrases in problem:

ˆ ”exactly k successes in n trials”

ˆ ”probability that k out of n...”

ˆ ”how many students work full-time” (out of total)

ˆ ”standard deviation of number who...”

ALL four conditions MUST be met:

1. Fixed number of trials (n)

2. Two outcomes only (success/failure)

3. Constant probability (p) for each trial

4. Trials are independent

Common scenarios:

ˆ Flipping coins n times

ˆ Sampling WITH replacement (or large population)

ˆ Survey responses (yes/no) from many people

ˆ Quality control with large batches

NOT Binomial if:

ˆ Sample WITHOUT replacement from small group → Hypergeometric

ˆ Counting until r-th success → Negative Binomial

ˆ More than 2 outcomes → Multinomial

Formulas
 
n k
P (X = k) = p (1 − p)n−k
k
E(X) = np
Var(X) = np(1 − p)
p
σ = np(1 − p)
Statistics Formulas & Problem-Solving Guide 7

Worked Example

Problem: ”18% of students work full-time. In sample of 35 students, find standard

deviation of number working full-time.”
Identify: n = 35, p = 0.18 → Binomial!
Solution: √ √
σ = 35 × 0.18 × 0.82 = 5.166 = 2.27

2.2 Negative Binomial Distribution

How to Identify Negative Binomial
Key phrases in problem:

ˆ ”probability that the r-th success occurs on trial k”

ˆ ”first two successes require exactly 4 trials”

ˆ ”how many trials until we get r successes”

ˆ Focus is on trial number, not success count

Structure:

ˆ Number of trials is variable (not fixed)

ˆ Number of successes is fixed (r)

ˆ Question: ”On which trial does r-th success occur?”

Example scenarios:

ˆ ”First 2 successful alignments in exactly 4 trials”

ˆ ”3rd defective item found on 10th inspection”

Formula
 
k−1 r
P (r-th success on trial k) = p (1 − p)k−r
r−1
Logic: Need r − 1 successes in first k − 1 trials, then success on k-th trial

Worked Example

Problem: ”Probability of success = 0.7. Find P(2nd success on exactly 4th trial)”
Identify: r = 2, k = 4, p = 0.7 → Negative Binomial!
Solution:
 
3
P = (0.7)1 (0.3)2 × 0.7 = 3 × 0.7 × 0.09 × 0.7 = 0.1323
1
Statistics Formulas & Problem-Solving Guide 8

2.3 Hypergeometric Distribution

How to Identify Hypergeometric
Key phrases in problem:

ˆ ”sample without replacement”

ˆ ”randomly select n from population of N ”

ˆ Population is small (usually < 100)

ˆ ”exactly k are [type]” from finite group

ALL conditions must be met:

1. Population size N is small/finite

2. Two types of items (success/failure)

3. Sampling WITHOUT replacement

4. Want probability of exactly k successes in n draws

Common scenarios:

ˆ Card games (52 cards, no replacement)

ˆ Quality control (small batches)

ˆ Committee selection

ˆ Lottery problems

When to use Binomial instead:

ˆ If N ≥ 20n (population is large) → use Binomial

ˆ If sampling WITH replacement → use Binomial

Formula
K N −K



k n−k
P (X = k) = N


n
Where:

ˆ N = total population

ˆ K = total successes in population

ˆ n = sample size

ˆ k = successes in sample
Statistics Formulas & Problem-Solving Guide 9

Worked Example
Problem: ”36 bacteria cells: 12 capable, 24 not. Select 7 without replacement. Find
P(exactly 3 capable)”
Identify: N = 36, K = 12, n = 7, k = 3, no replacement → Hypergeometric!
Solution:
12
× 24



3 4 220 × 10626
P (X = 3) = 36

= ≈ 0.28
7
8347680

2.4 Poisson Distribution

How to Identify Poisson Problems
Key phrases in problem:

ˆ ”average number of [events] per [time/space]”

ˆ ”mean of λ per day/hour/meter”

ˆ ”rare events” in fixed interval

ˆ ”number of arrivals/failures/calls per...”

Characteristics:

1. Events occur randomly over time/space

2. Events are rare (low probability)

3. Events are independent

4. Average rate λ is known

Common scenarios:

ˆ Number of power failures per day

ˆ Customer arrivals per hour

ˆ Typos per page

ˆ Accidents per month

ˆ Phone calls per minute

Typical questions:

ˆ ”Find P(exactly k events)”

ˆ ”P(no events)” = e−λ

ˆ ”P(at least 1)” = 1 − e−λ

Statistics Formulas & Problem-Solving Guide 10

Formula
e−λ λk
P (X = k) =
k!
Where λ = average rate per interval

Worked Example

Problem: ”Power company has mean 0.12 failures per day. Find P(exactly 2 failures in
a day)”
Identify: λ = 0.12, want P (X = 2) → Poisson!
Solution:
e−0.12 × (0.12)2 0.887 × 0.0144
P (X = 2) = = ≈ 0.006
2! 2
Statistics Formulas & Problem-Solving Guide 11

3 CONTINUOUS DISTRIBUTIONS
3.1 Uniform Distribution
How to Identify Uniform Problems
Key phrases in problem:

ˆ ”uniformly distributed over interval [a, b]”

ˆ ”equally likely” over a range

ˆ ”constant probability” between two values

ˆ ”random variable distributed over...”

Characteristics:

ˆ All values in range [a, b] equally likely

ˆ Rectangular probability density function

ˆ Often used for: wait times, measurement errors

Common scenarios:

ˆ Machine output between min and max

ˆ Arrival time within time window

ˆ Random number generation

Formula
For X ∼ Uniform(a, b):
1
f (x) = , a≤x≤b
b−a
d−c
P (c < X < d) =
b−a
b−c
P (X > c) =
b−a
d−a
P (X < d) =
b−a

Worked Example

Problem: ”Machine pumps uniformly between 9.75 and 11.25 gallons. Find P(pump ¿
10.65)”
Identify: Uniform on [9.75, 11.25] → Uniform!
Solution:
11.25 − 10.65 0.6
P (X > 10.65) = = = 0.4
11.25 − 9.75 1.5
Statistics Formulas & Problem-Solving Guide 12

3.2 Exponential Distribution

How to Identify Exponential Problems
Key phrases in problem:

ˆ ”wait time until event occurs”

ˆ ”time between events”

ˆ ”best approximated by exponential distribution”

ˆ Given: mean time µ

Characteristics:

ˆ Models waiting times

ˆ Memoryless property

ˆ Related to Poisson (time version)

Common scenarios:

ˆ Time until next customer arrives

ˆ Time on hold before answer

ˆ Lifespan of electronic components

ˆ Time between phone calls

Relationship to Poisson:

ˆ Poisson: counts events in fixed time

ˆ Exponential: time until next event

Formula
For mean µ:
P (X > x) = e−x/µ
P (X ≤ x) = 1 − e−x/µ

Worked Example

Problem: ”Wait time is exponential with mean 4.2 minutes. Find P(wait ¿ 1.8 minutes)”
Identify: µ = 4.2, want P (X > 1.8) → Exponential!
Solution:
P (X > 1.8) = e−1.8/4.2 = e−0.4286 ≈ 0.651
Statistics Formulas & Problem-Solving Guide 13

3.3 Normal Distribution

How to Identify Normal Problems
Key phrases in problem:

ˆ ”normally distributed with mean µ and standard deviation σ”

ˆ ”bell-shaped distribution”

ˆ ”assume normal distribution”

Question types:

1. Find probability: ”Find P(a < X < b)”

2. Find percentile: ”Find value where X% are above/below”

3. Sample mean problems: Use CLT if n ≥ 30

Common scenarios:

ˆ Heights, weights, test scores

ˆ Measurement errors

ˆ Manufacturing specifications

ˆ Natural phenomena

Formula
Convert to Z-score:
X −µ
Z=
σ
X̄−µ
For sample mean: Z = σ/ √
n
Then use Z-table or calculator

Worked Example

Problem: ”Beer bottles: µ = 16 oz, σ = 1.2 oz. Find P(12.5 < X < 14.5)”
Identify: Normal distribution given → Normal!
Solution:
12.5 − 16 14.5 − 16
Z1 = = −2.917, Z2 = = −1.25
1.2 1.2
P = Φ(−1.25) − Φ(−2.917) = 0.1056 − 0.0018 = 0.1039
Statistics Formulas & Problem-Solving Guide 14

3.4 Finding Percentiles

How to Identify Percentile Problems
Key phrases in problem:

ˆ ”Find value where X% are above”

ˆ ”Find value where X% are below”

ˆ ”90% of values are greater than...”

ˆ ”Top 10% of scores”

Direction:

ˆ Given: probability or percentage

ˆ Find: actual X value

ˆ This is reverse of normal probability

Formula

x=µ+Z ×σ
Where Z = invNorm(cumulative probability)
Common percentiles:

ˆ 90th percentile: Z = 1.28

ˆ 95th percentile: Z = 1.645

ˆ 99th percentile: Z = 2.33

Worked Example

Problem: ”µ = 12.4, σ = 4.3. Find value where 86% are above”

Identify: Finding X given probability → Percentile!
Solution:
P (X > x) = 0.86 ⇒ P (X < x) = 0.14
Z = invNorm(0.14) = −1.08
x = 12.4 + (−1.08)(4.3) = 7.8
Statistics Formulas & Problem-Solving Guide 15

4 SAMPLING DISTRIBUTIONS & CLT

4.1 Central Limit Theorem
How to Identify CLT Problems
Key phrases in problem:

ˆ ”average of n observations”

ˆ ”sample mean X̄” or ”mean of sample”

ˆ ”probability that average exceeds...”

ˆ ”n items are randomly selected”

Critical distinction:

ˆ Question about single value → Use population distribution

ˆ Question about average/mean → Use CLT

When CLT applies:

1. Sample size n ≥ 30, OR

2. Population is normal (any n)

Common scenarios:

ˆ Average score of 100 students

ˆ Mean time for 25 cars

ˆ Sample mean exceeds value

Formula
For sample mean X̄ from population with µ and σ:
 
σ
X̄ ∼ N µ, √
n
Standard Error: SE = √σ
n

X̄ − µ
Z= √
σ/ n
σ2
Variance of sample mean: Var(X̄) = n
Statistics Formulas & Problem-Solving Guide 16

Worked Example

Problem: ”Golfer scores: µ = 80, σ = 3. For 100 golfers, find P(X̄ > 80.5)”
Identify: Asking about average of 100 → CLT!
Solution:
3
SE = √ = 0.3
100
80.5 − 80
Z= = 1.667
0.3
P (Z > 1.667) ≈ 0.0478
Statistics Formulas & Problem-Solving Guide 17

5 CONFIDENCE INTERVALS
5.1 CI for Population Mean
How to Identify CI Problems
Key phrases in problem:

ˆ ”construct a [X]% confidence interval”

ˆ ”estimate the population mean”

ˆ ”interval estimate for µ”

ˆ Given sample data: n, x̄, s (or σ)

Decision tree:

1. Is σ known?

ˆ Yes → Use Z-interval

ˆ No → Use t-interval

2. Sample size?

ˆ n ≥ 30 → Can use Z even if σ unknown

ˆ n < 30 → Must use t

Formulas
Z-interval (known σ):
σ
x̄ ± zα/2 × √
n
t-interval (unknown σ):
s
x̄ ± tα/2,df × √
n
where df = n − 1
Common z-values:

ˆ 90% CI: z = 1.645

ˆ 95% CI: z = 1.96

ˆ 99% CI: z = 2.576

Statistics Formulas & Problem-Solving Guide 18

Worked Example

Problem: ”n = 28, x̄ = 82.5, s = 9.2. Find 95% CI for µ”

Identify: σ unknown, n < 30 → t-interval!
Solution:
9.2
SE = √ = 1.738
28
t0.025,27 ≈ 2.052
CI = 82.5 ± 2.052(1.738) = (78.93, 86.07)

5.2 CI for Proportion

How to Identify Proportion CI
Key phrases in problem:

ˆ ”proportion of population”

ˆ ”percentage who...”

ˆ Given: sample size n and number of successes x

ˆ ”construct CI for difference between two proportions”

Data format:

ˆ Sample proportion: p̂ = x
n

ˆ NOT means or averages

ˆ Usually percentages or rates

Formulas
One proportion: r
p̂(1 − p̂)
p̂ ± zα/2 ×
n
Two proportions:
s
p̂1 (1 − p̂1 ) p̂2 (1 − p̂2 )
(p̂1 − p̂2 ) ± zα/2 × +
n1 n2
Statistics Formulas & Problem-Solving Guide 19

Worked Example

Problem: ”Democrats: 92/205 voted yes. Republicans: 120/230 voted yes. Find 95%
CI for difference”
Identify: Two proportions → Two-proportion CI!
Solution:
p̂1 = 0.45, p̂2 = 0.52
r
0.45(0.55) 0.52(0.48)
SE = + = 0.0479
205 230
CI = (0.45 − 0.52) ± 1.96(0.0479) = (−0.16, 0.02)
Statistics Formulas & Problem-Solving Guide 20

6 HYPOTHESIS TESTING
6.1 Understanding Hypothesis Tests
How to Identify Hypothesis Test
Key phrases in problem:

ˆ ”test the claim that...”

ˆ ”is there evidence that...”

ˆ ”at significance level α”

ˆ ”test H0 : µ = µ0 ”

Structure:

1. Claim is made (null hypothesis)

2. Sample data is collected

3. Question: Is data consistent with claim?

Types by parameter:

ˆ One mean: H0 : µ = µ0

ˆ One proportion: H0 : p = p0

ˆ Two means: H0 : µ1 = µ2

ˆ Two proportions: H0 : p1 = p2

Type I vs Type II Errors

Type I Error (α): Reject H0 when H0 is true

ˆ False positive

ˆ ”Crying wolf when no wolf”

Type II Error (β): Fail to reject H0 when H0 is false

ˆ False negative

ˆ ”Missing the wolf that’s there”

Statistics Formulas & Problem-Solving Guide 21

6.2 Test for One Mean

How to Identify
Key phrases:

ˆ ”test if mean equals [value]”

ˆ ”claim: µ = µ0 ”

ˆ Given: x̄, s (or σ), n

Choose test statistic:

ˆ Known σ → Z-test

ˆ Unknown σ → t-test

Formulas
Z-test (known σ):
x̄ − µ0
Z= √
σ/ n
t-test (unknown σ):
x̄ − µ0
t= √ , df = n − 1
s/ n
Critical values (α = 0.05):

ˆ Two-tailed: ±1.96 (Z), varies (t)

ˆ One-tailed: ±1.645 (Z)

6.3 Test for One Proportion

How to Identify
Key phrases:

ˆ ”test if proportion equals [value]”

ˆ ”claim: at most X% are...”

ˆ Given: n, x (number of successes)

ˆ Claim about percentage or rate

Common scenarios:

ˆ ”15% of students are overweight”

ˆ ”More than 50% support policy”

ˆ ”Defect rate is 2%”

Statistics Formulas & Problem-Solving Guide 22

Formula
p̂ − p0
Z=p
p0 (1 − p0 )/n
Where: p̂ = nx , p0 = claimed proportion

Worked Example

Problem: ”Claim: at most 15% overweight. Sample: 28/175 overweight. Test at

α = 0.03”
Identify: Proportion test, one-tailed → Z-test for proportion!
Solution:
28
p̂ = = 0.16
175
0.16 − 0.15
Z=p = 0.37
0.15(0.85)/175
Critical value: Z0.03 = −1.88
Since 0.37 > −1.88, fail to reject H0

6.4 Two-Sample Tests

How to Identify Two-Sample Tests
Key phrases:

ˆ ”compare two groups”

ˆ ”test if µ1 = µ2 ” or ”µ1 > µ2 ”

ˆ ”difference between two means/proportions”

ˆ Given data from TWO independent samples

Types:

1. Two means, independent samples → Two-sample t-test

2. Two means, paired data → Paired t-test

3. Two proportions → Two-proportion Z-test

Statistics Formulas & Problem-Solving Guide 23

Two-Sample t-test Formula

x̄ − x̄2
t= q1
sp n11 + n12
Pooled standard deviation:
s
(n1 − 1)s21 + (n2 − 1)s22
sp =
n1 + n2 − 2

df = n1 + n2 − 2
Statistics Formulas & Problem-Solving Guide 24

7 SAMPLE SIZE DETERMINATION

How to Identify Sample Size Problems
Key phrases in problem:

ˆ ”how large a sample is needed”

ˆ ”minimum sample size required”

ˆ ”determine sample size for [X]% confidence”

ˆ Given: desired margin of error E

Information needed:

1. Confidence level (determines z or t)

2. Margin of error E

3. Population standard deviation σ (for mean)

4. Estimated proportion p (for proportion)

Formulas
Sample size for mean: 2
zα/2 × σ

n=
E
Sample size for proportion:
2
zα/2 × p(1 − p)
n=
E2
If p unknown: Use p = 0.5 (most conservative)
Always round UP to next integer

Worked Example

Problem: ”σ = 3.2, want 95% CI with margin of error = 2. Find n”

Identify: Finding sample size for mean → Sample size formula!
Solution:  2
1.96 × 3.2
n= = (3.136)2 = 9.83 ≈ 10
2
Statistics Formulas & Problem-Solving Guide 25

8 CORRELATION & REGRESSION

8.1 Correlation
How to Identify Correlation Problems
Key phrases in problem:

ˆ ”correlation coefficient r”

ˆ ”relationship between X and Y”

ˆ ”linear correlation”

ˆ Given: paired data (x, y)

ˆ Scatter plot shown

Types of questions:

1. Calculate r from data

2. Interpret scatter plot

3. Determine direction of correlation

Common scenarios:

ˆ Supply vs price (negative)

ˆ Study time vs grade (positive)

ˆ Height vs weight (positive)

Properties of Correlation

−1 ≤ r ≤ 1
Interpretation:

ˆ r > 0: Positive correlation

ˆ r < 0: Negative correlation

ˆ r = 0: No linear correlation

ˆ |r| close to 1: Strong relationship

ˆ |r| close to 0: Weak relationship

Important: Correlation ̸= Causation

Statistics Formulas & Problem-Solving Guide 26

8.2 Linear Regression

How to Identify Regression Problems
Key phrases in problem:

ˆ ”find regression line”

ˆ ”equation of estimated regression line”

ˆ ”predict Y from X”

ˆ Given: data table with X and Y values

Question types:

1. Find equation: ŷ = a + bx

2. Predict Y for given X

3. Test hypothesis about slope

Regression Line Formula

ŷ = a + bx
Slope: P
(xi − x̄)(yi − ȳ)
b= P
(xi − x̄)2
Intercept:
a = ȳ − bx̄

Worked Example

Given: x = [1, 2, 2, 5], y = [30, 40, 90, 120]

Calculate: x̄ = 2.5, ȳ = 70
After calculation: b = 21.11, a = 17.22
Regression line: ŷ = 17.22 + 21.11x
Statistics Formulas & Problem-Solving Guide 27

9 DESCRIPTIVE STATISTICS
9.1 Quartiles
How to Identify Quartile Problems
Key phrases in problem:

ˆ ”first quartile” or ”Q1”

ˆ ”third quartile” or ”Q3”

ˆ ”median” or ”Q2”

ˆ ”identify the quartile”

ˆ Given: list of data values

What quartiles represent:

ˆ Q1: 25th percentile

ˆ Q2: 50th percentile (median)

ˆ Q3: 75th percentile

Quartile Formulas
Position formulas:
n+1
Q1 position =
4
n+1
Q2 position =
2
3(n + 1)
Q3 position =
4
If position is not integer: Interpolate between adjacent values

Worked Example

Problem: ”Data (n = 25): 60, 61, ..., 92. Find Q1”

Identify: Finding quartile → Use position formula!
Solution:
26
Q1 position = = 6.5
4
65 + 66
Q1 = = 65.5
2
Statistics Formulas & Problem-Solving Guide 28

9.2 Stem-and-Leaf Diagram

How to Identify
Key phrases:

ˆ ”construct stem-and-leaf diagram”

ˆ Given: list of decimal numbers

ˆ Need to organize data visually

Purpose:

ˆ Shows distribution shape

ˆ Preserves original data

ˆ Easy to see patterns

How to Create
Steps:

1. Separate each number into stem (leading digits) and leaf (last digit)

2. List stems vertically in order

3. Write leaves horizontally in ascending order

Example: 15.2 → Stem: 15, Leaf: 2

Statistics Formulas & Problem-Solving Guide 29

10 QUICK DECISION FLOWCHARTS

10.1 Choosing the Right Distribution
If Problem Says... Use Key Feature
”exactly k successes in Binomial Fixed n, want P (X = k)
n trials”
”without replace- Hypergeometric No replacement
ment”, small N
”r-th success on k-th Neg. Binomial Variable trials
trial”
”mean λ per interval” Poisson Rare events
”uniformly between a Uniform Constant density
and b”
”waiting time”, mean Exponential Time between events
given
”normally dis- Normal Bell curve
tributed”
”average of n values” CLT/Normal Sample mean

10.2 Hypothesis Test vs Confidence Interval

Confidence Interval Hypothesis Test
”Construct CI” ”Test the claim”
”Estimate µ” ”Is µ = µ0 ?”
Gives range of plausible values Makes yes/no decision
Reports interval Reports test statistic, p-value

10.3 One-Sample vs Two-Sample

One-Sample Two-Sample
Compare sample to known value Compare two samples
”Is µ = 50?” ”Is µ1 = µ2 ?”
One dataset given Two datasets given
Z-test or t-test Two-sample t-test
Statistics Formulas & Problem-Solving Guide 30

11 COMMON MISTAKES TO AVOID

11.1 Distribution Selection
1. Confusing Binomial and Hypergeometric

ˆ Binomial: WITH replacement OR large population

ˆ Hypergeometric: WITHOUT replacement AND small population

2. Confusing Binomial and Negative Binomial

ˆ Binomial: Fixed trials, count successes

ˆ Negative Binomial: Fixed successes, count trials

3. Missing CLT requirement

√
ˆ For sample mean problems, divide σ by n
ˆ Don’t use population σ directly

11.2 Probability Concepts

1. Disjoint vs Independent

ˆ Disjoint events CANNOT be independent (unless P = 0)

ˆ If P (A ∩ B) = 0, they’re disjoint

2. Conditional Probability Direction

ˆ P (A|B) ̸= P (B|A) in general

ˆ Read carefully: ”given” comes after the bar

11.3 Hypothesis Testing

1. Type I vs Type II Errors

ˆ Type I: False positive (reject true null)

ˆ Type II: False negative (fail to reject false null)

2. One-tailed vs Two-tailed

ˆ Ha : µ ̸= µ0 → Two-tailed
ˆ Ha : µ > µ0 or µ < µ0 → One-tailed

3. Using Wrong Test Statistic

ˆ Known σ → Z-test
ˆ Unknown σ → t-test
Statistics Formulas & Problem-Solving Guide 31

11.4 Confidence Intervals

1. Sample Size Effect on Width

ˆ Larger n → Narrower CI
ˆ Higher confidence → Wider CI

2. Interpretation

ˆ Correct: ”95% confident true µ is in interval”

ˆ Wrong: ”µ has 95% probability of being in interval”
Statistics Formulas & Problem-Solving Guide 32

12 EXAM STRATEGY
12.1 Step-by-Step Problem Solving
General Approach:

1. Read carefully - Identify all given information

2. Recognize pattern - Use pattern boxes to identify type

3. Write knowns - List n, µ, σ, p, etc.

4. Choose formula - Select from formula boxes

5. Calculate - Show all steps

6. Check - Does answer make sense?

12.2 Time Management

ˆ Quick wins first: Do problems you recognize immediately

ˆ Skip and return: Don’t get stuck on one problem

ˆ Check units: Make sure answer has correct units

ˆ Use calculator: For normal/t distributions

12.3 Common Sense Checks

ˆ Probabilities must be between 0 and 1

ˆ Standard deviations are always positive

ˆ Sample means cluster around population mean

ˆ Larger samples give more precise estimates

ˆ Confidence intervals should be reasonable ranges

GOOD LUCK WITH YOUR EXAM!

Remember: Pattern recognition is key.
Practice identifying problem types before solving.

Study Tips:

ˆ Focus on pattern boxes to quickly identify problem types

ˆ Memorize critical z and t values

ˆ Practice CLT problems - they’re common

ˆ Understand when to use each distribution