Easy2Siksha

GNDU Question Paper-2022

B.A 5

Semester

QUANTITATIVE TECHNIQUES

(Quantitative Technqiues-V)

Time Allowed: Three Hours Max. Marks: 100

Note: Attempt Five questions in all, selecting at least One question from each section. The

Fifth question may be attempted from any section. All questions carry equal marks.

SECTION-A

1. What is the difference between estimate and an estimator ? What are the characteristic

features of a good estimator? Explain in detail.

2. What are the basic features of Binomial Distribution? Find the probability of three

heads in six flips of an unbiased coin by using binomial distribution.

SECTION-B

3. What do you understand by Z-distribution? Highlight the basic properties of Z-

distribution.

4. What are the characteristics features of F-distribution ?

SECTION-C

5. A total of 10 students were evaluated on the basis of their performance based on a test

with 50 maximum marks. They were given coaching for 6 months and again evaluated.

Easy2Siksha

The results are given below. Can you conclude that the students have benefitted by

the extra coaching?

Student

Code

III

Marks:

Before

coaching

After

coaching

Student

Code

VII

VIII

Marks:

Before

coaching

After

coaching

6. What is the difference between Parametric and Non-parametric Test? The Chi square

test falls under which category? Explain the mechanics of this test.

SECTION-D

7. Highlight the difference between One way and Two way ANOVA technique. Describe

the method of working out the critical difference for both.

8. In order to assess the significance of possible variation in productivity performance of

three different varieties (X, Y, Z) of wheat, an experiment was conducted on 12 equi-sized

plots. The related data is given below. Frame the hypothesis and test if there is any

significant difference in the productivity of these three varieties.

Easy2Siksha

GNDU Answer Paper-2022

B.A 5

Semester

QUANTITATIVE TECHNIQUES

(Quantitative Technqiues-V)

Time Allowed: Three Hours Max. Marks: 100

Note: Attempt Five questions in all, selecting at least One question from each section. The

Fifth question may be attempted from any section. All questions carry equal marks.

SECTION-A

1. What is the difference between estimate and an estimator ? What are the characteristic

features of a good estimator? Explain in detail.

Ans: What is the Difference Between Estimate and Estimator? What are the Characteristic

Features of a Good Estimator?

Lets imagine youre part of a team of treasure hunters trying to find a buried treasure on an

island. The only clue you have is a map with random dots, representing possible locations.

You can't dig everywhere, so you must make a smart guess based on the available

information. In statistics, this "smart guess" is what we call an estimate, and the method or

rule we use to make that guess is known as an estimator.

Lets understand these two terms properly and then move to what makes an estimator a

good one.

󹰤󹰥󹰦󹰧󹰨 1. Understanding Estimate and Estimator (With Examples)

󷃆󼽢 Estimator The Rule or Method (The Treasure Hunting Strategy)

An estimator is a mathematical rule or formula used to calculate an unknown value

(parameter) of a population based on the sample data.

Easy2Siksha

Think of it like a recipe you use to bake a cake. The recipe tells you what to do with the

ingredients. Similarly, the estimator tells you how to use the sample data to estimate

something about the entire population.

For example:

Sample Mean () is an estimator of the population mean (μ).

Sample Proportion (p



) is an estimator of the population proportion (p).

Sample Variance (s²) is an estimator of the population variance (σ²).

󷃆󼽢 Estimate The Result or Value (The Final Guess Where to Dig)

An estimate is the numerical value you get when you apply the estimator to a specific

sample.

Continuing with the treasure analogy, the estimator is your strategy, but the estimate is the

spot you choose to dig.

Example:

Lets say you collect data on the heights of 100 students to estimate the average height of

all students in your college.

Your estimator: Sample mean formula →  = (Σx)/n

Your estimate:  = 165.3 cm (actual number after calculation)

So,

Term

Meaning

Example

Estimator

Rule or formula (statistical method)

Sample mean  = Σx/n

Estimate

Actual calculated value from sample

165.3 cm

󹸯󹸭󹸮 2. Types of Estimators

There are two major types:

󹳴󹳵󹳶󹳷 Point Estimator:

It gives a single value estimate of a parameter.

E.g., Sample mean to estimate population mean.

󹳴󹳵󹳶󹳷 Interval Estimator:

It gives a range of values within which the parameter is likely to lie.

E.g., "We estimate the population mean is between 160 cm and 170 cm with 95%

confidence."

Easy2Siksha

󷗭󷗨󷗩󷗪󷗫󷗬 3. Characteristics of a Good Estimator (What Makes an Estimator Reliable?)

Now imagine you have multiple maps (different estimators). Which one will lead you to the

treasure most accurately? Thats where the qualities of a good estimator come into play.

A good estimator must have the following key properties:

󹻀 A. Unbiasedness "No Favoritism"

An estimator is said to be unbiased if its average value (expected value) is equal to the true

value of the parameter it is estimating.

󹻂 Meaning: It does not consistently overestimate or underestimate the actual value.

󹳴󹳵󹳶󹳷 Example:

If the true average weight of students is 60 kg and you keep taking many samples and

calculating the average weight, a good (unbiased) estimator will give you an average close

to 60 kg over time.

󷃆󼽢 A good estimator always hits the target on average.

󹻀 B. Consistency "The More You Try, The Better You Get"

An estimator is said to be consistent if it gives values that get closer to the true parameter

as the sample size increases.

󹻂 Meaning: As your sample becomes larger, your estimate becomes more accurate.

󹳴󹳵󹳶󹳷 Example:

Estimating the population height using a sample of 10 people may not be accurate, but

using 1,000 people will give a much closer result.

󷃆󼽢 A consistent estimator improves with more data.

󹻀 C. Efficiency "The Best Shot Among Many"

Out of all the unbiased estimators, the one with the smallest variance (least fluctuation in

repeated sampling) is considered the most efficient.

󹻂 Meaning: Among many estimators, the one that gives estimates closer to each other

and to the true value is the best.

Easy2Siksha

󹳴󹳵󹳶󹳷 Example:

If Estimator A gives values around 50, 52, 48, and Estimator B gives 60, 40, 70, then

Estimator A is more efficient.

󷃆󼽢 Less spread = More reliable = More efficient

󹻀 D. Sufficiency "Using All the Clues"

An estimator is sufficient if it uses all the available information in the sample to estimate the

parameter.

󹻂 Meaning: No information in the data is wasted.

󹳴󹳵󹳶󹳷 Example:

The sample mean is a sufficient estimator for the population mean when data follows a

normal distribution.

󷃆󼽢 A sufficient estimator doesn't leave anything important behind.

󹻀 E. Robustness "Staying Strong Against Outliers"

An estimator is robust if it is not too sensitive to extreme values or outliers.

󹻂 Meaning: Even if the data has some errors or unexpected values, a robust estimator will

still give a reliable estimate.

󹳴󹳵󹳶󹳷 Example:

The median is a more robust estimator than the mean because it is not affected much by

extremely large or small values.

󷃆󼽢 A robust estimator remains stable even with unusual data.

󹴡󹴵󹴣󹴤 4. Real-Life Example to Wrap Up

Lets take a real-world scenario. A university wants to estimate the average marks of

students in a semester. They can't check all students, so they collect marks from a sample of

100 students.

They use the sample mean () as an estimator.

After calculation, they get  = 72.5 marks, which is the estimate.

Now, how do we know this is a good estimator?

We check:

Easy2Siksha

Quality

Question

Answer

Unbiased

Is the average of many such estimates 72.5?

Yes → Unbiased

Consistent

Does it get better with more student samples?

Yes → Consistent

Efficient

Does it give values close to the true average?

Yes → Efficient

Sufficient

Does it use all the data collected?

Yes → Sufficient

Robust

Is it stable even if one or two marks are odd?

Maybe not (mean isn't robust)

Hence, if extreme marks are not a problem, sample mean is a good estimator.

󹲹󹲺󹲻󹲼󹵉󹵊󹵋󹵌󹵍 5. Conclusion

To summarize:

An estimator is a method or formula; an estimate is the actual value obtained.

A good estimator must be:

o Unbiased: No systematic error.

o Consistent: Improves with larger samples.

o Efficient: Least variance among all.

o Sufficient: Uses all sample information.

o Robust: Resistant to outliers (optional but helpful).

These characteristics ensure that the conclusions drawn from a sample are as close and

useful as possible to understanding the whole population.

2. What are the basic features of Binomial Distribution? Find the probability of three

heads in six flips of an unbiased coin by using binomial distribution.

Ans: What is Binomial Distribution? A Simple Story-Based Explanation

Imagine you are playing a game with your friend. You toss a coin six times and bet that

exactly three heads will come up. Now youre curious whats the mathematical probability

that you win this bet?

To solve this, we use a concept from probability theory called the Binomial Distribution.

Easy2Siksha

Why is it called Binomial?

The word Binomial comes from bi meaning two, and nomial referring to terms. So

binomial experiments are based on outcomes that have only two possibilities like:

Heads or Tails

Success or Failure

Yes or No

Each time you perform the experiment (like tossing a coin), it is called a trial.

Basic Features of Binomial Distribution

To really understand Binomial Distribution, lets go over its basic features step-by-step,

using real-life examples:

1. Fixed Number of Trials (n)

You perform the experiment a set number of times. In our case, you toss the coin 6 times,

so n = 6.

󷵻󷵼󷵽󷵾 Think of it like playing 6 rounds of a game. The total number of times you try doesnt

change.

2. Only Two Outcomes per Trial

Each trial must have exactly two possible results:

Success (e.g., getting a Head)

Failure (e.g., getting a Tail)

󷵻󷵼󷵽󷵾 You can't have a third outcome. For example, in a coin toss, the coin can't land standing

up!

3. Constant Probability of Success (p)

The probability of success (like getting a Head) stays the same for each trial.

If the coin is unbiased (fair), the probability of getting a head (success) is:

p = 0.5

So, the probability of failure (getting tails) is:

q = 1 - p = 0.5

Easy2Siksha

󷵻󷵼󷵽󷵾 This is like using the same coin for every toss the chances dont change.

4. Trials are Independent

The outcome of one toss does not affect the outcome of another.

󷵻󷵼󷵽󷵾 Just because you got a head on the first toss, doesnt mean youre more or less likely to

get heads on the second toss.

Each toss is like a fresh start.

5. Discrete Distribution

Binomial distribution is a discrete probability distribution, which means it deals with whole

number counts (like 0, 1, 2, n successes).

󷵻󷵼󷵽󷵾 You cant have 2.5 heads in a coin toss it has to be a whole number.

6. Mathematical Formula

The Binomial Probability Formula tells us the probability of getting exactly k successes in n

trials:



󰇛



󰇜

󰇡





󰇢



 



Where:

󰇡





󰇢is n choose k, or combinations =





󰇛



󰇜



n = number of trials

k = number of successes (e.g., number of heads)

p = probability of success

q = 1 p = probability of failure

Now Lets Apply This to Our Problem:

Question:

Find the probability of getting exactly three heads in six flips of an unbiased coin.

So from the problem:

n = 6 (total coin tosses)

Easy2Siksha

k = 3 (number of heads we want)

p = 0.5 (probability of getting heads)

q = 0.5 (probability of getting tails)

Step-by-Step Calculation:

Lets plug everything into the formula:



󰇛



󰇜

󰇡





󰇢 

󰇛



󰇜





󰇛



󰇜



Step 1: Calculate combinations 󰇡





󰇢

󰇡





󰇢





󰇛

  

󰇜





  

    



Step 2: Calculate powers of p and q

(0.5)

= 0.125 and (0.5)

= 0.125

Step 3: Multiply all together



󰇛



󰇜

   

󷃆󼽢 Final Answer:

The probability of getting exactly 3 heads in 6 coin tosses is 0.3125, or 31.25%.

󷗭󷗨󷗩󷗪󷗫󷗬 Lets Understand This Visually (Optional)

If you want to imagine the situation better, think of all the possible outcomes of flipping the

coin 6 times. There are 26=642^6 = 6426=64 total combinations (like HHTTHT, HTHTHH,

etc.).

Out of those 64 outcomes, 20 will contain exactly 3 heads and 20/64 = 0.3125.

Why is Binomial Distribution Important?

Binomial distribution is used widely in:

Business: Calculating customer response rates

Easy2Siksha

Biology: Finding probabilities of inheriting genes

Medicine: Testing success rates of treatments

Psychology: Analyzing behavior experiments

Quality control: Checking how many items in a batch are defective

In Short Key Takeaways for University Students

Feature

Explanation

Fixed trials (n)

Number of times the experiment is done (e.g., 6 tosses)

Two outcomes per trial

Like success/failure or heads/tails

Same probability (p)

Probability of success remains constant (e.g., 0.5)

Independent trials

One trial doesnt affect another

Whole number outcomes

No fractions, only full number of successes

Use of combinations

To count how many ways k successes can happen in n trials

Real-life application

Used in sciences, finance, psychology, and more

󷇴󷇵󷇶󷇷󷇸󷇹 Conclusion

Binomial distribution is like a smart calculator that helps us understand how likely a certain

number of "wins" or "successes" is when we repeat the same test multiple times. In our

example, we found that getting exactly 3 heads in 6 fair coin tosses has a probability of

0.3125 or 31.25%.

Whether youre studying statistics, biology, or economics, this concept will pop up often.

And now, youre ready to handle it with ease.

SECTION-B

3. What do you understand by Z-distribution? Highlight the basic properties of Z-

distribution.

Ans: Introduction A Story to Begin With

Easy2Siksha

Imagine you're a university student who just took a test with your entire class. When your

teacher hands back the results, you're curious: Did I do better than most students? How far

was I from the average? But here's the problem: your class is huge, and everyone got

different scores.

Now, your teacher tells you:

"Your test score is 1.5 standard deviations above the mean."

What does that even mean? To understand this, youll need to understand a concept from

statistics called the Z-distribution.

Lets dive into it one step at a time and see how this powerful tool helps us make sense

of data in everyday life.

1. What is Z-distribution?

Z-distribution is a special kind of normal distribution, also known as the standard normal

distribution. It is used in statistics to describe how individual values relate to the average

(mean) of a data set, when the data is normally distributed.

Heres the core idea:

Z-distribution is a normal distribution with a mean (average) of 0 and a standard deviation

of 1.

This distribution helps us standardize scores from different normal distributions so they can

be compared on the same scale.

2. Understanding with a Real-Life Analogy

Imagine two students Riya and Arjun who took tests in different subjects:

Riya scored 85 out of 100 in Math, where the class average was 80, and the standard

deviation was 5.

Arjun scored 78 out of 100 in History, where the class average was 70, and the

standard deviation was 4.

At first glance, Riya's score (85) is higher than Arjun's (78). But is that enough to say Riya

performed better compared to her class?

Lets find the Z-scores (more on that soon):

Riya: Z = (85 - 80) / 5 = 1.0

Arjun: Z = (78 - 70) / 4 = 2.0

Easy2Siksha

Arjun has a higher Z-score, meaning his performance was further above average compared

to Riya. This is the power of the Z-distribution it standardizes scores across different

conditions.

3. What is a Z-score?

A Z-score (also called a standard score) tells us how many standard deviations an individual

data point is from the mean of the distribution.

Formula:



󰇛

 

󰇜



Where:

X = individual score

μ = mean (average)

σ = standard deviation

Interpretation:

If Z = 0, the score is exactly at the average.

If Z > 0, the score is above the average.

If Z < 0, the score is below the average.

4. Why is Z-distribution important for university students?

As university students, whether you're studying psychology, economics, sociology, or even

education, youll often collect or analyze data. You might need to compare scores,

understand outliers, or interpret test results. Z-distribution helps in:

Comparing performances

Finding probabilities

Conducting hypothesis testing

Standardizing different data sets

Lets now look at the features or properties of Z-distribution in detail.

5. Basic Properties of Z-Distribution

(1) Mean is 0

Easy2Siksha

This means that after standardizing, the center of the Z-distribution is always 0. So, if a score

has a Z-value of 0, it means it's right at the average.

(2) Standard Deviation is 1

All scores are measured in terms of standard deviations from the mean. A Z-score of +1

means 1 standard deviation above the mean, and a Z-score of -2 means 2 standard

deviations below the mean.

(3) Bell-shaped Curve

The Z-distribution follows the famous bell-shaped curve of the normal distribution. This

means:

Most scores are close to the mean.

Fewer scores are found at the extremes.

(4) Symmetrical

The Z-distribution is perfectly symmetrical around the mean (0). So, the area under the

curve on the left of 0 is equal to the area on the right each side is 50% of the total area.

(5) Total Area under the Curve = 1

This property means that if you calculate the area under the curve, it will always be 1 (or

100%). This area is used to find probabilities.

(6) Empirical Rule (68-95-99.7 Rule)

In a Z-distribution:

About 68% of data lies within ±1 standard deviation (Z = -1 to +1).

About 95% lies within ±2 standard deviations (Z = -2 to +2).

About 99.7% lies within ±3 standard deviations (Z = -3 to +3).

This rule helps you understand how typical or rare a particular score is.

6. Applications of Z-distribution

(i) In Research and Hypothesis Testing

When researchers want to test a theory or hypothesis (for example, Does a new teaching

method improve scores?), they often use the Z-distribution to determine if the results are

statistically significant.

(ii) In Psychology

Psychologists use Z-scores to compare individuals scores on different psychological tests or

scales (like IQ or depression ratings).

Easy2Siksha

(iii) In Economics

Economists use it to analyze large data sets, like income levels, inflation patterns, or

consumer behavior, and compare across different groups.

(iv) In Academic Testing

Standardized tests like SAT, GRE, and other entrance exams often report scores as

percentiles or Z-scores, helping to rank students accurately.

7. Advantages of Using Z-distribution

Makes different types of data comparable.

Helps find percentiles and probabilities easily.

Used in many standard tables and software tools (like Excel, SPSS, R).

Makes it easier to visualize the relative position of a score in a data set.

8. Final Summary Why You Should Care

Think of the Z-distribution as the universal language of comparison in statistics. Whether

you're comparing student scores, research data, or economic values, the Z-distribution

levels the playing field by telling you how far something is from the average in a

standardized, meaningful way.

So the next time someone says, Your Z-score is 2, youll know:

You are two standard deviations above the mean,

Youve performed better than roughly 97.5% of others (based on the standard Z-

table),

And your performance is exceptional.

In short, Z-distribution is like a statistical compass it tells you where you are in a data set

and how far youve come.

4. What are the characteristics features of F-distribution ?

Ans: 󹴡󹴵󹴣󹴤 Understanding the F-Distribution: A Statistical Story

Imagine you're a university student studying statistics, and youre exploring the world of

hypothesis testing. One day, your professor introduces you to a curious new character in the

world of probability distributions the F-distribution. Lets take a walk into this world and

get to know this character closely.

Easy2Siksha

󷗛󷗜 What is the F-Distribution?

The F-distribution is a type of probability distribution that arises when we compare two

sample variances. In simpler terms, its a statistical tool used to test if two different groups

have similar variability (spread or dispersion of data).

Lets say you have two classes of students and you want to check if the variability in their

marks is the same. You can use the F-distribution to do that.

Its mainly used in:

ANOVA (Analysis of Variance)

Regression Analysis

Comparing two variances

But before diving deep, lets get to know this distribution's unique characteristics its

features, behaviors, and traits like youre getting to know someone personally.

󷇴󷇵󷇶󷇷󷇸󷇹 Characteristics Features of F-distribution

1. 󹳣󹳤󹳥 It is a Right-Skewed Distribution

Imagine a graph that starts high on the left and slowly trails off to the right thats the

shape of an F-distribution.

Why? Because it only includes positive values (more on that below), and extreme

values on the right are more likely than extreme values on the left.

In most practical uses, especially with small sample sizes, the F-distribution is

positively skewed.

󷵻󷵼󷵽󷵾 Analogy: Think of exam scores where most students score low to moderate marks, but a

few toppers score very high pulling the average to the right.

2. 󽅐 It Only Takes Positive Values

The F-distribution starts at zero and goes on forever towards the right.

Why cant it be negative? Because it is the ratio of two variances, and variance is

always a non-negative quantity (you cant have negative variability).

For example, if youre dividing one positive number by another, the result will always be

positive.

Easy2Siksha

3. 󷘄󷘅󷘆󷘇󷘈󷘉 It Depends on Two Different Degrees of Freedom (df1 and df2)

Unlike the normal distribution, which is defined by mean and standard deviation, the F-

distribution is defined by two parameters:

df1 (numerator degrees of freedom) related to the first sample

df2 (denominator degrees of freedom) related to the second sample

󷵻󷵼󷵽󷵾 These degrees of freedom come from the sample sizes you are comparing.

Example:

If you're comparing the variances of two groups with sizes n₁ = 10 and n₂ = 15, then:

df1 = n₁ - 1 = 9

df2 = n₂ - 1 = 14

The shape of the F-distribution changes based on these degrees of freedom.

4. 󷃆󹸊󹸋 It Is Not Symmetrical

The F-distribution is asymmetrical. This means:

It is not a mirror image on both sides.

It has a long tail on the right, especially with small degrees of freedom.

As df1 and df2 increase, the curve becomes more bell-shaped, but it never becomes

symmetrical.

5. 󺭬󺭭 The Mean of the F-distribution Depends on Degrees of Freedom

The mean of the F-distribution (when df2 > 2) is calculated using the formula:

Mean









This tells us:

The mean changes with df2.

The mean is always greater than 1, and approaches 1 as df2 increases.

6. 󹳦󹳤󹳧 It Has No Negative Values and the Minimum Value is Zero

Because it is the ratio of two squared quantities, the smallest value the F-statistic can take is

0. But even 0 is almost impossible, because variances are rarely exactly zero.

This is why F-tables start with positive values and dont include negative values.

Easy2Siksha

7. 󹳨󹳤󹳩󹳪󹳫 Used to Test Hypotheses about Variances

The F-distribution is the heart of several important hypothesis tests, such as:

Testing equality of two variances (F-test)

ANOVA (checking if means of multiple groups are different)

Model comparison in regression

In all of these, you calculate an F-ratio and then compare it to the critical value from the F-

distribution.

8. 󹴂󹴃󹴄󹴅󹴉󹴊󹴆󹴋󹴇󹴈 The Total Area Under the Curve Is 1

Like all probability distributions, the area under the F-distribution curve equals 1. This

represents the 100% probability of all possible outcomes.

9. 󷃆󹸊󹸋 As the Degrees of Freedom Increase, It Approaches the Normal Distribution

When df1 and df2 both become large (e.g., over 100), the shape of the F-distribution

becomes more symmetrical and similar to the normal distribution.

This is helpful because it allows statisticians to approximate results with normal curves

when sample sizes are large.

10. 󹸯󹸭󹸮 Highly Sensitive to Outliers and Sample Sizes

Because the F-distribution is based on variance, and variance is affected by outliers, the F-

distribution is also sensitive to extreme values.

Also, small sample sizes can make the F-distribution more skewed and less reliable.

󼨐󼨑󼨒 A Quick Recap: Summary of Key Features

Feature

Description

Shape

Right-skewed, asymmetrical

Range

Only positive values (0 to ∞)

Parameters

Two degrees of freedom (df1, df2)

Easy2Siksha

Feature

Description

Mean

Depends on df2, calculated as df2/(df2 - 2)

Uses

Comparing variances, ANOVA, regression

Tail

Long right tail (especially for small df)

Area under curve

Always 1

Sensitivity

Affected by outliers and small sample sizes

Relation to Normal

Becomes similar to normal as df increases

󹴷󹴺󹴸󹴹󹴻󹴼󹴽󹴾󹴿󹵀󹵁󹵂 Conclusion: Why the F-Distribution Matters to You

The F-distribution may sound technical at first, but it's actually quite practical. It helps

answer real-life questions like:

Are the teaching methods in two classrooms equally effective?

Is one group more variable than another?

Are multiple group averages significantly different?

The more you practice with it in problems involving ANOVA, F-tests, and regression

the more comfortable youll be in using this powerful statistical tool.

Just like characters in a novel, each distribution in statistics has its own personality. The F-

distribution is serious, right-leaning, and loves to compare groups. Once you understand its

nature, youll find it much easier to use in real-world applications.

SECTION-C

5. A total of 10 students were evaluated on the basis of their performance based on a test

with 50 maximum marks. They were given coaching for 6 months and again evaluated.

The results are given below. Can you conclude that the students have benefitted by

the extra coaching?

Student

Code

III

Marks:

Easy2Siksha

Before

coaching

After

coaching

Student

Code

VII

VIII

Marks:

Before

coaching

After

coaching

Ans: Imagine you are a teacher who has just completed a special 6-month coaching program

for a group of 10 students. You had assessed their performance before the coaching began

and again after it was completed. Now you are eager to know: Did the extra coaching really

help the students improve?

This is a very practical question in the field of education and psychology. To answer it in a

proper and logical way, we can use statistics, which gives us tools to measure and verify

improvements.

Let us now take this step-by-step like a story and learn some basic concepts in a very simple

manner.

󹴡󹴵󹴣󹴤 Understanding the Scenario

Here is the data of students scores before and after coaching:

Student

Marks Before Coaching

Marks After Coaching

III

Easy2Siksha

Student

Marks Before Coaching

Marks After Coaching

VII

VIII

Now the goal is to analyze whether this coaching actually helped the students to improve.

󼨐󼨑󼨒 Basic Concepts: What are We Trying to Do?

This type of problem can be solved using a "Paired t-test", which is a statistical method used

when we want to compare two sets of related scores in this case, the "before" and "after"

marks of the same students.

󹸯󹸭󹸮 What is a Paired t-test?

It is used when you have two sets of scores for the same group of individuals.

You want to know if there is a significant difference between the two sets.

It helps to determine whether an intervention (like coaching) had any real impact.

󼩕󼩖󼩗󼩘󼩙󼩚 Step-by-Step Process to Analyze

Lets go through this analysis like a recipe step by step.

Step 1: Find the Differences in Scores

We find the difference between After Coaching and Before Coaching for each student:

Student

Before

After

Difference (d = After - Before)

III

Easy2Siksha

Student

Before

After

Difference (d = After - Before)

VII

VIII

Step 2: Calculate the Mean of the Differences (d



)

To calculate the average improvement (mean difference), we add all the differences and

divide by the number of students:

Sum of differences

= 1 + 10 + 5 + 1 + 22 + 12 + 3 + 21 + 14 + 16

= 105

Number of students = 10

Mean difference (d) = 105 / 10 = 10.5

Step 3: Calculate the Standard Deviation of the Differences

We now calculate how spread out the differences are from the average.

1. Find each (d - d)²

d - d



(d - d



)²

-9.5

90.25

-0.5

0.25

-5.5

30.25

Easy2Siksha

d - d



(d - d



)²

-9.5

90.25

11.5

132.25

1.5

2.25

-7.5

56.25

10.5

110.25

3.5

12.25

5.5

30.25

2. Add all the (d - d



)² values:

Sum = 90.25 + 0.25 + 30.25 + 90.25 + 132.25 + 2.25 + 56.25 + 110.25 + 12.25 + 30.25

= 554.5

3. Divide by n - 1 (degrees of freedom = 10 - 1 = 9):

Variance = 554.5 / 9 = 61.61

Standard Deviation (Sd) = √61.61 ≈ 7.85

Step 4: Calculate the t-value

Now use the formula:

















Where:

d= 10.5 (mean difference)

= 7.85 (standard deviation of differences)

N = 10 (number of students)



 











  

   





  

  



Step 5: Compare the t-value with Critical Value

Easy2Siksha

From the t-distribution table for 9 degrees of freedom (n-1) and significance level of 0.05,

the critical value is approximately 2.262.

Our calculated t = 4.23 > 2.262

󷗭󷗨󷗩󷗪󷗫󷗬 Conclusion

Since the calculated t-value (4.23) is greater than the critical t-value (2.262), we reject the

null hypothesis.

󷃆󼽢 What does it mean?

It means the difference in marks before and after coaching is statistically significant. In

simple words:

Yes! The extra coaching has helped the students improve their performance.

󼨻󼨼 Why is This Important?

This kind of statistical analysis is very useful in many real-life scenarios:

Schools use it to test the effectiveness of teaching methods.

Psychologists use it to evaluate behavioral changes after therapy.

Companies use it to measure training results of employees.

It shows us how we can use numbers to draw meaningful conclusions, and not just rely on

guesses or opinions.

󹰤󹰥󹰦󹰧󹰨 Final Thoughts for Students

Even though statistics might sound difficult, its actually a logical way of

understanding the truth behind data.

Always start by understanding the question what are we trying to find?

Follow the steps like a recipe: difference → mean → standard deviation → t-value →

conclusion.

With practice, youll realize how helpful statistics is in making fair, logical, and

evidence-based decisions.

6. What is the difference between Parametric and Non-parametric Test? The Chi square

test falls under which category? Explain the mechanics of this test.

Ans: 󹴷󹴺󹴸󹴹󹴻󹴼󹴽󹴾󹴿󹵀󹵁󹵂 Understanding Parametric and Non-Parametric Tests: A Story for University

Students

Easy2Siksha

Imagine you're a university student named Riya. Youre studying statistics and your

professor tells you that todays class will cover a very important concept Parametric and

Non-Parametric Tests and also something called the Chi-Square Test.

You sigh, thinking this is going to be difficult. But then your professor starts explaining it in a

way that feels like a story, and everything becomes much clearer.

󼨽󼨾󼨿󼩁󼩀 Parametric vs Non-Parametric Tests Two Different Paths

Your professor begins:

"Imagine you are baking a cake, but before baking, you need to know whether your

ingredients are perfect. The same goes for statistical testing before testing your data, you

need to know what kind of data you have and what assumptions it follows."

This is where Parametric and Non-Parametric tests come in.

󷃆󽄾 What are Parametric Tests?

Riyas professor explains:

"Parametric tests are like baking with a strict recipe. You need to follow certain rules, or the

cake wont come out right. These tests assume that your data fits a specific shape usually

a normal distribution (the famous bell curve)."

󹳴󹳵󹳶󹳷 Key Assumptions of Parametric Tests:

1. The data is normally distributed.

2. The sample size is large enough.

3. The data is measured on an interval or ratio scale (like height, weight, temperature).

4. Variances in the data groups should be equal (called homogeneity of variance).

󼩕󼩖󼩗󼩘󼩙󼩚 Examples of Parametric Tests:

t-test (for comparing two means)

ANOVA (for comparing more than two means)

Pearson's correlation (for relationships between variables)

These tests are powerful, meaning they give precise results if all assumptions are met.

󷃆󽄾 What are Non-Parametric Tests?

Then the professor moves on:

Easy2Siksha

"Now imagine baking without a fixed recipe. You mix what you have and hope it works out.

Thats what non-parametric tests do they dont assume any specific shape or distribution

of data."

These tests are more flexible and can be used in situations where parametric tests cannot.

󹳴󹳵󹳶󹳷 Characteristics of Non-Parametric Tests:

1. No assumption of normal distribution.

2. Work well with small sample sizes.

3. Can be used for ordinal data (like rankings) or nominal data (like categories).

4. Less powerful than parametric tests but more robust when data doesnt follow rules.

󼩕󼩖󼩗󼩘󼩙󼩚 Examples of Non-Parametric Tests:

Chi-square test

Mann-Whitney U test

Kruskal-Wallis test

Wilcoxon signed-rank test

󽅄󽅅 So, Which Category Does the Chi-Square Test Fall Under?

Your professor pauses and asks the class:

So, students, where do you think the Chi-Square Test falls?

Riya proudly raises her hand and says, Non-Parametric Test!

Correct! 󷗭󷗨󷗩󷗪󷗫󷗬

The Chi-Square Test is a non-parametric test, used mainly with categorical data like colors,

brands, gender, education levels, etc.

󹸯󹸭󹸮 The Mechanics of the Chi-Square Test: Breaking it Down

Now the professor dives deep into the Chi-Square Test.

He explains:

Lets say you run a small café. You serve coffee, tea, and juice. You want to know: Do

people prefer one drink more than the others? Thats where you use the Chi-Square Test for

Goodness of Fit.

There are two main types of Chi-Square Tests:

Easy2Siksha

1. 󷃆󼽢 Chi-Square Test for Goodness of Fit

This checks whether your observed data (what actually happened) matches expected data

(what you thought would happen).

󼨽󼨾󼨿󼩁󼩀 Example:

You expect that out of 100 customers:

40 will choose coffee

30 will choose tea

30 will choose juice

But your actual results are:

50 chose coffee

20 chose tea

30 chose juice

To know whether this difference is just by chance or statistically significant, you apply the

Chi-Square Test.

󹴌󹴍󹴐󹴑󹴒󹴎󹴏󹴓󹴔󹴕 Formula:







󰇛

  

󰇜





Where:

O = Observed value

E = Expected value

You calculate this for each category, sum them up, and get a Chi-Square value. Then, you

compare it to a critical value from the Chi-Square distribution table based on your degrees

of freedom.

If your calculated value is greater than the table value, the difference is statistically

significant.

2. 󷃆󼽢 Chi-Square Test for Independence

This test helps answer questions like:

Is there a relationship between gender and preferred drink?

Here, you're testing independence between two variables.

󼨽󼨾󼨿󼩁󼩀 Example:

Easy2Siksha

Coffee

Tea

Juice

Total

Male

Female

Total

100

Now you want to know if gender affects drink choice.

Again, use the formula:







󰇛

  

󰇜





Where the expected frequency (E) is calculated using:



󰇛



󰇜

󰇛󰇜



You calculate E for each cell, find the difference from O, and compute the Chi-Square value.

󹸽 Interpreting Results

After calculating the value, compare it with the critical value from the Chi-Square table. If:

Chi-Square calculated > Chi-Square table → Reject null hypothesis

Chi-Square calculated ≤ Chi-Square table → Fail to reject null hypothesis

󹳴󹳵󹳶󹳷 Null Hypothesis (H₀):

For goodness of fit: No difference between observed and expected.

For independence: The variables are independent (no relationship).

󼿍󼿎󼿑󼿒󼿏󼿓󼿐󼿔 Advantages of Chi-Square Test

1. 󷃆󼽢 Easy to understand and apply.

2. 󷃆󼽢 Works on categorical data.

3. 󷃆󼽢 No need to assume a normal distribution.

4. 󷃆󼽢 Used widely in market research, education, sociology, and psychology.

󺠰󺠱 Limitations

Easy2Siksha

1. 󽅂 Cannot be used for small sample sizes (expected frequency in any cell should not

be < 5).

2. 󽅂 Only tells association, not cause-effect.

3. 󽅂 Not suitable for continuous data.

󼨐󼨑󼨒 Final Summary

Riya now understands:

Feature

Parametric Test

Non-Parametric Test

Assumptions

Requires normal distribution

No strict assumptions

Data Type

Interval/Ratio

Nominal/Ordinal

Power

More powerful

Less powerful but more robust

Examples

t-test, ANOVA

Chi-Square, Mann-Whitney U, etc.

Chi-Square Test Type

󽅂 Not parametric

󷃆󼽢 Non-parametric

󽄻󽄼󽄽 Conclusion

So now, when someone asks Riya:

What is the difference between parametric and non-parametric tests? And what is the Chi-

Square test?

She smiles and replies with confidence, like you will too:

Parametric tests are for data that follows rules, like a recipe. Non-parametric tests dont

need such assumptions. And the Chi-Square test is a non-parametric test used to analyze

categorical data. It helps us know whether our observed results are significantly different

from what we expected, or whether two variables are related.

Easy2Siksha

SECTION-D

7. Highlight the difference between One way and Two way ANOVA technique. Describe

the method of working out the critical difference for both.

Ans: Imagine you are a university student working on your research project. Youve

collected some data, maybe comparing students' scores from different classes or analyzing

the impact of diet and exercise on weight loss. Now, you want to test whether the

differences you observe in the data are real (statistically significant) or just due to chance.

This is where ANOVA comes in a powerful statistical tool.

Lets break it down like a story to make it easy and engaging.

What is ANOVA?

ANOVA stands for Analysis of Variance. It is a statistical technique used to compare the

means of different groups and check if at least one group mean is significantly different

from the others.

But why not just use a t-test?

Well, t-tests are great for comparing two groups, but what if you have more than two? For

example, comparing scores of students from 3 different classes or analyzing the

performance of people under 4 different diets? Thats where ANOVA shines it allows you

to compare three or more groups simultaneously.

There are two main types of ANOVA:

One-Way ANOVA

Two-Way ANOVA

Lets understand them one by one.

One-Way ANOVA: The Simpler Road

What is it?

Imagine you are testing whether students from three different teaching methods (Method

A, Method B, and Method C) score differently in an exam. The only thing you're comparing

is one factor: the teaching method. This is a one-way ANOVA, because you are analyzing the

impact of one independent variable (teaching method) on the dependent variable (exam

scores).

When to Use One-Way ANOVA?

Use it when:

Easy2Siksha

You have one independent variable (with two or more groups).

You want to compare the means of those groups.

Your dependent variable is quantitative (e.g., test scores, height, weight).

Example:

Lets say:

Group A: Students taught with Method A

Group B: Students taught with Method B

Group C: Students taught with Method C

You collect test scores and want to know: Is there a significant difference in mean scores

across these three groups?

One-way ANOVA will help you answer this.

Two-Way ANOVA: A Step Ahead

What is it?

Now, imagine you're not just interested in teaching methods, but also in the gender of

students. You want to see:

How teaching method affects scores,

How gender affects scores,

AND whether there's an interaction between the two.

This is where Two-Way ANOVA is used it allows you to examine the effects of two

independent variables and their interaction on a dependent variable.

When to Use Two-Way ANOVA?

Use it when:

You have two independent variables (e.g., teaching method and gender).

You want to compare the individual effects and interaction effect of both.

Your dependent variable is quantitative.

Example:

Teaching Method

Gender

Test Score

Male

Easy2Siksha

Teaching Method

Gender

Test Score

Female

Male

Female

Male

Female

A two-way ANOVA will tell you:

Does teaching method affect scores?

Does gender affect scores?

Do teaching method and gender interact in a way that affects scores?

󹸯󹸭󹸮 Key Differences Between One-Way and Two-Way ANOVA

Feature

One-Way ANOVA

Two-Way ANOVA

Number of

Independent

Variables

One

Two

Example Variables

Teaching method

Teaching method and gender

Interaction Effect

Not studied

Studied

Purpose

To test the effect of one

factor

To test the effects of two factors and

their interaction

Complexity

Simpler

More complex

Data Structure

One factor with multiple

levels (e.g., Method A, B, C)

Two factors with multiple levels (e.g.,

Method A/B/C and Male/Female)

󷗭󷗨󷗩󷗪󷗫󷗬 What is Critical Difference (CD)?

So far, ANOVA only tells us whether there is a significant difference. But it doesn't say

between which groups the difference lies. Thats where Critical Difference (CD) or Post-Hoc

Tests come in.

Think of it like this:

Easy2Siksha

ANOVA is like a teacher telling you that someone in class cheated but not who. CD helps

you identify who exactly is different from the rest.

Step-by-Step: How to Work Out the Critical Difference (CD)

Once ANOVA shows that there's a significant difference, we compute CD using tests like

Tukeys HSD (Honestly Significant Difference) or Least Significant Difference (LSD). Heres

how we usually proceed:

󷃆󹹳󹹴󹹵󹹶 Steps for Calculating Critical Difference (LSD method)

1. Compute Mean Square Error (MSE)

From the ANOVA table, locate the error mean square (i.e., the residual variance).

2. Calculate the Standard Error (SE):









Where n = number of observations in each group.

3. Find the t-value from the t-table at the required significance level (e.g., 0.05), using

the degrees of freedom from the error term.

4. Calculate Critical Difference (CD):

CD = t x SE

5. Compare Mean Differences:

Subtract the group means from each other. If the difference is greater than CD, then

that pair of groups is significantly different.

󷃆󼽢 Example: One-Way ANOVA CD Calculation

Suppose:

MSE = 10

n = 10 per group

t (from t-table) = 2.101 (for df = 27, at 0.05 level)





  







CD = 2.101 x 1.41 = 2.96

Now, compare group mean differences. If Group A mean = 50 and Group B mean = 54, then:

Easy2Siksha

54−50=4>2.96Significant Difference

󼨐󼨑󼨒 For Two-Way ANOVA

Two-way ANOVA CD is a little more complex. You may need to compute:

CD for the main effects (e.g., teaching method)

CD for the second factor (e.g., gender)

CD for the interaction effect

The same basic steps apply:

Extract MSE from the error term.

Calculate SE for each factor (this might vary if sample sizes are unequal).

Use the corresponding t-value.

Compute CD and compare group differences.

󹳴󹳵󹳶󹳷 Summary and Key Takeaways

Concept

One-Way ANOVA

Two-Way ANOVA

Independent

Variables

Measures

One factors effect

Two factors and their interaction

Output

Whether at least one group

mean is different

Whether each factor and their

combination affect the outcome

Use of CD

To find which groups differ

To find differences in each factor and

interaction

Post-Hoc Tests

LSD, Tukeys HSD

Similar, but done separately for each

main effect and interaction

󷕘󷕙󷕚 Final Thoughts

Understanding One-Way and Two-Way ANOVA is essential for university-level research and

data analysis. Think of One-Way ANOVA as asking Which group is different? and Two-Way

ANOVA as asking Do two factors matter, and do they work together?

Easy2Siksha

And always remember: ANOVA shows that a difference exists, but Critical Difference (CD)

shows where it exists.

By mastering these tools, you'll not only enhance your academic research but also gain a

deeper understanding of the real-world complexities of data and decision-making.

8. In order to assess the significance of possible variation in productivity performance of

three different varieties (X, Y, Z) of wheat, an experiment was conducted on 12 equi-sized

plots. The related data is given below. Frame the hypothesis and test if there is any

significant difference in the productivity of these three varieties.

Ans: 󷊀󷊁󷊂󷊃 The Story Begins: A Wheat Field Experiment

Once upon a time in a university research centre, a group of agricultural scientists wanted to

improve wheat production in the country. They had developed three new varieties of wheat

named X, Y, and Z. But before recommending them to farmers, they had a big question:

Is there any significant difference in the productivity of these three varieties?

To answer this question, they conducted an experiment on 12 plots of land. Each plot was of

equal size to ensure fairness. They planted:

Variety X in 4 plots

Variety Y in 4 plots

Variety Z in 4 plots

After the harvest, the data collected (in terms of productivity, say in quintals per plot) was:

Variety

Productivity

14, 16, 18,

14, 13, 15, 22

18, 16, 19, 20

Easy2Siksha

Oops! One value for Variety X is missing but for the sake of simplicity and story, lets

assume that missing value is 15 (you can replace this with any imputed value in actual

calculations). So now, we complete the data:

Variety

Productivity

14, 16, 18, 15

14, 13, 15, 22

18, 16, 19, 20

Now, the scientists needed to test whether the observed differences in productivity are real

(statistically significant) or just due to chance. This is where ANOVA (Analysis of Variance)

comes in.

󹸯󹸭󹸮 Step 1: Setting the Hypotheses

Before we jump into the math, lets understand the logic behind hypothesis testing.

Null Hypothesis (H₀): There is no significant difference in the mean productivity of

the three wheat varieties.

Alternative Hypothesis (H₁): At least one variety is significantly different in mean

productivity.

Our goal is to test if the null hypothesis can be rejected using statistical evidence.

󹳨󹳤󹳩󹳪󹳫 Step 2: What is ANOVA?

ANOVA (Analysis of Variance) is a method used when we want to compare more than two

groups (in our case: X, Y, Z) to see if their means are significantly different.

Rather than comparing means pair-by-pair (which would be tedious), ANOVA helps us check

all groups together at once.

It works by comparing two types of variations:

1. Between-group variation How much the group means differ from the overall

mean.

2. Within-group variation How much the values within each group vary among

themselves.

If between-group variation is much greater than within-group variation, we conclude that

not all varieties perform the same.

Easy2Siksha

󹴌󹴍󹴐󹴑󹴒󹴎󹴏󹴓󹴔󹴕 Step 3: Calculating ANOVA Step-by-Step

Lets compute it step-by-step using our example.

󹻂 Step 3.1: Total observations (N)

There are 4 values in each group → Total = 4 + 4 + 4 = 12 observations

󹻂 Step 3.2: Group Means

Mean of X = (14 + 16 + 18 + 15)/4 = 15.75

Mean of Y = (14 + 13 + 15 + 22)/4 = 16

Mean of Z = (18 + 16 + 19 + 20)/4 = 18.25

󹻂 Step 3.3: Grand Mean (Overall Average)

Total sum = (14 + 16 + 18 + 15 + 14 + 13 + 15 + 22 + 18 + 16 + 19 + 20) = 220

Grand Mean = 220 / 12 = 18.33

Wait, lets recheck the sum of all:

X: 14 + 16 + 18 + 15 = 63

Y: 14 + 13 + 15 + 22 = 64

Z: 18 + 16 + 19 + 20 = 73

Total = 63 + 64 + 73 = 200

So Grand Mean = 200 / 12 = 16.67

󹻂 Step 3.4: Between Group Sum of Squares (SSB)

SSB = n * [(mean of X - GM)² + (mean of Y - GM)² + (mean of Z - GM)²]

Where n = number of observations per group = 4

SSB = 4 * [(15.75 - 16.67)² + (16 - 16.67)² + (18.25 - 16.67)²]

= 4 * [0.8464 + 0.4489 + 2.5609]

= 4 * 3.8562 = 15.4248

󹻂 Step 3.5: Within Group Sum of Squares (SSW)

SSW = Sum of squared differences within each group.

For X:

(14 - 15.75)² + (16 - 15.75)² + (18 - 15.75)² + (15 - 15.75)²

= 3.0625 + 0.0625 + 5.0625 + 0.5625 = 8.75

For Y:

(14 - 16)² + (13 - 16)² + (15 - 16)² + (22 - 16)²

= 4 + 9 + 1 + 36 = 50

Easy2Siksha

For Z:

(18 - 18.25)² + (16 - 18.25)² + (19 - 18.25)² + (20 - 18.25)²

= 0.0625 + 5.0625 + 0.5625 + 3.0625 = 8.75

Total SSW = 8.75 + 50 + 8.75 = 67.5

󹳨󹳤󹳩󹳪󹳫 Step 4: Degrees of Freedom

Between groups (df1) = k - 1 = 3 - 1 = 2

Within groups (df2) = N - k = 12 - 3 = 9

Where k = number of groups, N = total observations

󹸱󹸲󹸰 Step 5: Mean Squares

MSB (Mean Square Between) = SSB / df1 = 15.42 / 2 = 7.71

MSW (Mean Square Within) = SSW / df2 = 67.5 / 9 = 7.5

󼨐󼨑󼨒 Step 6: F-Ratio

F = MSB / MSW = 7.71 / 7.5 = 1.028

󹴂󹴃󹴄󹴅󹴉󹴊󹴆󹴋󹴇󹴈 Step 7: Compare with F-table Value

At 5% level of significance, for df1 = 2 and df2 = 9, the critical F-value is approximately 4.26

(from F-distribution table).

Since our calculated F (1.028) < critical F (4.26), we fail to reject the null hypothesis.

󷃆󼽢 Final Conclusion:

There is no statistically significant difference in the productivity of the three varieties of

wheat at the 5% significance level. Any small differences observed are likely due to chance.

󹰤󹰥󹰦󹰧󹰨 Real-World Interpretation:

What this means in simple terms is:

Although variety Z looks a little better than the rest based on raw numbers, this difference is

not strong enough to conclude it performs better consistently. So, from a statistical point of

view, we can treat all three varieties as equally productive.

Easy2Siksha

󹴷󹴺󹴸󹴹󹴻󹴼󹴽󹴾󹴿󹵀󹵁󹵂 Key Learnings for University Students:

1. Hypothesis Testing helps in decision-making based on data.

2. ANOVA is useful when comparing more than two group means.

3. The F-ratio tells us whether between-group variation is significantly higher than

within-group variation.

4. Even if one group seems better visually, it might not be statistically significant.

5. Always check your conclusions against the critical value from the F-table.

This paper has been carefully prepared for educational purposes. If you notice any mistakes or

have suggestions, feel free to share your feedback.