MMPC-005: Quantitative Analysis For Managerial Applications

INDIRA GANDHI NATIONAL OPEN UNIVERSITY

School of Management Studies
Master of Business Administration (MBA)
MMPC-005 — Quantitative Analysis For Managerial Applications

Assignment Code: MMPC-005/TMA/JULY/2024

Coverage: All Blocks
Session: January 2025
Last Date of Submission: 30th April, 2025

Question 1: Questionnaire Method of Collecting Primary Data

1.1 — The Questionnaire Method
The questionnaire method is one of the most widely used techniques for collecting primary
data in business and social research. A questionnaire is a structured set of questions,
presented in a definite sequence, designed to elicit specific information from respondents. It
may be administered directly (self-completion), through a personal interview, by mail,
telephone, or increasingly via online platforms. The method is particularly valuable when the
researcher needs standardised data from a large, geographically dispersed sample within a
limited time frame.
In this method, the researcher formulates questions that correspond precisely to the research
objectives. Respondents fill in the questionnaire independently or with the assistance of an
investigator. The data collected is typically amenable to quantitative analysis, which makes
the questionnaire method especially suitable for statistical inference and managerial decision-
making.

1.2 — Types of Questionnaires

Questionnaires can be broadly classified into two types based on question structure:
Structured Questionnaires: These contain closed-ended questions with fixed response
categories (e.g., Yes/No, multiple choice, Likert scale). They are easy to administer, code,
and analyse statistically. They are best suited for large-scale surveys.
Unstructured Questionnaires: These contain open-ended questions that allow respondents
to answer in their own words, yielding richer qualitative insights. They are more difficult to
analyse but capture nuanced opinions.
Semi-structured Questionnaires: A combination of both, offering flexibility while
maintaining some standardisation.

1.3 — Essentials of a Good Questionnaire

The quality of data collected through a questionnaire depends largely on how well it is
designed. A good questionnaire must satisfy the following essential criteria:
(i) Clarity and Simplicity: Each question must be worded clearly, using plain language that
every respondent can understand. Ambiguous, vague, or technical terms should be avoided
unless the target audience is a specialist group. The meaning of each question must be
unambiguous to all respondents.

IGNOU – School of Management Studies Page 1

MMPC-005: Quantitative Analysis For Managerial Applications

(ii) Relevance: Every question must be directly relevant to the research objectives. Irrelevant
questions increase the length of the questionnaire, reduce response rates, and dilute data
quality. A useful test is to ask: 'What will I do with the answer to this question?' If no clear
use exists, the question should be dropped.
(iii) Brevity: A good questionnaire is as short as possible while still capturing all necessary
information. Respondents are more likely to complete a concise questionnaire. Length should
be determined by the scope of the study, not by the researcher's curiosity.
(iv) Logical Sequencing: Questions should follow a logical order. They should begin with
simple, non-threatening questions to build respondent confidence, progress to more specific
or sensitive questions, and end with personal or demographic items. This flow, often called
the 'funnel approach,' enhances completion rates.
(v) Avoidance of Leading and Loaded Questions: Questions must not suggest a desired
answer (leading) or contain emotionally charged language (loaded). For example, 'Don't you
think the management is unfair?' is both leading and loaded. A neutral phrasing such as 'How
would you rate the management's fairness?' is preferable.
(vi) Avoidance of Double-Barrelled Questions: Each question should address only one
issue. A double-barrelled question such as 'Are you satisfied with the salary and working
conditions?' should be split into two separate questions.
(vii) Appropriate Question Format: The choice between open-ended and closed-ended
formats should match the information required. Closed questions are preferable for factual or
categorical data; open questions are better for opinions or explanations.
(viii) Pre-testing (Pilot Study): Before full deployment, the questionnaire should be tested
on a small sample representative of the target population. Pre-testing identifies ambiguities,
poor question sequencing, missing response categories, and time estimation issues. Revisions
based on pilot feedback significantly improve data quality.
(ix) Instructions and Layout: Clear instructions for completing each section must be
provided. The layout should be clean, with adequate spacing, consistent formatting, and
logical grouping of related questions. A professional appearance encourages completion.
(x) Confidentiality Assurance: Respondents are more likely to answer honestly if they are
assured that their responses will remain confidential or anonymous. The questionnaire should
state clearly how the data will be used and protected.

In summary, a well-designed questionnaire is the researcher's most powerful tool for

collecting reliable, valid, and analysable primary data. Investing time in thoughtful design,
careful wording, and rigorous pre-testing pays dividends in the quality of insights generated.

Question 2: Importance of Measuring Variability for Managerial

Decision-Making
2.1 — Introduction to Variability
Measures of central tendency — mean, median, and mode — describe the 'typical' or
'average' value in a data set. However, they tell only part of the story. Two data sets can have
identical means yet exhibit vastly different distributions. Variability (also called dispersion or
spread) measures the degree to which data values deviate from the central value. The
principal measures of variability include the Range, Mean Absolute Deviation (MAD),
Variance, Standard Deviation, and Coefficient of Variation.

IGNOU – School of Management Studies Page 2

MMPC-005: Quantitative Analysis For Managerial Applications

For managers, knowing the average is rarely sufficient for sound decision-making.
Understanding variability is equally — and often more — important, because it reflects risk,
reliability, and consistency.

2.2 — Significance of Variability in Managerial Decisions

(i) Risk Assessment and Financial Planning: In finance, the standard deviation of returns is
the most widely used measure of investment risk. Two investment portfolios with identical
expected returns may carry very different risk profiles. A manager choosing between them
must examine variability: a high standard deviation signals volatile, unpredictable returns,
while a low standard deviation indicates stability. Without measuring variability, financial
risk is invisible.
(ii) Quality Control and Process Management: In manufacturing and operations,
variability in product dimensions, weights, or performance characteristics directly reflects
process quality. Statistical Process Control (SPC) uses standard deviation to set control limits.
A production process with high variability produces more defects and customer complaints.
Managers use variability data to identify when a process is 'out of control' and intervene
before defects reach customers.
(iii) Human Resource Management: Variability in employee performance metrics —
productivity, attendance, sales achievement — helps HR managers design targeted
interventions. High variability in performance may indicate inconsistent supervision, unclear
expectations, or motivational issues. Understanding dispersion guides decisions about
training, appraisal calibration, and incentive design.
(iv) Demand Forecasting and Inventory Management: In supply chain management,
demand variability determines optimal safety stock levels. High demand variability requires
larger safety stocks to prevent stockouts, increasing holding costs. Managers must balance the
cost of excess inventory against the cost of stockouts — a trade-off that cannot be quantified
without measuring demand variability.
(v) Comparing Relative Variability Across Different Units: The Coefficient of Variation
(CV = Standard Deviation / Mean × 100) enables managers to compare variability between
data sets measured in different units or with different means. For example, comparing the
consistency of two suppliers — one supplying components at an average of ₹500 and another
at ₹5,000 — requires relative variability, not absolute variability.
(vi) Setting Realistic Targets and Tolerances: Understanding variability helps managers set
achievable performance targets and realistic tolerance bands. Setting a target at the mean
without accounting for standard deviation will result in roughly half the workforce or
processes being classified as 'below target' even in a well-functioning system. Variability-
informed targets are fairer and more motivating.

2.3 — Conclusion
In conclusion, measures of variability transform raw averages into actionable intelligence.
They enable managers to assess risk, ensure quality, optimise inventories, benchmark
performance, and make decisions that are robust to uncertainty. A manager who ignores
variability is like a navigator who knows the average depth of a river but not the variance —
liable to drown in a spot that averages three feet deep. Sound quantitative analysis demands
attention to both central tendency and dispersion.

IGNOU – School of Management Studies Page 3

MMPC-005: Quantitative Analysis For Managerial Applications

Question 3: Probability — Stock Price Movement

3.1 — Given Information
An investment consultant provides the following odds:
• Odds against the price going UP: 2 : 1
• Odds in favour of the price remaining the SAME: 1 : 3
We are required to find the probability that the stock price will go DOWN during the next
week.

3.2 — Converting Odds to Probabilities

Recall: If the odds against an event E are a : b, then P(E) = b / (a + b).
If the odds in favour of an event E are a : b, then P(E) = a / (a + b).
Step 1 — Probability that price goes UP:
Odds against price going up = 2 : 1
P(Up) = 1 / (2 + 1) = 1/3 ≈ 0.3333
Step 2 — Probability that price remains the SAME:
Odds in favour of price remaining the same = 1 : 3
P(Same) = 1 / (1 + 3) = 1/4 = 0.25
Step 3 — Probability that price goes DOWN:
Since the three outcomes (Up, Same, Down) are mutually exclusive and exhaustive, their
probabilities must sum to 1:
P(Up) + P(Same) + P(Down) = 1
1/3 + 1/4 + P(Down) = 1
P(Down) = 1 - 1/3 - 1/4
P(Down) = 12/12 - 4/12 - 3/12
P(Down) = 5/12 ≈ 0.4167

3.3 — Summary Table

Event Probability
Price goes UP 1/3 ≈ 0.3333
Price remains SAME 1/4 = 0.2500
Price goes DOWN 5/12 ≈ 0.4167
Total 12/12 = 1.0000

3.4 — Conclusion
The probability that the stock price will go DOWN during the next week is 5/12 ≈ 0.4167
(approximately 41.67%). This is the highest single probability among the three outcomes,

IGNOU – School of Management Studies Page 4

MMPC-005: Quantitative Analysis For Managerial Applications

indicating that the price decline is the most likely scenario according to the consultant's
assessment.

Question 4: Decision-Making Under Uncertainty — Criteria Without

Probabilities
4.1 — Decision-Making Under Uncertainty
In many real-world situations, the probabilities of various outcomes cannot be objectively
determined — either because historical data is unavailable, the situation is entirely novel, or
experts disagree fundamentally. This condition is known as decision-making under
uncertainty (as opposed to decision-making under risk, where probabilities are known). Four
classical criteria guide decision-making in such situations:

4.2 — Criteria for Decision-Making Under Uncertainty

(i) Maximax Criterion (Criterion of Optimism): The decision-maker selects the alternative
that offers the maximum of the maximum payoffs (the 'best of the best'). This criterion is
used by optimists who assume that the best possible outcome will occur. For each alternative,
identify the maximum payoff; then choose the alternative with the highest of these maximum
payoffs. While this can lead to bold, high-reward choices, it ignores the risk of poor outcomes
and is unsuitable for risk-averse decision-makers.
(ii) Maximin Criterion (Criterion of Pessimism / Wald's Criterion): The decision-maker
identifies the minimum (worst) payoff for each alternative and then selects the alternative
with the maximum of these minimum payoffs — the 'best of the worst.' This is a
conservative, risk-averse approach attributed to Abraham Wald. It is appropriate when the
decision-maker wants to avoid the worst possible outcome at all costs. It is criticised for
being overly cautious and potentially forgoing significant gains.
(iii) Minimax Regret Criterion (Savage's Criterion): This criterion minimises the
maximum regret (opportunity loss) the decision-maker might experience. Regret is defined as
the difference between the actual payoff achieved and the best payoff that could have been
achieved under the same state of nature. A regret matrix is constructed from the original
payoff table; the decision-maker then selects the alternative with the minimum of the
maximum regrets. This criterion is psychologically motivated — it minimises post-decision
disappointment.
(iv) Laplace Criterion (Criterion of Insufficient Reason / Criterion of Equal
Likelihood): When no information is available about the probabilities of states of nature, the
Laplace criterion assumes equal probability for all states (invoking the principle of
insufficient reason, attributed to Laplace). The expected payoff of each alternative is
computed as the simple average of its payoffs across all states. The alternative with the
highest average payoff is selected. This criterion transforms the uncertainty problem into a
risk problem by assigning uniform probabilities, and is intuitive but controversial — it treats
all unknown outcomes as equally likely, which may not reflect reality.

4.3 — Illustrative Example

Consider a manager choosing between three investment strategies (A, B, C) across three
possible economic conditions (Boom, Stable, Recession), with the following payoff matrix (₹
Lakhs):

IGNOU – School of Management Studies Page 5

MMPC-005: Quantitative Analysis For Managerial Applications

Strategy Boom Stable Recession

A 50 30 −10
B 40 35 10
C 30 25 20

Maximax: Max payoffs: A=50, B=40, C=30 → Choose Strategy A (highest maximum)
Maximin: Min payoffs: A=−10, B=10, C=20 → Choose Strategy C (highest minimum = 20)
Laplace: Averages: A=(50+30−10)/3=23.33; B=(40+35+10)/3=28.33; C=(30+25+20)/3=25
→ Choose Strategy B

4.4 — Conclusion
The choice of criterion depends on the decision-maker's attitude toward risk. Optimists
favour Maximax; pessimists prefer Maximin; those concerned with regret use Minimax
Regret; and those who treat all outcomes as equally probable apply the Laplace criterion. In
practice, a prudent manager considers multiple criteria before making a final decision,
especially when the stakes are high and information is limited.

Question 5: Hypothesis Testing — Hardness of Castings

5.1 — Problem Statement
The hardness of castings from any supplier is known to be normally distributed with:
Population Mean (μ₀): 20.25
Population Standard Deviation (σ): 2.5
A purchase manager picks 100 samples (n = 100) from a particular supplier who claims that
his castings have heavier hardness (i.e., higher mean hardness). The sample mean is found to
be x̄ = 20.50.
We must test whether the supplier's claim of higher hardness is tenable at the standard 5%
level of significance.

5.2 — Setting Up the Hypotheses

Since the supplier claims that his castings have heavier (higher) hardness, this is a one-tailed
(right-tailed) test.
H₀: μ = 20.25 (Null Hypothesis — supplier's castings are no
different)
H₁: μ > 20.25 (Alternative Hypothesis — supplier's castings
have higher hardness)

5.3 — Test Statistic

Since the population standard deviation (σ) is known and n = 100 (large sample), we use the
Z-test:
Z = (x̄ − μ₀) / (σ / √n)
Substituting the values:

IGNOU – School of Management Studies Page 6

MMPC-005: Quantitative Analysis For Managerial Applications

Z = (20.50 − 20.25) / (2.5 / √100)

Z = 0.25 / (2.5 / 10)
Z = 0.25 / 0.25
Z = 1.00

5.4 — Critical Value and Decision Rule

For a one-tailed (right-tailed) test at α = 0.05, the critical value from the standard normal
distribution is:
Z_critical = 1.645
Decision Rule: Reject H₀ if Z_calculated > Z_critical (i.e., if Z > 1.645).

5.5 — Decision
Since the calculated Z-value (1.00) is less than the critical value (1.645):
Z_calculated (1.00) < Z_critical (1.645)
We fail to reject the Null Hypothesis (H₀).

5.6 — Conclusion
At a 5% level of significance, there is insufficient statistical evidence to support the supplier's
claim that his castings have heavier (higher mean) hardness than the standard population
mean of 20.25. The observed sample mean of 20.50, while slightly higher than the population
mean, could plausibly have arisen due to random sampling variation alone. The supplier's
claim is therefore NOT tenable at the 5% level of significance.

Note: If we were to test at a more lenient 10% significance level (Z_critical = 1.28), the result
would be the same — Z = 1.00 < 1.28 — still failing to reject H₀. The supplier's claim lacks
sufficient statistical support.

References
Anderson, D.R., Sweeney, D.J. & Williams, T.A. (2018). Statistics for Business and Economics (13th
ed.). Mason: Cengage Learning.
Gupta, C.B. & Gupta, V. (2019). An Introduction to Statistical Methods (23rd ed.). New Delhi: Vikas
Publishing House.
Levin, R.I. & Rubin, D.S. (2010). Statistics for Management (7th ed.). New Delhi: Pearson Education.
Raghavachari, M. (2001). Operations Research: A Practical Approach. New Delhi: Tata McGraw-
Hill.
Sharma, J.K. (2016). Business Statistics (3rd ed.). New Delhi: Pearson Education.
Taha, H.A. (2017). Operations Research: An Introduction (10th ed.). Harlow: Pearson Education.

IGNOU – School of Management Studies Page 7