1.

Processing Operations
Explanation

These are the steps followed after data collection to make it ready for analysis.

a) Editing

Definition: Checking collected data for errors, completeness, and consistency.

Example Problem:
A survey asks for respondents' age. One participant writes 150 years. What should researchers
do?

Answer:
The value is impossible. During editing, the researcher should either:
✔ Recheck with respondent
✔ Remove or correct the value based on logic or average age.

b) Coding

Definition: Assigning numerical or symbolic codes to data for easy analysis.

Example Problem:
Code the following responses for marital status: Single, Married, Divorced, Widowed.

Answer (Codebook):

Marital Status Code

Single 1
Married 2
Divorced 3
Widowed 4

c) Classification

Definition: Grouping data based on characteristics (qualitative or quantitative).

Example Problem:
Classify monthly income: ₹5000, ₹8000, ₹12000, ₹17000, ₹25000 into income groups.

Answer:

Income Range No. of People

Below ₹10,000 ₹5000, ₹8000 = 2
₹10,000–₹20,000 ₹12000, ₹17000 = 2
Above ₹20,000 ₹25000 = 1

d) Tabulation

Definition: Systematic presentation of data into rows and columns.

Example Problem:
Tabulate data for number of mobile phones used by 10 people:
1, 2, 2, 1, 3, 2, 1, 1, 2, 3

Answer:

No. of Mobiles Frequency

1 4
2 4
3 2

2. Statistics in Research
A. Measures of Central Tendency

1) Mean (Average)

Formula:

Problem:
Marks of 5 students: 56, 60, 70, 80, 84.
Calculate Mean.
Answer:
Mean = (56+60+70+80+84)/5 = 350/5 = 70

2) Median

Definition: Middle value when data is arranged.

Problem:
Find Median: 20, 25, 30, 28, 18, 22, 40

Answer:
Arranged: 18, 20, 22, 25, 28, 30, 40
Median = Middle value = 25

3) Mode

Definition: Value that occurs most frequently.

Problem:
Find Mode: 10, 12, 14, 12, 15, 18, 12, 16

Answer:
12 appears most → Mode = 12

B) Measures of Dispersion

1) Range

Problem:
Find range of weight: 55, 58, 60, 65, 70, 75

Answer:
Range = 75–55 = 20

2) Standard Deviation

Formula:

Problem:
Find SD for: 2, 4, 6, 8, 10
Answer (in brief steps):
Mean = 6
Squared deviations: 16, 4, 0, 4, 16
SD = √(40/5)= √8 = 2.82

C. Measures of Asymmetry (Skewness)

Distribution Shape Example

Positive Skew Tail on right Income

Negative Skew Tail on left Retirement age

Symmetrical Even Height

Problem:
Salary distribution: ₹10k, ₹12k, ₹15k, ₹18k, ₹50k → Skewed?

Answer:
One very high value (50k) → Positively skewed

D. Measures of Relationship

1) Correlation

Definition: Measures direction and strength of relationship.

Problem:
Study hours vs Score: Calculate correlation (interpretation only)

Hours Score

2 40

4 50

6 60

8 80

Answer:
As study hours increase, marks increase → Positive correlation

2) Simple Regression

Formula:
 Y = a + bX

(Y = dependent, X = independent)

Problem:
Sales = a + b × Advertisement
Ads = ₹10,000 → Sales = ₹70,000
Ads = ₹15,000 → Sales = ₹85,000
Find relationship.

Answer:
When ads increase, sales increase → Positive linear regression

Symmetric vs Asymmetric Dispersion of Data with an example

Dispersion refers to how spread out the data values are around the central value
(mean/median).

Skewness helps us identify whether the distribution is symmetric or asymmetric.

1. Symmetric Dispersion

✔ Definition

A distribution is symmetric when:

 The left and right sides of the distribution are mirror images.
 Mean = Median = Mode
 No skewness → Skewness = 0

✔ Characteristics

 Data is evenly spread around the center.

 No tail is longer than the other.

✔ Example

Data: {10, 12, 14, 16, 18}

 Mean = 14
 Median = 14
 The data values above and below 14 are equally spaced.
✔ Interpretation

This is a symmetrical distribution → no skew.

2. Asymmetric Dispersion (Skewed Distribution)

A distribution is asymmetric when:

 The two sides of the curve are not mirror images

 Mean ≠ Median ≠ Mode
 The distribution has skewness (positive or negative)

There are two types:

A. Positively Skewed (Right Skewed)

✔ Definition

The right tail (larger values) is longer.

✔ Order of averages

Mean > Median > Mode

✔ Example

Data: {5, 6, 7, 8, 50}

 Mean ≈ 15.2
 Median = 7
 Mode = none

One large value (50) pulls the mean towards the right → positive skew.

✔ Interpretation

More values are on the lower side, with a few very high values.
B. Negatively Skewed (Left Skewed)

✔ Definition

The left tail (smaller values) is longer.

✔ Order of averages

Mean < Median < Mode

✔ Example

Data: {2, 20, 22, 25, 27}

 Mean ≈ 19.2
 Median = 22
 Mode = none

One small value (2) pulls the mean towards the left → negative skew.

✔ Interpretation

More values are on the higher side, with a few very low values.

TESTING OF HYPOTHESIS
Definition

A hypothesis is a tentative assumption to be tested.

Example Hypothesis:
“There is no significant difference in performance of boys and girls.”

Basic Concepts

Term Meaning

Null Hypothesis (H₀) No difference/effect

Alternative Hypothesis (H₁) There is a difference

 Term Meaning

Level of Significance (α) Probability of rejecting true H₀ (usually 5%)

Type I Error Reject true H₀

Type II Error Accept false H₀

P-value Probability of observed result

Procedure of Hypothesis Testing (Steps)

1⃣ Set hypothesis (H₀ and H₁)

2⃣ Choose significance level (α)
3️⃣ Select a suitable statistical test (t-test, Z-test, chi-square)
4⃣ Calculate test statistic
5⃣ Compare with critical value or p-value
6️⃣ Accept or reject H₀

Flow Diagram

Start
↓
Define Hypothesis (H₀, H₁)
↓
Select significance level (α)
↓
Choose appropriate test
↓
Compute test statistic
↓
Compare with critical/p-value
↓
Accept or reject H₀
 ↓
Conclusion

Explanation of flow diagram:

1. Start
This is the beginning of the hypothesis testing procedure. You start with a research question—for
example, “Does a new teaching method improve student performance?”

2. Define Hypothesis (H₀, H₁)

You set up two statements:

 Null Hypothesis (H₀):

The claim that there is no effect or no difference.
Example: The new teaching method has no effect.
 Alternative Hypothesis (H₁):
The claim you want to test or prove.
Example: The new teaching method improves performance.

These two hypotheses must be clear and mutually exclusive.

3. Select Significance Level (α)

The significance level is the probability of making a Type I error (rejecting a true H₀).
Common choices are:

 α = 0.05 (5% risk)

 α = 0.01 (1% risk)

This α value decides the critical region for the test.

4. Choose Appropriate Test

You choose a statistical test depending on:
  Type of data (quantitative/qualitative)
 Sample size
 Whether population variance is known
 Number of groups

Examples:

 Z-test → large samples, known variance

 t-test → small samples, unknown variance
 Chi-square test → categorical data
 ANOVA → comparing more than 2 means

5. Compute Test Statistic

Using the sample data, calculate the test statistic value (like z, t, χ², F).
This value tells you how far your sample is from the expected value under H₀.

6. Compare with Critical Value / p-value

You make a decision using either:

Critical Value Method

 If test statistic > critical value, reject H₀.

 If test statistic ≤ critical value, fail to reject H₀.

p-value Method

 If p ≤ α, reject H₀.
 If p > α, fail to reject H₀.

This step checks whether the sample evidence is strong enough to disprove H₀.

7. Accept or Reject H₀
Based on the comparison:

 Reject H₀ → Evidence supports H₁

  Fail to Reject H₀ → Not enough evidence to support H₁ (but H₀ is not proved)

Remember: We never “accept” H₁ fully; we only see if evidence suggests rejecting H₀.

8. Conclusion
Finally, you interpret the statistical decision in simple, real-world terms.

Example:
“At 5% significance level, the new teaching method significantly improves student
performance.”

 Comparison of Common Tests of Hypothesis

Test
Purpose Example
Name
t-test Compare means of two groups Male vs female marks
Mean comparison (large
Z-test Compare salary levels
sample)
Relationship between gender and product
Chi-square Test for frequency/association
preference
ANOVA Compare means of 3+ groups Compare 3 teaching methods
1) t-Test
A t-test is a statistical test used to decide whether the mean of a sample is significantly
different from:

 a known population mean, or

 the mean of another sample.

It is used when:

 sample size is small (n < 30), and

 population standard deviation is unknown.

The test uses the t-distribution instead of the normal distribution.

Types of t-tests

1. One-sample t-test
Checks if the sample mean differs from a known population mean.
 2. Independent two-sample t-test
Compares means of two independent groups (e.g., boys vs girls marks).
3. Paired t-test
Compares before–after measurements on the same subjects.

Key Formula

For a one-sample t-test:

A t-test helps determine whether observed differences in means are due to chance or statistically
significant.

2) Z-Test

A Z-test is a statistical test used to check whether the sample mean or proportion is
significantly different from the population mean/proportion.
It is based on the standard normal (Z) distribution.

It is used when:

 Sample size is large (n ≥ 3️0), OR

 Population standard deviation (σ) is known

When Do We Use a Z-Test?

1. Test for population mean (mean difference)

Example: Checking if the average height of students differs from 160 cm.
2. Test for population proportion
Example: Checking if the proportion of left-handed people is 10%.
3. Two-sample Z-test
Compare means or proportions of two large independent samples.

Formula for One-Sample Z-Test

Why Use Z-Test?

 Helps determine if differences between sample and population are real or due to
sampling error.
 Useful for large samples where the distribution approaches normal.

3) Chi-Square (χ²) Test

The Chi-Square test is a statistical test used to determine whether there is a significant
difference between:

 Observed values (what you actually see in the data)

 Expected values (what you expect based on theory or past data)

It is commonly used with categorical data (e.g., gender, yes/no, types of products, colors, etc.).

Two Main Types of Chi-Square Tests

1. Chi-Square Test of Independence

Used to check whether two categorical variables are related.

Example:
Is gender related to preference for online vs offline shopping?

2. Chi-Square Goodness-of-Fit Test

Used to check whether observed frequencies follow a theoretical distribution.

Example:
Do students prefer four different sports equally?

When is Chi-Square Used?

 Data is categorical
 Frequencies (counts) are used, not averages
 Sample size is moderately large
 Expected values should not be too small

Problems on chi square:

1) Online vs Offline shopping (2×2 table)

Question: A sample of 120 shoppers was asked whether they prefer online or offline shopping.
The results:

Gender Online Offline Total

Male 40 20 60

Female 30 30 60

Total 70 50 120

Test at 5% significance whether shopping preference is independent of gender.

Solution

Step 1 — Hypotheses
H₀: Preference and gender are independent.
H₁: Preference and gender are not independent.

Step 2 — Expected frequencies (E = row total × column total / grand total)

  Male–Online: E = 60 × 70 / 120 = 35
 Male–Offline: E = 60 × 50 / 120 = 25
 Female–Online: E = 60 × 70 / 120 = 35
 Female–Offline: E = 60 × 50 / 120 = 25

Step 3 — χ² calculation

Compute each cell:

 (40−3️5)² / 3️5 = 25 / 3️5 = 0.7143️

 (20−25)² / 25 = 25 / 25 = 1.0000
 (3️0−3️5)² / 3️5 = 25 / 3️5 = 0.7143️
 (3️0−25)² / 25 = 25 / 25 = 1.0000


Step 4 -- Degrees of freedom

df = (rows − 1) × (cols − 1) = 1 × 1 = 1

Step 5 — Critical value (α = 0.05)

χ²crit (df=1, 0.05) ≈ 3.841 (Refer the table below)

Decision: χ²calc = 3️.429 < 3️.841 → Fail to reject H₀.

Note: Compare the values to take decision in chi square :

 If your calculated 𝜒2 value is greater than the critical value, you reject the null
hypothesis. This indicates there is a statistically significant difference or relationship.
 If your calculated 𝜒2 value is less than or equal to the critical value, you fail to reject
the null hypothesis.
 Some common critical values at 5% significance level i.e. α = 0.05
Degrees of Freedom (df) Critical Value (at α = 0.05)
1 3.841
2 5.991
3 7.815
4 9.488
5 11.070
10 18.307
15 24.996

2) Survey of 150 people on news source preference by age group produced:

Age Group TV Online Newspaper Row total

Young 20 30 10 60

Middle 25 20 15 60

Old 10 5 15 30

Total 55 55 40 150

Test at 5% significance whether news source preference is associated with age group.

Solution:

Step 1 — Hypotheses
H₀: Age and news source are independent.
H₁: They are not independent.

Step 2 — Expected frequencies (E = row total × column total / grand total)

Compute a few examples (all computed similarly):

 Young–TV: E = 60 × 55 /150 = 22.0

 Young–Online: E = 60 × 55 /150 = 22.0
 Young–Newspaper: E = 60 × 40 /150 = 16.0
 Middle–TV: E = 60 × 55 /150 = 22.0
 Middle–Online: E = 60 × 55 /150 = 22.0
 Middle–Newspaper: E = 60 × 40 /150 = 16.0
 Old–TV: E = 30 × 55 /150 = 11.0
 Old–Online: E = 30 × 55 /150 = 11.0
 Old–Newspaper: E = 30 × 40 /150 = 8.0

Step 3 — χ² calculation (O = observed, E = expected)

Calculate (O−E)²/E for each of the 9 cells and sum:

Cells (O, E, (O−E)²/E):

  Young–TV: (20, 22): (−2)²/22 = 4/22 = 0.1818
 Young–Online: (30, 22): 8²/22 = 64/22 = 2.9091
 Young–Newspaper: (10,16️): (−6️)²/16️ = 3️6️/16️ = 2.25
 Middle–TV: (25,22): 3²/22 = 9/22 = 0.4091
 Middle–Online: (20,22): (−2)²/22 = 4/22 = 0.1818
 Middle–Newspaper: (15,16️): (−1)²/16️ = 1/16️ = 0.06️25
 Old–TV: (10,11): (−1)²/11 = 1/11 = 0.0909
 Old–Online: (5,11): (−6️)²/11 = 3️6️/11 = 3️.2727
 Old–Newspaper: (15,8): 7²/8 = 49/8 = 6.1250

Sum χ² ≈ 0.1818+2.9091+2.25+0.4091+0.1818+0.06️25+0.0909+3️.2727+6️.1250
χ²calc ≈ 15.4830

Step 4 — Degrees of freedom

df = (3️ −1) × (3️ −1) = 2 × 2 = 4

Step 5 — Critical value (α = 0.05)

χ²crit (df=4, 0.05) ≈ 9.488 (Refer the critical value table above )

Decision: χ²calc = 15.483️ > 9.488 → Reject H₀.

3) A spinner with 4 equal sectors (expected equally likely) was spun 80 times. Observed
counts:

Sector A B C D Total

Frequency 18 22 20 20 80

Test at 5% significance whether the spinner is fair (i.e., all sectors equally likely).

Solution

Step 1 — Hypotheses
H₀: Spinner is fair (each sector probability = 1/4).
H₁: Spinner is not fair.

Step 2 — Expected frequencies

E for each sector = 80 / 4 = 20

Step 3 — χ² calculation

Cells:

 A: (18−20)²/20 = 4/20 = 0.20

 B: (22−20)²/20 = 4/20 = 0.20
  C: (20−20)²/20 = 0
 D: (20−20)²/20 = 0

χ² = 0.20 + 0.20 + 0 + 0 = 0.40

Step 4 — Degrees of freedom

df = k − 1 = 4 − 1 = 3

Step 5 — Critical value (α = 0.05)

χ²crit (df=3️, 0.05) ≈ 7.815 (Refer the critical value table above )

Decision: χ²calc = 0.40 < 7.815 → Fail to reject H₀.

4) A bag of 100 candies is supposed to have colours (Red : Green : Blue) in ratio 2:3:5. A
sample of 100 candies yielded counts:

Colour Red Green Blue Total

Observed (O) 10 35 55 100

Test at 5% significance whether the observed counts fit the expected ratio 2:3:5.

Solution

Step 1 — Hypotheses
H₀: Observed distribution follows the ratio 2:3️:5.
H₁: It does not follow that ratio.

Step 2 — Expected frequencies

Total parts = 2 + 3 + 5 = 10. Expected counts:

 Red: E = 100 × (2/10) = 20

 Green: E = 100 × (3/10) = 30
 Blue: E = 100 × (5/10) = 50

Step 3 — χ² calculation

 Red: (10 − 20)² / 20 = 100 / 20 = 5.0000

 Green: (3️5 − 3️0)² / 3️0 = 25 / 3️0 = 0.83️3️3️
 Blue: (55 − 50)² / 50 = 25 / 50 = 0.5000

χ² = 5.0000 + 0.83️3️3️ + 0.5000 = 6.3333

Step 4 — Degrees of freedom

df = k − 1 = 3️ − 1 = 2
Step 5 — Critical value (α = 0.05)
χ²crit (df=2, 0.05) ≈ 5.991 (Refer the critical value table above )

Decision: χ²calc = 6️.3️3️3️ > 5.991 → Reject H₀.