Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 5

Research Methods II

Ayesha Arshad

362475

MS CP-21

Assignment # 2

Submitted to

Dr. Sumara Masood

Department of Behavioral Sciences

School of social sciences and economics

National University of Science and Technology


1.1. Procedure of Dummy coding:
Dummy variables of the categorical variable of “Highest Education level” were created to

represent the categories of this independent variable. Primary education has 6 categories:

Primary, Some Secondary, Completed High School, Some Additional Training, Completed

Undergraduate, and Postgraduate Completed. Since there are 6 categories, 5 dummy categories

(K-1) were created by using primary education as a reference category. The reference category

would be the ‘missing’ category.

Following are the steps for dummy coding for this variable

 Checking the original values on the variable “Highest Education level”

 From the menu at the top, click transform, select recode into different variable

 In the input variable --->output variable drag the original education variable

 Write the name and label you want to assign to the new variable (e.g. in the name writing

“somesec” and labelling it as “some secondary education”, “someAT and labeling it as

“some additional training”).

 In the column left “old values”, write the old value for some secondary education i.e.,

“2”, on the right in the column for the new value “1”, click on “add.

 Selecting “all other values” in the old value column and giving them the new value of

“0”, Click on “add”.

 Select system-or-user missing values from the left and system-missing on the right

column. Click “add:

 Click on continue

 Go on the “variable view” and check if your new variable is added.


 Repeat the same procedure for “some secondary education”, “undergraduate completed”

and postgraduate completed” separately i.e. labeling them as “1” and all others as “0”

(primary education being the reference category remains zero for all categories)

1.2. Variance caused in Negative Affect by all the variables

Overall, the variables are contributing 31.7% variance in Negative Affect, given by value of

R2=.31

Variables R2
Total PCOISS .234
Total Mastery .295
Some additional training .307
Age .317

1.3. Best predictor of Negative affect?

Best predictor of dependent variable, Negative Affect, is “Perceived control of internal

states” with standardized coefficient, β = -.33, which is higher than standardized coefficients i.e.

-.294, -.102, -.104 of the other three variables.

1.4. Do all the variables are Statistically significance contribution in predicting negative

affect?

All variables are statistically significant with p> 0.05 i.e. Total PCOISS (p=.000), Total

Mastery (p= .000), SAT (p= .012). and Age (p= .014).
Table 1.1
Stepwise Regression Results for Negative Affect

Variable B 95% CI β t R2 Δ R2

LL UL

Step 1
(Constant) 36.70*** 33.67 37.94 23.74 .23 .23
PCOISS -.28*** -.34 -.24 -.48 -11.40

Step 2
(Constant) 42.56*** 39.08 46.03 24.06 .29 .06
PCOISS -.20*** -.25 -.14 -.33 -6.96
Mastery -.52*** -.68 -.35 -.29 -6.09

Step 3
Constant 43.14*** 39.66 46.62 24.38 .30 .01
PCOISS -.20*** -.25 -.14 -.33 -7.01
Mastery -.52*** -.69 -.36 -.29 -6.21
Some Additional Training -1.70* -2.96 -.44 -.11 -2.65

Note CI= confidence interval; LL= lower limit; UL= upper limit

***p<.001, **p<.01, *p<.05

Step 1= F (1, 426) = 130.04; Sig-F= .000 ; R=.48; Adj. R2= .23 ; F Change = 130.04 ; Sig-F Change= .000

Step 2= F (2,425) = 89.09 ; Sig-F= .000 ; R=.54 ; R= Adj. R2= .29 ; F Change = 37.13 ; Sig-F Change= .000

Step 3= F (3, 424) = 62.57 ; Sig-F= .000 ; R=.55 ; R= Adj. R2=. 30 ; F Change = 6.99 ; Sig-F Change= .008

Step 4= F (4, 423) = 49.00 ; Sig-F= .000 ; R=.56 ; R= Adj. R2= .31; F Change = 6.06; ; Sig-F Change= .014
1.5. Interpretation

Stepwise regression was conducted to find which predictor i.e., PCOISS, Total mastery,

SAT, Age contributes significantly in sequence towards Negative Affect. Preliminary analyses

were conducted to ensure no violation of assumptions of normality, linearity, multicollinearity,

and homoscedasticity. In Step 1, R2 value of .23 revealed that PCOISS explained 23% variance

in Negative affect with F (1, 426) = 130.042, p < .001. The findings revealed that PCOISS

negatively predicted Negative Affect (β= -.48, p < .001). In step 2, after entering Total Mastery

and PCOISS, the total variance explained by the model as a whole was 29%, F (2, 425) = 89.09,

p < .001. The findings revealed that PCOISS (β= -.33, p < .001) and Total Mastery (β= -.29, p

< .001) negatively predicted Negative Affect. In step 3, R2 value of .30 revealed that PCOISS,

after adding Total Mastery and Some Additional Training 30% variance was explained in

Negative affect with F (3,424) = 62.57, p < .001. Results revealed that PCOISS (β= -.33, p

< .001), Total Mastery (β= -.29, p < .001) and Some Additional Training (β= -.11, p < .001)

negatively predicted Negative Affect. In step 4, after the entry of age, some additional training,

total mastery and total PCOISS, the total variance contributed by model as a whole was 31%, F

(4, 423) =49.00, p < .001. Results revealed that PCOISS (β= -.33, p < .001), Total Mastery (β=

-.29, p < .001) and Some Additional Training (β= -.11, p < .001) and Age (β= -.10, p < .001)

negatively predicted Negative Affect. The remaining categories of Education i.e. Some

Secondary (t= .641, p> .001), Completed High School (t= .405, p > .001), Completed

Undergraduate (t= -1.021, p > .001), and Postgraduate Completed (t= .417, p > .001) were

excluded from the analysis as they were not significantly contributing towards Negative Affect.

You might also like