Frequency Distribution
Business Statistics
1. What is a Frequency Distribution?
A frequency distribution is a summary table that organises raw data into classes (groups) and
shows how often each value or group of values occurs. It transforms a disorganised list of numbers into
a meaningful, readable summary — making patterns, concentrations, and outliers immediately visible.
In healthcare and business management, frequency distributions are foundational tools for analysing
any measured variable — patient waiting times, sales figures, employee productivity, complaint
volumes, and more.
Key Terms
• Class / Class Interval – A range of values grouped together (e.g., 0–4 complaints per day)
• Frequency (f) – The count of observations that fall within a class
• Relative Frequency – Frequency divided by total observations; expressed as a proportion or
percentage
• Cumulative Frequency – A running total of frequencies from the first class to the current class
• Class Width (W) – The span covered by each class: (Largest value − Smallest value) ÷ Number
of classes
2. Building a Frequency Distribution: The Tally Technique
The tally technique is the simplest and most intuitive way to construct a frequency distribution. For
each data point, you mark a vertical stroke (|) in the corresponding category. Every fifth observation is
drawn diagonally across the previous four (||||͟), creating bundles of five that are easy to count.
Example: Blood Group Distribution in a Hospital OPD
A hospital's Out-Patient Department recorded the blood groups of 30 patients visiting a general
medicine clinic on a single day. The raw data is:
A O B O AB O O B A O B O O A B O AB B O B A O B AB O
B O A O O
Step 1 — Identify the Categories
The four blood groups — A, B, AB, and O — form our classes. These are nominal (qualitative)
categories, not numbers.
Step 2 — Mark Tallies
Read through each entry. For each patient, place one tally mark in the matching group. After four marks
(||||), cross the fifth diagonally (||||͟) to create a bundle of 5.
Step 3 — Count and Record
Count each group's tally bundle(s), record the frequency, and calculate relative frequency = f ÷ 30.
Blood Group Tally Marks Frequency Relative Freq. (%)
A ||||| 5 16.7%
B ||||||| 7 23.3%
AB ||| 3 10.0%
O ||||͟ ||||͟ ||||͟ 15 50.0%
Total 30 100%
Observation: Blood Group O is the most common at 50%. This directly informs blood bank inventory
decisions — O-type blood should always be maintained at highest stock levels.
3. Grouped Frequency Distribution: Using the 2k ≥ n Rule
When data has many distinct numerical values, listing each value separately results in a table too long
to be useful. We instead group values into classes. The challenge is deciding: how many classes?
How wide should each class be? The 2k ≥ n rule provides a systematic answer.
Problem 1 — Cell Phone Calls per Day
The following table shows the number of cell phone calls made per day by 30 individuals. The values
range from 0 to 14. Construct a grouped frequency distribution.
Cell Phone Calls per Day (Raw Data, n = 30)
4 5 1 0 7 8
3 6 8 3 0 9
2 12 14 5 5 10
7 2 11 9 4 3
1 5 7 3 5 6
Step 1 — Determine Number of Classes (2k ≥ n Rule)
With n = 30, we find the smallest k such that 2k ≥ 30:
k (classes) 2^k Sufficient for n ≤
4 16 16
5 32 32
6 64 64
7 128 128
2⁴ = 16 (less than 30 — insufficient). 2⁵ = 32 (≥ 30 ✓). Therefore, k = 5 classes.
Step 2 — Calculate Class Width
W = (Largest value − Smallest value) ÷ Number of classes = (14 − 0) ÷ 5 = 2.8 ≈
3
Always round up to ensure no value is left unclassified.
Step 3 — Define Class Boundaries
• Class 1: 0–2
• Class 2: 3–5
• Class 3: 6–8
• Class 4: 9 – 11
• Class 5: 12 – 14
Step 4 — Tally, Count, and Build the Table
Class (Calls/Day) Tally Frequency Relative Freq. Cumulative Freq.
0–2 |||||| 6 20.0% 6
3–5 ||||||||| 9 30.0% 15
6–8 |||||||| 8 26.7% 23
9 – 11 |||| 4 13.3% 27
12 – 14 ||| 3 10.0% 30
Total 30 100% —
Step 5 — Interpret
• Modal class is 3–5 (30%), indicating most people make between 3 and 5 calls per day.
• The distribution is right-skewed — values concentrate at the low end with a tail toward high call
counts.
• 76.7% of individuals make fewer than 9 calls per day (cumulative frequency check).
4. Business Example — Customer Complaints at a Call Centre
The following problem applies the same grouped frequency distribution methodology to a business
operations context. This is particularly relevant in service quality management, where tracking
complaint volumes helps identify demand patterns, allocate staffing, and trigger corrective action
protocols.
The Problem
A retail company's customer service call centre recorded the number of customer complaints
received each day over a period of 30 working days. The data is shown below:
Customer Complaints per Day — Call Centre (n = 30)
3 7 12 18 5 9
14 2 20 11 6 15
0 8 22 4 17 10
13 1 9 24 7 16
5 19 11 3 8 14
Task: Develop a grouped frequency distribution to analyse the daily complaint pattern.
Solution: Step-by-Step
Step 1 — Count and Identify Range
n = 30 observations. Minimum = 0, Maximum = 24. Range = 24 − 0 = 24.
Step 2 — Apply the 2k ≥ n Rule
With n = 30, the same logic applies: 2⁵ = 32 ≥ 30, therefore k = 5 classes.
Step 3 — Calculate Class Width
W = (24 − 0) ÷ 5 = 4.8 ≈ 5 (round up to 5)
Step 4 — Define Class Boundaries
• Class 1: 0 – 4 (0 to 4 complaints per day)
• Class 2: 5 – 9 (5 to 9 complaints per day)
• Class 3: 10 – 14 (10 to 14 complaints per day)
• Class 4: 15 – 19 (15 to 19 complaints per day)
• Class 5: 20 – 24 (20 to 24 complaints per day)
Step 5 — Tally the Data and Build the Frequency Table
Class (Complaints/Day) Tally Frequency Relative Freq. Cumulative Freq.
0–4 |||||| 6 20.0% 6
5–9 ||||||||| 9 30.0% 15
10 – 14 ||||||| 7 23.3% 22
15 – 19 ||||| 5 16.7% 27
20 – 24 ||| 3 10.0% 30
Total 30 100% —
Step 6 — Interpret the Results
Business Insights from the Frequency Distribution:
• The modal class is 5–9 complaints per day (30%), suggesting this is the 'normal' daily complaint
load.
• Days with 0–9 complaints account for 50% of all days — on half the working days, the call
centre receives relatively light complaint loads.
• Heavy complaint days (15–24) account for 26.7% of all days. These require either additional
staffing or an escalation protocol.
• The distribution is right-skewed: most days fall in the lower complaint range, but a minority of
high-complaint days significantly impact operations.
• Management Action: Staff the call centre for 10–14 complaints per day as a standard baseline;
have an overflow protocol ready for the ~27% of days exceeding 14 complaints.
Note: Notice how the same 2k ≥ n rule and class width calculation process is used identically here as in
Problem 1 — the methodology transfers directly across different business contexts.
5. Inclusive vs. Exclusive Methods of Class Boundaries
When constructing a grouped frequency distribution, one important decision is: how do we define the
boundary between two adjacent classes? This determines where a value that sits exactly on a
boundary will be placed. There are two conventions: the Inclusive Method and the Exclusive Method.
5.1 The Inclusive Method
In the Inclusive Method, both the lower and upper limits of each class are included in that class. This
means the classes are written with a visible gap — the upper limit of one class is NOT the same
number as the lower limit of the next class.
Key Characteristics:
• Both boundaries are included: a value equal to either boundary belongs to that class
• There is always a gap of at least 1 unit between the upper limit of one class and the lower limit
of the next
• Class width = Upper Limit − Lower Limit + 1
• Best suited for discrete data — data that takes only whole number values (e.g., marks, counts,
scores)
Example of inclusive classes (width = 10): 40–49, 50–59, 60–69, 70–79, 80–89
A student scoring 49 goes into the 40–49 class. A student scoring 50 goes into the 50–59 class. There
is no ambiguity, and no value falls between the classes.
5.2 The Exclusive Method
In the Exclusive Method, the upper boundary of each class is excluded — that is, a value equal to the
upper limit belongs to the next class, not the current one. Classes are therefore written so that the
upper limit of one class equals the lower limit of the next.
Key Characteristics:
• Lower boundary is included; upper boundary is excluded — written formally as [lower, upper)
• No gap between classes — the upper of one class = lower of the next
• Class width = Upper Limit − Lower Limit
• Best suited for continuous data — data that can take any value including decimals (e.g., height,
weight, time, temperature)
Example of exclusive classes (width = 10): 40–50, 50–60, 60–70, 70–80, 80–90
A value of 50 goes into the 50–60 class (NOT the 40–50 class, since 50 is the excluded upper limit). A
value of 49.7 goes into the 40–50 class.
5.3 Demonstrating Both Methods on the Same Problem
The following problem uses the same dataset to construct a frequency distribution both inclusively and
exclusively, so you can directly compare the two approaches. This makes the structural difference very
concrete.
The Problem — Business Statistics Test Marks
The test scores of 25 MBA students in a Business Statistics examination are recorded below.
Construct a grouped frequency distribution using (a) the Inclusive Method and (b) the Exclusive
Method.
Business Statistics Test Marks — MBA Batch (n = 25)
42 55 63 78 85
49 61 74 88 52
67 71 46 58 80
64 75 89 53 69
43 76 82 57 66
Range: Minimum = 42, Maximum = 89, Range = 89 − 42 = 47
Number of classes: n = 25. Apply 2k ≥ 25 → 2⁵ = 32 ≥ 25 ✓ → k = 5 classes
Class width: W = 47 ÷ 5 = 9.4 ≈ 10 (round up to 10)
Solution (a): Inclusive Method — Classes: 40–49, 50–59, 60–69, 70–79, 80–89
Both boundary values are included in each class. The gap between consecutive class limits (49 then
50, 59 then 60, etc.) eliminates any boundary ambiguity. All marks are whole numbers here, so every
value fits cleanly into exactly one class.
Class (Marks) Tally Frequency Relative Freq. Cumulative Freq.
40 – 49 |||| 4 16.0% 4
50 – 59 ||||| 5 20.0% 9
60 – 69 |||||| 6 24.0% 15
70 – 79 ||||| 5 20.0% 20
80 – 89 ||||| 5 20.0% 25
Total 25 100% —
How to verify: Check that 40 + 49 + 1 = 10 (class width = upper − lower + 1 = 10 ✓). No value can fall
in the gap because our data consists of integers, and the gap between 49 and 50 is exactly 1 unit.
Solution (b): Exclusive Method — Classes: 40–50, 50–60, 60–70, 70–80, 80–90
The upper boundary is excluded from each class. So a mark of exactly 50 belongs to the 50–60 class,
not 40–50. Classes are continuous, with no gap between them — the end of one class is the start of the
next.
Class (Marks) Tally Frequency Relative Freq. Cumulative Freq.
40 – 50 |||| 4 16.0% 4
50 – 60 ||||| 5 20.0% 9
60 – 70 |||||| 6 24.0% 15
70 – 80 ||||| 5 20.0% 20
80 – 90 ||||| 5 20.0% 25
Total 25 100% —
How to verify: Class width = upper − lower = 50 − 40 = 10 ✓. Notice the classes are contiguous (no
gap). A hypothetical score of 59.5 would go into 50–60; a score of 60 would go into 60–70 under this
method.
Important Observation: In this dataset, no student scored exactly 50, 60, 70, or 80 — the boundary
values themselves. This is why both methods produce identical frequencies here (4, 5, 6, 5, 5). In real-
world continuous data where a value like 60.0 is possible, the two methods would produce different
placements — which is precisely why selecting the right method matters.
5.4 Side-by-Side Comparison
Feature Inclusive Method Exclusive Method
Class boundaries Both ends included (e.g., 40–49) Upper end excluded (e.g., 40–50)
Gap between classes Gap of 1 unit between upper and No gap — upper of one = lower of
lower limit next
Data type suited for Discrete (whole numbers, counts) Continuous (heights, weights,
time)
Value at boundary Belongs to the LOWER class Belongs to the UPPER class
Class width Upper limit − Lower limit + 1 Upper limit − Lower limit
Example usage Exam marks, age groups, error Waiting times, blood pressure,
counts rainfall
Practical Rule of Thumb: Use the Inclusive Method for exam marks, age groups, error counts, and
any data that is clearly discrete (whole numbers only). Use the Exclusive Method for weights, heights,
temperatures, waiting times, and any data that is continuous or could include decimal values.
6. Summary: Steps to Construct a Grouped Frequency
Distribution
1. Step 1: Count total observations (n)
2. Step 2: Find the range: Maximum value − Minimum value
3. Step 3: Apply the 2k ≥ n rule to determine the number of classes (k)
4. Step 4: Calculate class width: W = Range ÷ k — always round up
5. Step 5: Choose Inclusive or Exclusive method based on data type (discrete vs. continuous)
6. Step 6: Define class boundaries starting from the minimum value
7. Step 7: Tally each data value into its appropriate class
8. Step 8: Count tallies to record frequency (f) for each class
9. Step 9: Calculate relative frequency = f ÷ n × 100
10. Step 10: Calculate cumulative frequency (running total of f)
11. Step 11: Interpret: identify modal class, skewness, and operational insights
7. Healthcare Application
In healthcare operations management, frequency distributions are a fundamental descriptive
statistics tool. They convert raw operational data into actionable management intelligence:
• Patient waiting time analysis: Group times into classes (0–10 min, 11–20 min, etc.) to identify
OPD bottlenecks — use the Exclusive Method since time is continuous.
• Medication error tracking: Tally errors per ward per week using the Inclusive Method (discrete
counts) to prioritise quality improvement interventions.
• Bed occupancy patterns: Analyse daily occupancy numbers to optimise staffing schedules —
modal class indicates the most common occupancy level.
• Readmission frequency: Group 30-day readmission counts by diagnosis to design targeted
discharge planning protocols.
• Customer complaint management (as in Section 4): Apply to any service environment to design
staffing and escalation protocols based on complaint load distribution.
Selecting the correct boundary method matters even in healthcare: waiting time in minutes
(continuous) should use the Exclusive Method, while number of medication doses (discrete) should
use the Inclusive Method.