Scribd
Upload
30 views

Assignment 3

The document discusses multiple regression analysis and calculations. It addresses topics like direct and indirect paths between independent and dependent variables, calculating coefficients, issues like multicollinearity, computing R-squared and effect sizes, and testing hypotheses about individual predictor coefficients.

Uploaded by

Trishla Gandhi
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
30 views

Assignment 3

The document discusses multiple regression analysis and calculations. It addresses topics like direct and indirect paths between independent and dependent variables, calculating coefficients, issues like multicollinearity, computing R-squared and effect sizes, and testing hypotheses about individual predictor coefficients.

Uploaded by

Trishla Gandhi
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Trishla Kamleshbhai Gandhi- Assignment 3

R12

B2 B1

Ans 7 : x1 and x2 are independent variables. Path from x1 to x2 is represented by


r12. B1 and b2 are coefficient showing strength of direct or unique relation of x1
and x2 to y. b1 is partial slope. It indicates how strongly x1 is related to y taking in
account the indirect relation of x1 via x2. The correlation of x1 and x2 should be
corrected to make sure nothing is double counted.

If there are multiple steps the coefficients long the path must be multiplied.
After calculating strength of association the , we have to sum the contributions
across path.

Direct path from x1 to y = b1

Indirect path from x1 to y = via x2 and 2 coefficients r12 and b2.

Total correlation = direct path+ indirect path.

Y = b1+ r12* b2

Likewise, for x2 to y = y = b2 + r12*b1


B1 = y- r12 y2 / 1-r2 12.

Also, if b1= 0 then entire path and correlation can be explained via indirect path
of x2.

Ans 8 : for the equation r1y = b1 + r12 * b2 ,

Total correlation = direct path+ indirect path.

Direct path = b1

Indirect path = r12 * b2

Ans 9 : Yes, for a regression it is possible to have significant R but non significant
b coefficient . it means that variables x1 and x2 collectively have the predictive
power but it’s not possible to determine b coefficient effectively.

Ans 10 : The addition of more independent variables creates more relationships


amongst them. So not only independent variables potentially related to
dependent variable, they also potentially related to each other. This is called
multicollinearity. If this happened, it creates difficulty in autonomous or separate
relapse coefficients. It can limit R size. It makes unique contributions of predictors
difficult. Confidence intervals around estimates of b are wider resulting in power
of statistical significance to be lower.

Ans 11: effect size for overall model – proportion of variance in y that is
predictable from x1 and x2 combined – is estimated by computation of R 2.

R2 = ssregression/ sstotal.

R2 shows linear relationship. R2 estimates the proportion of variance in y scores


that can be predicted from independent variables. It is an index of strength of
linear relationship.
Different effect size for overall regression equation = f2 = R2 / (1-R2)

Ans 12: For each predictor value in the regression for xi :

H0 : bi = 0

b = unknown population raw-score slope that is estimated by sample slope. If b i


coefficient for predictor x is statistically significant then there is significant
increase in predicted y values that is uniquely associated with x i. If we get a low p-
value/significance level, this means they're not likely to be 0, i.e. null
hypothesis can be rejected at low significant levels.

You might also like