Unit 1: Descriptive Statistics

Complete Method Guide — Every Formula with Full Step-by-Step Worked

Examples

1. Mean (Arithmetic Average)

The mean is the sum of all values divided by the number of values. It gives the "balance
point" of a dataset.

1A. Mean for Ungrouped Data

Formula:
M = Sum of all observations (SX) / Total number (N)
Where: X = each observation, N = total number of observations

Problem: Find the mean of: 16, 14, 12, 18, 21, 22, 13, 15, 16, 18
Step 1: Count the observations. N = 10
Step 2: Add up all the values. Sum = 16 + 14 + 12 + 18 + 21 + 22 + 13 + 15 + 16 + 18 =
165
Step 3: Divide by N. M = 165 / 10 = 16.5
Answer: Mean = 16.5

1B. Mean for Grouped Data — Direct Method

Used when data is given as a frequency distribution table (class intervals and
frequencies).
Formula:
Mean = Sum(xi x fi) / Sum(fi)
Where: xi = midpoint of each class = (Upper Limit + Lower Limit) / 2, fi = frequency of
that class

Problem: Find the mean of the following grouped data using the Direct Method.
Class Interval 0-10 10-20 20-30 30-40 40-50
Frequency (fi) 9 13 8 15 10

Step 1: Find midpoints (xi) for each class.

0-10: xi = 5, 10-20: xi = 15, 20-30: xi = 25, 30-40: xi = 35, 40-50: xi = 45
Step 2: Multiply each midpoint by its frequency (xi x fi).
 Class fi xi xi x fi
0-10 9 5 45
10-20 13 15 195
20-30 8 25 200
30-40 15 35 525
40-50 10 45 450
Total 55 1415

Step 3: Apply the formula. Mean = 1415 / 55 = 25.73

Answer: Mean = 25.73

1C. Mean for Grouped Data — Assumed Mean Method

Used to simplify calculations when midpoints are large.
Formula:
Mean = a + Sum(di x fi) / Sum(fi)
Where: a = assumed mean (any convenient midpoint, usually the middle one), di = xi
minus a

Problem: Find the mean using the Assumed Mean Method.

Class Interval 0-10 10-20 20-30 30-40 40-50
Frequency 12 15 10 25 20

Step 1: Find midpoints (xi): 5, 15, 25, 35, 45

Step 2: Choose assumed mean (a). Let a = 25 (midpoint of the middle class).
Step 3: Calculate deviations di = xi minus a:
5-25=-20, 15-25=-10, 25-25=0, 35-25=+10, 45-25=+20
Step 4: Calculate fi x di for each class.
Class fi xi di = xi - 25 fi x di
0-10 12 5 -20 -240
10-20 15 15 -10 -150
20-30 10 25 0 0
30-40 25 35 +10 +250
40-50 20 45 +20 +400
 Total 82 260

Step 5: Apply the formula. Mean = 25 + 260/82 = 25 + 3.17 = 28.17

Answer: Mean = 28.17

1D. Mean for Grouped Data — Step Deviation Method

Used when all class widths are equal. Simplifies the Assumed Mean method by dividing
deviations by the class width.
Formula:
M = AM + (Sum(f x x') / N) x i
Where: AM = Assumed Mean, x' = (xi - AM) / i (coded deviation), i = class width, N =
total frequency

Problem: Find the mean using the Step Deviation Method.

Class 9.5-14.5 14.5-19.5 19.5-24.5 24.5-29.5 29.5-34.5 34.5-39.5
Frequency 3 6 8 11 8 4
(f)

Step 1: Midpoints: 12, 17, 22, 27, 32, 37

Step 2: Choose AM = 27 (midpoint with highest frequency, f=11). Class width i = 5.
Step 3 & 4: Calculate coded deviations x' = (xi - 27) / 5.
Class f xi x' = (xi-27)/5 f x x'
9.5-14.5 3 12 -3 -9
14.5-19.5 6 17 -2 -12
19.5-24.5 8 22 -1 -8
24.5-29.5 11 27 0 0
29.5-34.5 8 32 +1 +8
34.5-39.5 4 37 +2 +8
Total 40 -13

Step 5: M = 27 + (-13/40) x 5 = 27 - 1.625 = 25.375

Answer: Mean = 25.375
2. Median (Middle Value)
The median divides an ordered dataset into two equal halves. 50% of data falls below it
and 50% above. It is a measure of position.

2A. Median for Ungrouped Data

When n is ODD:
Median = value at position (n + 1) / 2
When n is EVEN:
Median = [(n/2)th value + (n/2 + 1)th value] / 2

Problem: Example 1 — ODD number of values: Find the median of 12, 2, 16, 8, 14,
10, 6
Step 1: Arrange in ascending order: 2, 6, 8, 10, 12, 14, 16
Step 2: Count observations. n = 7 (odd).
Step 3: Find the position. Position = (7+1)/2 = 4th value.
Step 4: The 4th value in sorted data is 10.
Answer: Median = 10

Problem: Example 2 — EVEN number of values: Find the median of 7, 9, 3, 4, 11,

1, 8, 6, 1, 4
Step 1: Sorted: 1, 1, 3, 4, 4, 6, 7, 8, 9, 11
Step 2: n = 10 (even). Middle positions are 5th and 6th.
Step 3: 5th value = 4, 6th value = 6.
Step 4: Median = (4 + 6) / 2 = 5
Answer: Median = 5

2B. Median for Grouped Data

Formula:
Median = l + [(N/2 - F) / f] x i
Where: l = lower limit of median class, N = total frequency, F = cumulative frequency of
the class BEFORE the median class, f = frequency of the median class, i = width of the
median class
Finding the Median Class: Calculate N/2. The first class whose cumulative frequency is
greater than or equal to N/2 is the median class.

Problem: Find the median for the following distribution.

Class 9.5-14.5 14.5-19.5 19.5-24.5 24.5-29.5 29.5-34.5 34.5-39.5
 Frequency 3 6 8 11 8 4
(f)

Step 1: Build the cumulative frequency (CF) table.

Class f Cumulative Frequency
(CF)
9.5-14.5 3 3
14.5-19.5 6 9
19.5-24.5 8 17
24.5-29.5 11 28
29.5-34.5 8 36
34.5-39.5 4 40

Step 2: N = 40, so N/2 = 20.

Step 3: First CF >= 20 is 28, which belongs to class 24.5-29.5. This is the median class.
Step 4: l = 24.5, N/2 = 20, F = 17, f = 11, i = 5
Step 5: Median = 24.5 + [(20-17)/11] x 5 = 24.5 + 1.36 = 25.86
Answer: Median = 25.86

3. Mode (Most Frequent Value)

The mode is the value that occurs the most number of times in a dataset.

3A. Mode for Ungrouped Data

Method: Count the frequency of each value. The one with the highest count is the
mode.

Problem: Find the mode of: 34, 46, 45, 39, 43, 22, 27, 37, 46, 35, 34, 39, 40, 30, 30,
41, 37, 46, 39, 29, 34, 39, 35, 43, 30
Step 1: Sort in ascending order and count frequencies:
30 appears 3 times, 34 appears 3 times, 39 appears 4 times (highest), 46 appears 3
times.
Step 2: The value with the highest frequency wins.
Answer: Mode = 39
If 1 value has the highest frequency: Unimodal. If 2 values tie: Bimodal. If 3 or more tie:
Multimodal.
3B. Mode for Grouped Data
Formula:
Mode = l + [(f1 - f0) / (2*f1 - f0 - f2)] x h
Where: l = lower limit of modal class, f1 = frequency of modal class, f0 = frequency of
class just before, f2 = frequency of class just after, h = class width

Problem: Find the mode for the following family size data.
Size of Family 1-3 3-5 5-7 7-9 9-11
No. of Families 7 8 2 2 1

Step 1: Identify the modal class. The highest frequency is 8, which belongs to class 3-5.
Step 2: Extract values: l = 3, f1 = 8, f0 = 7, f2 = 2, h = 2
Step 3: Mode = 3 + [(8-7)/(2x8-7-2)] x 2 = 3 + [1/7] x 2 = 3 + 0.286 = 3.286
Answer: Mode = 3.286

4. Mean Deviation
Mean deviation measures how far, on average, each data point is from the mean. We
use absolute values (ignore negative signs) because positive and negative deviations
would cancel out otherwise.

4A. Mean Deviation for Ungrouped Data

MD = Sum|xi - mu| / n
Where: xi = each observation, mu = mean, | | = absolute value, n = number of
observations

Problem: Find the mean deviation for: 5, 3, 7, 8, 4, 9

Step 1: Mean = (5+3+7+8+4+9) / 6 = 36/6 = 6
Step 2: Find the absolute deviation of each value.
xi xi - mean |xi - mean|
5 -1 1
3 -3 3
7 +1 1
8 +2 2
4 -2 2
 9 +3 3
Sum 12

Step 3: MD = 12 / 6 = 2
Answer: Mean Deviation = 2

4B. Mean Deviation for Grouped Data (Discrete Frequency

Distribution)
MD = Sum(fi * |xi - mu|) / Sum(fi)

Problem: Calculate the mean deviation from the mean.

x 12 9 6 18 10
f 7 3 8 1 2

Step 1: Mean = Sum(fi*xi)/Sum(fi) = (84+27+48+18+20)/(7+3+8+1+2) = 197/21 = 9.381

Step 2: Build the deviation table.
x f xxf |x - 9.381| f x |x - 9.381|
12 7 84 2.619 18.333
9 3 27 0.381 1.143
6 8 48 3.381 27.048
18 1 18 8.619 8.619
10 2 20 0.619 1.238
Total 21 197 56.381

Step 3: MD = 56.381 / 21 = 2.685

Answer: Mean Deviation = 2.685

5. Variance
Variance is the average of squared deviations from the mean. Squaring ensures all
deviations are positive and gives more weight to larger deviations.
5A. Variance for Ungrouped Data
Population Variance:
sigma^2 = Sum(xi - mu)^2 / N
Sample Variance:
s^2 = Sum(xi - x_bar)^2 / (n - 1)
We divide by n-1 for samples to correct for underestimation. This is called Bessel's
correction.

Problem: Find the variance and standard deviation of: 4, 2, 5, 8, 6

Step 1: Mean = (4+2+5+8+6) / 5 = 25/5 = 5
Step 2: Find each deviation and square it.
xi xi - mean (xi - mean)^2
4 -1 1
2 -3 9
5 0 0
8 +3 9
6 +1 1
Sum 20

Population Variance = 20 / 5 = 4 | Sample Variance = 20 / (5-1) = 5

Answer: Population Variance = 4, Sample Variance = 5

5B. Variance for Grouped Data (Discrete) — Shortcut Formula

s^2 = [Sum(f*x^2) - (Sum(f*x))^2 / n] / (n - 1)

Problem: Find sample variance.

X 0 1 2 3 4
Frequency 1 5 10 6 3

x f f*x f*x^2
0 1 0 0
1 5 5 5
2 10 20 40
 3 6 18 54
4 3 12 48
Total n=25 Sum(f*x)=55 Sum(f*x^2)=147

s^2 = [147 - (55)^2/25] / (25-1) = [147 - 121] / 24 = 26/24 = 1.083

Answer: Sample Variance = 1.083

6. Standard Deviation
Standard Deviation = Square Root of Variance. It tells us the spread of data in the
SAME UNITS as the original data (unlike variance which is in squared units).

Population SD: sigma = Square root of [Sum(xi - mu)^2 / N]

Sample SD: s = Square root of [Sum(xi - x_bar)^2 / (n-1)]

Examples (continuing from the Variance examples above):

From ungrouped example (data: 4, 2, 5, 8, 6):
• Population SD = Square root of 4 = 2
• Sample SD = Square root of 5 = 2.236

From discrete grouped example:

• Sample SD = Square root of 1.083 = 1.041

Key insight: SD is ALWAYS just the square root of variance.

7. Quartile Deviation
Quartile Deviation (QD) is half the distance between Q3 and Q1. It represents the
spread of the middle 50% of the data.
QD = (Q3 - Q1) / 2

7A. Quartile Deviation for Ungrouped Data

Q1 = ((n+1)/4)th value -- median of the lower half
Q2 = ((n+1)/2)th value -- median (the middle value)
Q3 = (3(n+1)/4)th value -- median of the upper half

Problem: Find QD of: 17, 2, 7, 27, 15, 5, 14, 8, 10, 24, 48, 10, 8, 7, 18, 28
Step 1: Sort in ascending order: 2, 5, 7, 7, 8, 8, 10, 10, 14, 15, 17, 18, 24, 27, 28, 48. n
= 16.
Step 2: Median Q2 = (8th + 9th) / 2 = (10 + 14) / 2 = 12
Step 3: Lower half (first 8 values): 2, 5, 7, 7, 8, 8, 10, 10
Q1 = (4th + 5th) / 2 = (7 + 8) / 2 = 7.5
Step 4: Upper half (last 8 values): 14, 15, 17, 18, 24, 27, 28, 48
Q3 = (4th + 5th) / 2 = (18 + 24) / 2 = 21
Step 5: QD = (21 - 7.5) / 2 = 13.5 / 2 = 6.75
Answer: Quartile Deviation = 6.75

7B. Quartile Deviation for Grouped Data

Qr = l1 + [r(N/4) - c] / f x (l2 - l1)
Where: r = 1 for Q1, r = 3 for Q3, l1 = lower limit of quartile class, l2 = upper limit, f =
frequency of quartile class, c = cumulative frequency of class before it, N = total
frequency

Problem: Find QD for the following distribution.

Class 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-
100
Freq 5 3 4 3 3 4 7 9 7 8

Class f Cumulative Frequency

(CF)
0-10 5 5
10-20 3 8
20-30 4 12
30-40 3 15
40-50 3 18
50-60 4 22
60-70 7 29
70-80 9 38
80-90 7 45
90-100 8 53

N = 53. Finding Q1: N/4 = 13.25. First CF >= 13.25 is 15, which belongs to class 30-40.
Q1 = 30 + [(13.25 - 12) / 3] x 10 = 30 + 4.167 = 34.167
Finding Q3: 3N/4 = 39.75. First CF >= 39.75 is 45, which belongs to class 80-90.
Q3 = 80 + [(39.75 - 38) / 7] x 10 = 80 + 2.5 = 82.5

QD = (82.5 - 34.167) / 2 = 48.333 / 2 = 24.167

Answer: Quartile Deviation = 24.167

8. Central Moments
Moments are a unified framework for computing statistical characteristics. The r-th
moment tells us about the r-th power of deviations.

8A. Types of Moments

Type 1 — Moments about an Arbitrary Point A (Grouped Data):
mu'_r = Sum(fi * (xi - A)^r) / N for r = 1, 2, 3, 4...
Step Deviation Shortcut (equal class width h):
Let di = (xi - A) / h
mu'_r = Sum(fi * di^r) / N x h^r

Type 2 — Moments about Origin (A = 0):

mu'_r = (1/N) * Sum(fi * xi^r)

Type 3 — Central Moments (about Mean):

mu_r = Sum(fi * (xi - x_bar)^r) / N for r = 0, 1, 2, 3, 4
Key facts: mu_0 = 1 (always), mu_1 = 0 (always), mu_2 = sigma^2 (this is the variance)

8B. Converting Moments from Arbitrary Point to Mean

mu_2 = mu'_2 - (mu'_1)^2
mu_3 = mu'_3 - 3*mu'_2*mu'_1 + 2*(mu'_1)^3
mu_4 = mu'_4 - 4*mu'_3*mu'_1 + 6*mu'_2*(mu'_1)^2 - 3*(mu'_1)^4

8C. Karl Pearson's Coefficients

Beta_1 = mu_3^2 / mu_2^3 (measures skewness)
Beta_2 = mu_4 / mu_2^2 (measures kurtosis)
Gamma_1 = mu_3 / sigma^3 = sqrt(Beta_1)
Gamma_2 = Beta_2 - 3 (excess kurtosis)
Problem: For the following distribution, calculate the first four moments about the
mean and find Beta_1, Beta_2, Gamma_1, Gamma_2.
Marks 5 10 15 20 25 30 35
(x)
Frequen 4 10 20 36 16 12 2
cy (f)

Step 1: Choose A = 20, h = 5. Calculate d = (x - 20) / 5.

x f d=(x- fd fd^2 fd^3 fd^4
20)/5
5 4 -3 -12 36 -108 324
10 10 -2 -20 40 -80 160
15 20 -1 -20 20 -20 20
20 36 0 0 0 0 0
25 16 +1 16 16 16 16
30 12 +2 24 48 96 192
35 2 +3 6 18 54 162
Total Sum(fd^ Sum(fd^ Sum(fd^
N=100 Sum(fd) 2)=178 3)=-42 4)=874
=-6

Step 2: Calculate moments about A using step deviation.

mu'_1 = (-6/100) x 5 = -0.3
mu'_2 = (178/100) x 25 = 44.5
mu'_3 = (-42/100) x 125 = -52.5
mu'_4 = (874/100) x 625 = 5462.5

Step 3: Convert to central moments.

mu_2 = 44.5 - (-0.3)^2 = 44.5 - 0.09 = 44.41 (this is sigma^2)
mu_3 = -52.5 - 3(44.5)(-0.3) + 2(-0.3)^3 = -52.5 + 40.05 - 0.054 = -12.504
mu_4 = 5462.5 - 4(-52.5)(-0.3) + 6(44.5)(0.09) - 3(0.0081) = 5423.506

Step 4: sigma = sqrt(44.41) = 6.664

Step 5: Calculate Beta and Gamma coefficients.

Beta_1 = (-12.504)^2 / (44.41)^3 = 156.35 / 87616 = 0.00178
Gamma_1 = -12.504 / (6.664)^3 = -0.0422 (slight negative skew)
Beta_2 = 5423.506 / (44.41)^2 = 2.750
Gamma_2 = 2.750 - 3 = -0.250 (Platykurtic: flatter than normal)
Answer: mu_2=44.41, mu_3=-12.504, mu_4=5423.506 | Beta_1=0.00178,
Beta_2=2.750, Gamma_2=-0.250 (Platykurtic)

9. Skewness
Skewness tells you whether data is spread out more on one side of the mean.

Type Tail Relationship of Mean, Median,

Mode
Positive Skew (right tail Extends right Mean > Median > Mode
longer)
Negative Skew (left tail Extends left Mean < Median < Mode
longer)
Symmetric (no skew) Equal both sides Mean = Median = Mode

Pearson's Mode Skewness:

Skewness = (X_bar - Mode) / s
Pearson's Median Skewness (when mode is unclear):
Skewness = 3 x (X_bar - Median) / s
Using Moments:
Gamma_1 = mu_3 / sigma^3
Gamma_1 > 0 means positive skew. Gamma_1 < 0 means negative skew. Gamma_1 = 0
means symmetric.

Problem: Using Pearson's Mode formula: Mean = 40, Mode = 35, SD = 10

Skewness = (40 - 35) / 10 = 5/10 = 0.5 (Positive Skew — longer tail to the right)
Answer: Skewness = +0.5 (Positive skew)

10. Kurtosis
Kurtosis measures how peaked or flat a distribution is compared to the normal (bell)
curve.
Beta_2 = mu_4 / mu_2^2
Gamma_2 = Beta_2 - 3
 Condition Type Shape Description
Beta_2 = 3 (Gamma_2 = Mesokurtic Normal bell curve
0)
Beta_2 > 3 (Gamma_2 > Leptokurtic Tall, sharp peak; heavy tails
0) (more outliers)
Beta_2 < 3 (Gamma_2 < Platykurtic Flat, broad peak; light tails (fewer
0) outliers)

Problem: Example 1: mu_1=0, mu_2=2.5, mu_3=0.7, mu_4=18.75

Beta_1 = (0.7)^2 / (2.5)^3 = 0.49/15.625 = 0.031 (nearly symmetric)
Beta_2 = 18.75 / (2.5)^2 = 3 --> Gamma_2 = 0 --> Mesokurtic
Answer: Beta_2 = 3 means the distribution is Mesokurtic (same shape as the
normal curve).

Quick Reference: Which Method to Use When?

Situation Best Method to Use
Raw data, few values Ungrouped formulas (direct calculation)
Frequency table with no class intervals Discrete grouped formulas
Class intervals with small midpoints Direct Method for mean
Class intervals with large midpoints Assumed Mean Method
Class intervals with equal widths Step Deviation Method (fastest)
Need spread in the same units as data Standard Deviation (= square root of
Variance)
Need spread of the middle 50% Quartile Deviation
Need to check symmetry of distribution Skewness (Gamma_1)
Need to check peak shape of distribution Kurtosis (Beta_2 or Gamma_2)
Full distribution characterisation Central Moments then Beta_1, Beta_2,
Gamma_1, Gamma_2

End of Complete Method Guide — Unit 1, Descriptive Statistics