Multivariate Analysis

Read Complete Research Material

MULTIVARIATE ANALYSIS

An account of a multivariate analysis of a provided data file, using multiple regressions



An account of a multivariate analysis of a provided data file, using multiple regressions

Assignment - 1

Requirement A:

We have use general linear model on SPSS to test the impact of independent variable on dependent variable. GLM is used when the independent variable is scale but the independent variables are combination of scale and categorical variable. Therefore, there were some categorical independent variables like gender, group, handed etc and the result is discussed below:

Tests of Between-Subjects Effects

Dependent Variable:time1

Source

Type III Sum of Squares

df

Mean Square

F

Sig.

Corrected Model

462986.598a

7

66140.943

6.893

.000

Intercept

282159.824

1

282159.824

29.407

.000

keen

17519.719

3

5839.906

.609

.610

group

.000

0

gender

320452.985

1

320452.985

33.398

.000

hand2

34805.584

1

34805.584

3.628

.058

dumgroup

.000

0

age

65934.457

1

65934.457

6.872

.009

Error

2341144.875

244

9594.856

Total

5308755.000

252

Corrected Total

2804131.472

251

a. R Squared = .165 (Adjusted R Squared = .141)

The above table “Tests of Between-Subjects Effects” shows the impact of independent variable on time, which is our study variable. Here in this table the impact is shown as whole mean this table does not show the impact of an individual category. For example, the gender has impact on the time taken by the observant but it does not tell which gender takes more time and which takes less time and there might be no impact of either gender. Now, we just simply discuss the impact of independent variables understudy as whole. The keenness has no impact on the time taken by the observant as sig value for this variable is above 0.05 (0.61) but that does not mean that any keen level that we are studying have no impact on the time taken, therefore we discuss individual impact of each category in further analysis. Similarly, being a right handed or left handed has nothing to do with the time take as this variable also has p value greater than 0.05. On the other hand, intercept gender and age has significant impact on the study variable, so we can say that variation in gender and age can influences that time taken by the coffee drinker.

Parameter Estimates

Dependent Variable:time1

Parameter

B

Std. Error

t

Sig.

95% Confidence Interval

Lower Bound

Upper Bound

Intercept

279.517

44.205

6.323

.000

192.446

366.589

[keen=1]

-15.662

18.102

-.865

.388

-51.317

19.994

[keen=2]

-17.770

17.705

-1.004

.317

-52.644

17.103

[keen=3]

.684

17.723

.039

.969

-34.226

35.594

[keen=4]

0a

[group=1]

-38.955

12.770

-3.051

.003

-64.107

-13.802

[group=2]

0a

[gender=1]

-76.752

13.281

-5.779

.000

-102.911

-50.592

[gender=2]

0a

[hand2=1]

-26.674

14.005

-1.905

.058

-54.261

.912

[hand2=2]

0a

[dumgroup=0]

0a

[dumgroup=1]

0a

age

-4.397

1.677

-2.621

.009

-7.702

-1.093

a. This parameter is set to zero because it is redundant.

The above table parameter estimates has benchmarked one category form each level as to measure the impact of other category (ies) in respect variable against the benchmarked variable. This table shows that any level of keenness does effect the time taken by the coffee drinker. Similarly hand used by the coffee drinker while taking the coffee does not influence the time taken. On the other hand group (drank coffee as part of the experiment) shows greater influence on the time taken than the other group (drank decaffeinated coffee as part of the experiment) that is set as benchmark. The beta of the group 1 is -38.955, which mean this group takes less time than other one (p<0.05). Similarly, gender (female) takes less time than he males so as age of the coffee drinker.

Requirement B:

Tests of Between-Subjects Effects

Dependent Variable:time10

Source

Type III Sum of Squares

df

Mean Square

F

Sig.

Corrected Model

1187.956a

7

169.708

9.050

.000

Intercept

6.057

1

6.057

.323

.570

keen

106.825

3

35.608

1.899

.130

group

.000

0

gender

623.560

1

623.560

33.254

.000

hand2

1.717

1

1.717

.092

.762

dumgroup

.000

0

age

446.364

1

446.364

23.804

.000

Error

4575.314

244

18.751

Total

27122.000

252

Corrected Total

5763.270

251

a. R Squared = .206 (Adjusted R Squared = .183)

Here in this table, we just simply discuss the impact of independent variables understudy as ...
Related Ads