Decision science

Presentation On Theory Of Decision Science

regression model with three explanatory variable

Of Life satisfaction.(Y)

Presented by:- ▪ Suhail Manjardekar 05

▪ Amar Itagi 47

▪ Shardul Thakker 38

▪ Kunal Sharma 61

Independent variables (X’s)

▪ Income (X1)

▪ Sprit (X2)

▪ Socio economic status of parents (X3)

Data collection

N Y X1 X2 X3

Data analysis using SPSS

1-fitting the regression model

2-Overall significance of the model & ANOVA table

Interpretation of the model

▪ Apriority Analysis.

▪ Statistical Analysis.

▪ Econometric Analysis.

Apriority Analysis

It is Assumed that

▪ Life satisfaction (Y) is α to Income (X1)

▪ Life satisfaction (Y) is α to Spirit (X2).

▪ Life satisfaction (Y) is α to Socio-economic status of parents(X3).

Statistical Analysis

▪ The regression model is “Y=16.472+0.123 X1+0.158 X2+0.174 X3”

1-ELASTICITY

▪ η1 (β1)= 0.169 under elastic (<1).

▪ η2 (β2)= 0.179 under elastic(<1).

▪ η3 (β3)= 0.1797 under elastic(<1).

2-OVER ALL SIGNIFICANCE OF THE MODEL

R square=0.264

Since R square < 0.7 the overall significance of the model is not good.

3-ANOVA table

▪ Since F value < F table therefore do not reject H0 and conclude that β1,β2,β3 are insignificant i.e. they are 0, and model is not good.

4-INDIVIDUAL TEST(t-test @5%level of significance)

▪ β0=1.676 < 2.120 (16) conclude that β0 is insignificant.

▪ β2=1.698 < 2.120 (16) conclude that β1 is insignificant.

▪ β3=0.982 < 2.120 (16) conclude that β2 is insignificant.

▪ β1=0.978 < 2.120 (16) conclude that β3 is insignificant.

Econometric Analysis

1-AUTOCORELATION

Since it lies between du and 2,there is no autocorrelation.

0 0.998 1.676 2 2.324 3.012 4

0 du dl 2 4-du 4-dl 4

d=1.821

2-MULTICOLINEARITY

▪ Since in F test and individual test we are not rejecting H0 i.e. in both the cases the β’s are insignificant or zero; there is no multicolinearity in the model.

Also

▪ VIF (variance inflation factors) for β0=1.062, β2=1.091, β3=1.037 are < 10 therefore there is no multicolinearity exsist in the model.