Gemini
Tribhuvan University — MPA 509: Statistics for Public Administration
Exam-Oriented Smart Notes & Revision Guide
Exam Strategy & Question Pattern (TU Format)
TU MPA examinations typically assess both theoretical concepts and numerical
problem-solving:
Group A (Long Questions): Analytical questions requiring step-by-step
hypothesis testing or multi-step numerical solutions (e.g., Chi-square, t -test,
Regression).
Group B (Short Questions): Numerical calculations (e.g., Correlation,
Estimation, Central Tendency) and theoretical definitions/distinctions (e.g., Type I
vs. Type II errors, Parametric vs. Non-parametric tests).
Unit 1: Introduction to Statistics
1. Key Theoretical Concepts
Statistics in Public Administration: The science of collecting, organizing,
analyzing, and interpreting numerical data to inform public policy, resource
allocation, and program evaluation.
Descriptive vs. Inferential Statistics:
o Descriptive: Methods for summarizing and presenting data (e.g., mean,
charts, standard deviation).
o Inferential: Methods for making generalizations or drawing conclusions
about a population based on a sample (e.g., hypothesis testing, confidence
intervals).
2. Measures of Central Tendency
Mean ( X ): Arithmetic average.
X=
∑ X or weighted mean: X = ∑ w X
∑w
w
n
Median ( M d ): Middle value dividing data into two equal halves. Positional average
unaffected by extreme outliers.
Mode ( M o): Most frequently occurring value. Useful for categorical data in public
service preferences.
3. Measures of Dispersion
Variance ( s ): Average of squared deviations from the mean.
2
Standard Deviation ( s or σ ): Absolute measure of dispersion in original units.
s=
√ ∑ ( X −X )2
n−1
Coefficient of Variation (CV ): Relative measure of dispersion used to compare
variability between two datasets with different units.
s
CV = × 100 %
X
> Exam Tip: Lower CV indicates greater consistency/uniformity in data.
Unit 2: Determining Bi-Variate Relationships
1. Correlation Analysis
Measures the strength and direction of the linear association between two variables ( X
and Y ).
Karl Pearson’s Coefficient of Correlation ( r ): Parametric measure for
interval/ratio data. Range: −1 ≤r ≤+1.
n ∑ X Y −( ∑ X )( ∑ Y )
r=
√ [ n∑ X −(∑ X ) ][ n ∑ Y −(∑ Y ) ]
2 2 2 2
Spearman’s Rank Correlation ( ρ or r s): Non-parametric measure used when data
are in ranks or ordinal scale.
6∑ d
2
r s=1−
n ( n2−1 )
(where d=R 1−R2 is the difference between ranks).
2. Regression Analysis
Used to predict the value of a dependent variable (Y ) based on an independent variable ( X
).
Regression Equation of Y on X :
Y =a+bX
n ∑ X Y −( ∑ X )( ∑ Y )
o Slope Coefficient (b ): b= 2
n ∑ X −( ∑ X )
2
o Intercept (a ): a=Y −b X
Unit 3: Probability Theory
1. Core Approaches & Terminology
Sample Space ( S): The set of all possible outcomes of a random experiment.
Approaches to Probability:
Favorable Outcomes
1. Classical: P ( A )= Total Possible Outcomes (assumes equally likely outcomes).
2. Relative Frequency: Based on empirical historical data over repeated trials.
3. Subjective: Based on expert judgment, intuition, or personal belief.
2. Laws of Probability
Addition Rule:
o Mutually Exclusive: P ( A ∪ B )=P ( A ) + P ( B )
o Non-Mutually Exclusive: P ( A ∪ B )=P ( A ) + P ( B ) −P ( A ∩ B )
Multiplication Rule & Conditional Probability:
o Independent Events: P ( A ∩B )=P ( A ) × P ( B )
P ( A ∩B )
o Conditional Probability: P ( A∨B ) =
P( B)
3. Theoretical Probability Distributions
Binomial Distribution: Discrete distribution with n independent trials, 2 outcomes
(success/failure), constant probability p.
()
P ( X=k )= n p (1−p ) , Mean=np , Variance=np ( 1−p )
k
k n−k
Normal Distribution: Continuous, symmetric, bell-shaped distribution defined by
mean ( μ) and standard deviation (σ ).
X−μ
Z= ∼ N (0 , 1)
σ
Unit 4: Estimation
1. Key Concepts
Point Estimate: Single numerical value used to estimate a population parameter
(e.g., using X for μ).
Interval Estimate: A range of values within which the true parameter is expected
to lie with a specified confidence level ( 1−α ) .
Standard Error ( SE): Standard deviation of a sampling distribution. For sample
σ s
mean: SE ( X )= or .
√n √n
2. Criteria of a Good Estimator
1. Unbiasedness: Expected value equals the true parameter ( E ( θ́ ) =θ).
2. Consistency: Estimator approaches the true parameter as sample size increases.
3. Efficiency: Has the minimum variance among all unbiased estimators.
4. Sufficiency: Utilizes all information contained in the sample.
3. Confidence Interval Formulas
For Population Mean ( μ):
s
X ± Z α /2 ⋅ ( if n≥ 30 )
√n
For Population Proportion ( P):
4. Sample Size Determination
p ± Zα/ 2
√ p ( 1−p )
n
( )
2
Z α /2 ⋅σ
For Estimating Mean: n=
E
2
Z α / 2 ⋅ p ( 1− p )
For Estimating Proportion: n= 2
E
Unit 5: Test of Hypothesis (High Weightage)
1. Core Terminology & Errors
Null Hypothesis ( H 0 ): Hypothesis of no effect, no difference, or status quo.
Alternative Hypothesis ( H 1): Claim to be tested ( directional ≠ ,<,>¿ ).
Type I Error (α ): Rejecting H 0 when it is actually true (Producer's risk).
Type II Error ( β ): Failing to reject H 0 when it is actually false (Consumer's risk).
2. The Standard 6-Step Test Procedure
1. State H 0 and H 1.
2. Select significance level (α , usually 0.05 or 0.01).
3. Identify test statistic ( Z , t , χ 2, or H ).
4. Determine critical value / rejection region.
5. Calculate test statistic from sample data.
6. Make decision: If ¿ Calculated∨¿ Critical, reject H 0; otherwise do not reject H 0.
3. Quick Reference Table for Statistical Tests
Test Type Application / Conditions Test Statistic Formula
Z-Test (Large n ≥ 30, Single or double X−μ0
Z=
Sample) mean/proportion s / √n
X 1− X 2
Small samples (n<30 ), unknown t=
Independent t -Test σ , comparing 2 means sp
√ 1 1
+
n1 n2
Dependent samples d
Paired t -Test (before/after measurements on t= where d= X 1−X 2
sd / √ n
same subjects)
Categorical data, comparing
Chi-Square ( χ 2) ( O−E )2
observed vs expected χ 2= ∑
Goodness of Fit E
distributions
Row Total ×Column Total
Chi-Square ( χ 2) Categorical data, contingency E= ,
Grand Total
Independence table (r × c) df =( r−1 )( c−1 )
Test Type Application / Conditions Test Statistic Formula
Non-parametric alternative to 2
Kruskal-Wallis 12 Ri
Test ( H )
one-way ANOVA (>2 H=
N ( N +1 )
∑ ni
−3 ( N +1 )
independent samples)
Parametric vs. Non-Parametric Tests
Feature Parametric Tests Non-Parametric Tests
Assumes normal distribution & Distribution-free, suitable for
Assumptions
interval/ratio metric scale ordinal/nominal data
Sample Size Preferred for large samples Useful for small samples
Power Higher statistical power Slightly lower statistical power
Z -test, t -test, F -test
Chi-square, Kruskal-Wallis,
Examples
Spearman Rank
Solution
Tribhuvan University — MPA 509: Statistics for Public Administration
Exam-Oriented Smart Notes & Revision Guide
Exam Strategy & Question Pattern (TU Format)
TU MPA examinations typically assess both theoretical concepts and numerical
problem-solving:
Group A (Long Questions): Analytical questions requiring step-by-step
hypothesis testing or multi-step numerical solutions (e.g., Chi-square, t -test,
Regression).
Group B (Short Questions): Numerical calculations (e.g., Correlation,
Estimation, Central Tendency) and theoretical definitions/distinctions (e.g., Type I
vs. Type II errors, Parametric vs. Non-parametric tests).
Unit 1: Introduction to Statistics
1. Key Theoretical Concepts
Statistics in Public Administration: The science of collecting, organizing,
analyzing, and interpreting numerical data to inform public policy, resource
allocation, and program evaluation.
Descriptive vs. Inferential Statistics:
o Descriptive: Methods for summarizing and presenting data (e.g., mean,
charts, standard deviation).
o Inferential: Methods for making generalizations or drawing conclusions
about a population based on a sample (e.g., hypothesis testing, confidence
intervals).
2. Measures of Central Tendency
Mean ( X ): Arithmetic average.
X=
∑ X or weighted mean: X = ∑ w X
∑w
w
n
Median ( M d ): Middle value dividing data into two equal halves. Positional average
unaffected by extreme outliers.
Mode ( M o): Most frequently occurring value. Useful for categorical data in public
service preferences.
3. Measures of Dispersion
Variance ( s ): Average of squared deviations from the mean.
2
Standard Deviation ( s or σ ): Absolute measure of dispersion in original units.
s=
√ ∑ ( X −X )2
n−1
Coefficient of Variation (CV ): Relative measure of dispersion used to compare
variability between two datasets with different units.
s
CV = × 100 %
X
> Exam Tip: Lower CV indicates greater consistency/uniformity in data.
Unit 2: Determining Bi-Variate Relationships
1. Correlation Analysis
Measures the strength and direction of the linear association between two variables ( X
and Y ).
Karl Pearson’s Coefficient of Correlation ( r ): Parametric measure for
interval/ratio data. Range: −1 ≤r ≤+1.
n ∑ X Y −( ∑ X )( ∑ Y )
r=
√ [ n∑ X −(∑ X ) ][ n ∑ Y −(∑ Y ) ]
2 2 2 2
Spearman’s Rank Correlation ( ρ or r s): Non-parametric measure used when data
are in ranks or ordinal scale.
6∑ d
2
r s=1−
n ( n2−1 )
(where d=R 1−R2 is the difference between ranks).
2. Regression Analysis
Used to predict the value of a dependent variable (Y ) based on an independent variable ( X
).
Regression Equation of Y on X :
Y =a+bX
n ∑ X Y −( ∑ X )( ∑ Y )
o Slope Coefficient (b ): b= 2
n ∑ X −( ∑ X )
2
o Intercept (a ): a=Y −b X
Unit 3: Probability Theory
1. Core Approaches & Terminology
Sample Space ( S): The set of all possible outcomes of a random experiment.
Approaches to Probability:
Favorable Outcomes
1. Classical: P ( A )= Total Possible Outcomes (assumes equally likely outcomes).
2. Relative Frequency: Based on empirical historical data over repeated trials.
3. Subjective: Based on expert judgment, intuition, or personal belief.
2. Laws of Probability
Addition Rule:
o Mutually Exclusive: P ( A ∪ B )=P ( A ) + P ( B )
o Non-Mutually Exclusive: P ( A ∪ B )=P ( A ) + P ( B ) −P ( A ∩ B )
Multiplication Rule & Conditional Probability:
o Independent Events: P ( A ∩B )=P ( A ) × P ( B )
P ( A ∩B )
o Conditional Probability: P ( A∨B ) =
P( B)
3. Theoretical Probability Distributions
Binomial Distribution: Discrete distribution with n independent trials, 2 outcomes
(success/failure), constant probability p.
()
P ( X=k )= n p (1−p ) , Mean=np , Variance=np ( 1−p )
k
k n−k
Normal Distribution: Continuous, symmetric, bell-shaped distribution defined by
mean ( μ) and standard deviation (σ ).
X−μ
Z= ∼ N (0 , 1)
σ
Unit 4: Estimation
1. Key Concepts
Point Estimate: Single numerical value used to estimate a population parameter
(e.g., using X for μ).
Interval Estimate: A range of values within which the true parameter is expected
to lie with a specified confidence level ( 1−α ) .
Standard Error ( SE): Standard deviation of a sampling distribution. For sample
σ s
mean: SE ( X )= or .
√n √n
2. Criteria of a Good Estimator
1. Unbiasedness: Expected value equals the true parameter ( E ( θ́ ) =θ).
2. Consistency: Estimator approaches the true parameter as sample size increases.
3. Efficiency: Has the minimum variance among all unbiased estimators.
4. Sufficiency: Utilizes all information contained in the sample.
3. Confidence Interval Formulas
For Population Mean ( μ):
s
X ± Z α /2 ⋅ ( if n≥ 30 )
√n
For Population Proportion ( P):
4. Sample Size Determination
p ± Zα/ 2
√ p ( 1−p )
n
( )
2
Z ⋅σ
For Estimating Mean: n= α /2
E
2
Z α/ 2 ⋅ p ( 1− p )
For Estimating Proportion: n= 2
E
Unit 5: Test of Hypothesis (High Weightage)
1. Core Terminology & Errors
Null Hypothesis ( H 0 ): Hypothesis of no effect, no difference, or status quo.
Alternative Hypothesis ( H 1): Claim to be tested ( directional ≠ ,<,>¿ ).
Type I Error (α ): Rejecting H 0 when it is actually true (Producer's risk).
Type II Error ( β ): Failing to reject H 0 when it is actually false (Consumer's risk).
2. The Standard 6-Step Test Procedure
1. State H 0 and H 1.
2. Select significance level (α , usually 0.05 or 0.01).
3. Identify test statistic ( Z , t , χ 2, or H ).
4. Determine critical value / rejection region.
5. Calculate test statistic from sample data.
6. Make decision: If ¿ Calculated∨¿ Critical, reject H 0; otherwise do not reject H 0.
3. Quick Reference Table for Statistical Tests
Test Type Application / Conditions Test Statistic Formula
Z-Test (Large n ≥ 30, Single or double X−μ0
Z=
Sample) mean/proportion s / √n
X 1− X 2
Small samples (n<30 ), unknown t=
Independent t -Test σ , comparing 2 means sp
√ 1 1
+
n1 n2
Dependent samples d
Paired t -Test (before/after measurements on t= where d= X 1−X 2
sd / √ n
same subjects)
Categorical data, comparing
Chi-Square ( χ 2) ( O−E )2
observed vs expected χ =∑
2
Goodness of Fit E
distributions
Row Total ×Column Total
Chi-Square ( χ 2) Categorical data, contingency E= ,
Grand Total
Independence table (r × c) df =( r−1 )( c−1 )
Non-parametric alternative to 2
Kruskal-Wallis 12 Ri
Test ( H )
one-way ANOVA (>2 H=
N ( N +1 )
∑ ni
−3 ( N +1 )
independent samples)
Parametric vs. Non-Parametric Tests
Feature Parametric Tests Non-Parametric Tests
Assumes normal distribution & Distribution-free, suitable for
Assumptions
interval/ratio metric scale ordinal/nominal data
Sample Size Preferred for large samples Useful for small samples
Power Higher statistical power Slightly lower statistical power
Z -test, t -test, F -test
Chi-square, Kruskal-Wallis,
Examples
Spearman Rank