Scribd
Upload
207 views

Linear Regression Analysis

The document provides instructions for conducting and presenting a linear regression analysis for an economics project. It describes 4 key steps: 1) selecting dependent and independent variables and collecting data, 2) specifying the regression model equation, 3) analyzing the data output, and 4) interpreting the results. The analysis is demonstrated on a model of exchange rate (dependent variable) regressed on inflation and trade openness (independent variables). The output shows coefficients, R-squared, and model significance. The interpretation identifies the significant factors, direction of impact, and goodness of fit.

Uploaded by

Ehtasham ul haq
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
207 views

Linear Regression Analysis

The document provides instructions for conducting and presenting a linear regression analysis for an economics project. It describes 4 key steps: 1) selecting dependent and independent variables and collecting data, 2) specifying the regression model equation, 3) analyzing the data output, and 4) interpreting the results. The analysis is demonstrated on a model of exchange rate (dependent variable) regressed on inflation and trade openness (independent variables). The output shows coefficients, R-squared, and model significance. The interpretation identifies the significant factors, direction of impact, and goodness of fit.

Uploaded by

Ehtasham ul haq
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Linear Regression Analysis

For Final Project Economics


You are required to follow these steps for doing and presenting Linear Regression Analysis.

Step # 01: Selection of Variables


In this step, you are required to identify the specific factors (independent variables) which
affect the (dependent variable). For the purpose of identification, you must study research
articles relevant to your chosen topic. (You are also advised to seek guidance from the
instructor before finalizing the variables).

After the selection, you will provide brief introduction of the selected variables (dependent as
well as independent variables) along with the sources from where the data has been collected. It
is highly recommended to collect data from World Development Indicators (WDI) at least for
the recent FIVE years or more and test for factors affecting the dependent variable. Also find
evidence that out of the selected independent variables which variable is/are significant in
determining the dependent variable by using Multiple Linear Regression Analysis.

This step should be presented in the following way

Data Data
Dependent Independent Year Under Analysis
Analysis Collection
Variable Variables Consideration Technique
Software Source
20XX – 20YY Regression MS-Excel/ WDI
E-Views
*Choose at least 3 independent variables for regression analysis.

Step # 02: Model Specification


In this step you will be required to present your model in mathematical equation form.
For example,

ER = α + β1INF + β2INT+ β3TRO + ε

Where:
ER = Exchange Rate
INF = Inflation Rate
INT = Interest Rate
TRO = Trade Openness
ε = Error/disturbance/residual
You can also do regression one by one by selecting each independent variable separately in
each model to avoid the problem of multi-collinearity. Search articles related to your chosen
topic to see the methodology.

Depending to the nature of the variables and data, you can also take the model in logarithm
form to make the same units of all variables. In that case, the model can be written as:

Log(ER) = α + β1log(INF) + β2log(INT)+ β3log(TRO) + ε

After taking log, the coefficients will be interpreted as elasticities in percentage.

Step # 03: Data Analysis


In this step, you are required to show the output results of the linear regression. Procedure of
how to run a Linear Regression in Excel is given in later part of the document.

SUMMARY
OUTPUT

Regression Statistics
Multiple R 0.5173425
R Square 0.2676432
Adjusted R
Square 0.0235243
Standard
Error 2.217208
Observations 9

ANOVA
Significance
df SS MS F
F
Regression 2 10.779481 5.3897409 1.0963644 0.3927968
Residual 6 29.496073 4.916012
Total 8 40.275555

Standard Upper Lower Upper


Coefficients t Stat P-value Lower 95%
Error 95% 95% 95%
Intercept 1.6181742 1.9731876 0.8200812 0.4435207 -3.2100412 6.4463904 -3.210041 6.4460468
INF 0.2159459 0.3262453 0.6619129 0.5326100 -0.5823475 1.0142395 -0.582347 1.0142335
TRO 0.3276435 0.3317540 0.9876097 0.361481 -0.4841294 1.1394165 -0.484129 1.1394165
Step # 04: Interpretation of the Results
In this step you are required to determine the significant factors that have an impact on
dependent variable (Exchange Rate) and also identify their impact level. So, these findings
should be based on the interpretation of:
1. Intercept Coefficient
2. Slope Coefficients along with direction of relationship (being
positive/negative)
3. R square
4. Model Significance (as per Prob (F-statistic))

1. a = 1.618 (Intercept Coefficient)


2. b1 = 0.216 (Coefficient of Inflation Rate)
b2 = 0.327 (Coefficient of Trade Openness)

So our regression equation is ER = 1.618 + 0.216(INF) + 0.327(TRO)

b1 = 0.216 means that one percent increase in inflation rate brings 0.216 percent increase in
exchange rate, on average.

Note: In the above mentioned way, you are required to interpret all the slope coefficients.

Also in the regression statistics output gives the goodness of fit measure i.e. R square.

3. R square = 0.267 (This means 26.7% of variation in exchange rate is determined by


inflation rate and trade openness and the rest is unexplained due to the factors/variables not
considered/included in the study).

4. Model Significance (as per probability of F-statistics) = 1.096 (0.392) means that
model/results are not significant as p-value of F ‘0.392’ is greater than 0.05 (significance
level). So we can infer that there is no signification relationship among the dependent and
independent variables.

Important Notes
 Regression analysis based on the data older or less than recent five years will NOT be
accepted.
 You are required to submit you the Raw Data (excel file) which contains the data of
variables along with your project report.

Linear Regression using Excel

Installing the Analysis ToolPak


1. Open Microsoft Excel, click the File menu and select Options

The following window will open.


2. In the Microsoft Office button, go to excel options to click Add-ins
3. In the Add-Ins box, select Analysis ToolPak and click Go...

4. In the Add-Ins available box, check the Analysis ToolPak and then click OK.
5. If Analysis ToolPak is not listed in the Add-Ins available box, click Browse to locate
it.

Linear Regression using the Data Analysis Add-In


After installation of Analysis ToolPak you will find Data analysis button on the right corner
of data menu. Click this button
Following window will open.

1. In the data analysis tool select the regression and then click Ok.
Suppose we want to determine whether X is a function of Y and Z.

Xi = a + b1Yi + b2Zi+ errori


Where;

Xi = value of X for observation i


a = mean value of X (dependent variable) when Y and Z (independent Variable) is zero
(intercept coefficient)
b = average change in X given a one unit change in Y and Z, i.e. (slope of Y and Z)
b1 = value of Y for observation i
b2 = value of Z for observation i
2. After clicking the regression, a regression box appears.
3. Select the Inputs for Y range (Dependent Variable) in first box.
4. Select the Inputs for X range (Independent Variable) in 2nd box.
4. Select the place where you want your output
5. Check the Labels and then click Ok

1
2

The output is given in the coefficients column in the last set of output

1. Intercept Coefficient
2. Slope Coefficients
3. R square
4. Model Significance (as per Prob (F-statistic))

You might also like