Bivariate and Multivariate Analysis Exam Questions and Answers
Bivariate and Multivariate Analysis Exam Q&A
Unit 1: Important Exam Questions with Answers
Q1. Correlation Analysis: Types and Properties
Answer:
Meaning of Correlation: Correlation is a statistical technique used to measure the degree and direction of the relationship between two variables. It shows whether variables move together or in opposite directions.
The coefficient of correlation is denoted by: r
Its value lies between: -1 and +1
Types of Correlation
- 1. Positive Correlation: When two variables move in the same direction, it is called positive correlation.
Example: An increase in income increases expenditure. - 2. Negative Correlation: When one variable increases and the other decreases, it is called negative correlation.
Example: An increase in price decreases demand. - 3. Zero Correlation: When no relationship exists between variables, it is called zero correlation.
Example: Weight of a person and petrol prices.
Properties of Correlation Coefficient
- Range: The value of the correlation coefficient lies between -1 and +1.
- Independent of Units: Correlation is not affected by units of measurement.
- Symmetrical Nature: rxy = ryx
- Indicates Direction:
- a. Positive sign shows a direct relationship.
- b. Negative sign shows an inverse relationship.
- Independent of Origin and Scale: Adding or multiplying constants does not affect correlation.
- Measures Degree of Relationship: A higher value indicates a stronger relationship.
Importance of Correlation
- a. Helps in prediction
- b. Useful in business forecasting
- c. Helps in research analysis
- d. Useful in economics and finance
Conclusion: Correlation analysis helps researchers and businesses understand relationships between variables and supports effective decision-making.
Q2. Regression Analysis: Assumptions and Objectives
Answer:
Meaning of Regression: Regression is a statistical method used to estimate or predict the value of one variable based on another variable.
General Equation: Y = a + bX
Where:
- Y = Dependent variable
- X = Independent variable
- a = Constant
- b = Regression coefficient
Objectives of Regression Analysis
- Prediction of Future Values: Regression helps forecast sales, demand, profits, etc.
- Studying Cause and Effect Relationship: It explains how one variable affects another.
- Business Planning: Used in budgeting and planning.
- Estimation: Helps estimate unknown values.
Assumptions of Regression Analysis
- Linear Relationship: Variables should have a linear relationship.
- No Measurement Error: Data should be accurate.
- Normally Distributed Errors: Residual errors should follow a normal distribution.
- Independence of Variables: Variables should not strongly depend on each other.
Significance of Regression Analysis
- Helps in Decision Making: Businesses use regression for planning.
- Sales Forecasting: Companies predict future sales.
- Financial Analysis: Useful in investment and risk analysis.
- Market Research: Helps understand customer behavior.
Conclusion: Regression analysis is an important statistical tool for prediction, estimation, and business decision-making.
Q3. Simple vs. Multiple Regression Analysis
Answer:
Simple Regression: Simple regression studies the relationship between one dependent variable and one independent variable.
Equation: Y = a + bX
Where:
- Y = Dependent variable
- X = Independent variable
Example: Relationship between advertisement expenditure and sales.
Features of Simple Regression
- a. One dependent variable
- b. One independent variable
- c. Linear relationship
- d. Easy interpretation
Multiple Regression: Multiple regression studies the relationship between one dependent variable and two or more independent variables.
Equation: Y = β0 + β1X1 + β2X2 + … + βnXn + e
Example: Sales affected by:
- a. Price
- b. Advertisement
- c. Income level
- d. Competition
Features of Multiple Regression
- a. Multiple independent variables
- b. More accurate prediction
- c. Used in business research
- d. Complex analysis
Difference between Simple and Multiple Regression
| Basis | Simple Regression | Multiple Regression |
|---|---|---|
| Independent Variables | One | More than one |
| Complexity | Simple | Complex |
| Prediction Accuracy | Lower | Higher |
| Variables | Two variables | More than two variables |
Conclusion: Simple regression studies one predictor variable while multiple regression studies several predictor variables simultaneously.
Q4. Multicollinearity, VIF, and R-Square
Answer:
Meaning of Multicollinearity: Multicollinearity occurs when independent variables in regression are highly correlated with each other. It creates problems in estimating accurate regression coefficients.
Problems of Multicollinearity
- a. Reduces reliability of regression model
- b. Creates unstable coefficients
- c. Difficult to identify impact of variables
- d. Increases standard errors
Variance Inflation Factor (VIF): VIF measures the degree of multicollinearity among independent variables.
Formula: VIF = 1 / (1 – R2)
Interpretation of VIF:
R-Square (Coefficient of Determination): R-square measures the proportion of variation in the dependent variable explained by independent variables.
Value lies between: 0 and 1
A higher value indicates a better model fit.
Adjusted R-Square: Adjusted R-square modifies R-square according to the number of variables included in the model. It provides a more accurate result in multiple regression.
Difference between R-Square and Adjusted R-Square
| Basis | R-Square | Adjusted R-Square |
|---|---|---|
| Meaning | Measures explained variation | Adjusted measure of variation |
| Effect of Variables | Always increases | May increase or decrease |
| Accuracy | Less accurate | More accurate |
| Use | Simple models | Multiple regression |
Conclusion: Multicollinearity affects regression accuracy while VIF, R-square, and Adjusted R-square help evaluate model quality.
Q5. Partial Correlation and Stepwise Regression
Answer:
Partial Correlation: Partial correlation measures the relationship between two variables while controlling the effect of another variable. It shows the pure relationship between variables.
Example: Suppose a researcher studies the relationship between salary and performance while controlling for experience. Partial correlation removes the influence of experience.
Importance of Partial Correlation
- a. Removes unwanted influence
- b. Improves accuracy
- c. Useful in research analysis
- d. Helps identify true relationship
Stepwise Regression: Stepwise regression is a method of selecting important independent variables in regression analysis. Variables are added or removed automatically based on statistical significance.
Process of Stepwise Regression
- Selection of Most Significant Variable: The most important variable enters first.
- Addition of Variables: Additional significant variables are added.
- Removal of Insignificant Variables: Unimportant variables are removed.
Advantages of Stepwise Regression
- a. Reduces unnecessary variables
- b. Improves prediction accuracy
- c. Saves time
- d. Simplifies model
Disadvantages
- a. May ignore theoretical importance
- b. Can create unstable models
Conclusion: Partial correlation identifies pure relationships while stepwise regression selects the best variables for prediction models.
Q6. Durbin-Watson Statistic in Regression
Answer:
Meaning of Durbin-Watson Statistic: The Durbin-Watson statistic is used to detect autocorrelation in residual errors of a regression model. Autocorrelation means residual errors are correlated with each other.
The value of the Durbin-Watson statistic lies between: 0 and 4
Interpretation:
| Value | Meaning |
|---|---|
| 2 | No autocorrelation |
| Less than 2 | Positive autocorrelation |
| Greater than 2 | Negative autocorrelation |
Importance of Durbin-Watson Statistic
- Detects Autocorrelation: Checks whether residuals are independent.
- Improves Accuracy: Helps improve the reliability of the regression model.
- Important in Time Series Analysis: Widely used in economic and financial data.
- Validates Regression Assumptions: Ensures assumptions are satisfied.
Example: A company analyzing monthly sales data may use the Durbin-Watson statistic to detect correlation among residual errors.
Conclusion: The Durbin-Watson statistic is an important tool used to test autocorrelation and improve the reliability of regression analysis.
Q7. Bivariate vs. Multivariate Analysis
Answer:
| Basis | Bivariate Analysis | Multivariate Analysis |
|---|---|---|
| Meaning | Analysis of two variables | Analysis of more than two variables |
| Variables | One dependent and one independent variable | Multiple variables |
| Complexity | Simple | Complex |
| Objective | Study relationship between two variables | Study combined effect of multiple variables |
| Techniques | Correlation, simple regression | Multiple regression, factor analysis |
| Example | Income and expenditure | Sales affected by price, income, and advertisement |
| Nature | Limited analysis | Comprehensive analysis |
| Accuracy | Less accurate | More accurate |
Importance of Multivariate Analysis
- a. Better prediction
- b. Studies complex relationships
- c. Widely used in research
- d. Helps business decisions
Conclusion: Bivariate analysis studies the relationship between two variables while multivariate analysis studies the relationship among multiple variables simultaneously.