Model fitted in Step 5 is more adequate . But there is very little increment in step 5 from step 4.
(Male) = 1 is respondent is male other wise 0.
(Business) = 1 if respondent from business school otherwise 0
(Liberal Studies) = 1 if respondent from Liberal Studies school otherwise 0
(Science) = 1 if respondent from Science school otherwise 0
(Assistant Professor) = 1 if respondent is assistant professor otherwise 0
(Professor) = 1 if respondent is professor otherwise 0
If respondent is male then there is $3538.509 increment in salary.
If respondent is from business school then there is $26979.21 increment in salary.
If respondent is from liberal study school then there is $8059.37 decrement in salary.
If respondent is from science school then there is $6197.97 decrement in salary.
If respondent is assistant professor then there is $10153.4 decrement in salary.
If respondent is professor then there is $23431.59 increment in salary.
If year of service increased by 1 year salary increased by $521.9534 and if age is increases by 1 year then salary is increased by 72.63425
Significance of Coefficient:
Here we test the significance of coefficient, it is two sided hypothesis. Above table show the P-value of significance test. IF P-Value < 0.05, we claim that variable is significant otherwise not.
So,
Business School, Liberal school, Assistant professor and Professor are significant variables.
6).
From correlation analysis, salary is significantly related with age but in step 5, we came across the conclusion that age is not significant factor for salary.
Summary Report:
We have data of 199 academics from the particular college. We noted the school in which they work, their rank, gender, age, year of service and salary. There are 66 female and 133 male academics in the data. We observed that the mean salary of male academics is more than female academics. We also noted that there is more variation in the male salary than female salary.
We used two sample t-test for comparing mean salary of male and female. We observed that male academics have more pay than female academics. We also compare mean salary of male and female for assistant professor, associate professor and professor. We observed that there is no significant difference between male assistant professor and female assistant professor whereas male associate professor and professor earns more salary than female associate professor and professor.
From the correlation analysis of salary, age and years of service. We observed that there is positive and significant relationship between this variables.
We used stepwise regression to the salary amount we add one by one variable gender, school, rank, years of service and age. We used female, health and associate professor as a reference variable for nominal variables. We observed that R2 is increased from 0.17 to 0.7. In step 1, we used gender as predictor variable and it is found to be significant. In step 2, we add school variable, in third we and rank and so on. The regression equation for the step 5 is as follows:
Salary = 81256.74 + 3538.509 × (Male) + 26979.21 × (Business) – 8059.37 × (Liberal Studies) – 6197.97 × (Science) – 10153.4 × (Assistant Professor) + 23431.59 × (Professor) + 521.9534 × Years of Service + 72.63425 × Age
where
(Male) = 1 is respondent is male other wise 0.
(Business) = 1 if respondent from business school otherwise 0
(Liberal Studies) = 1 if respondent from Liberal Studies school otherwise 0
(Science) = 1 if respondent from Science school otherwise 0
(Assistant Professor) = 1 if respondent is assistant professor otherwise 0
(Professor) = 1 if respondent is professor otherwise 0
In step 5, we found that school and rank are only the significant variables for predicting the salary of the employee. Gender, year of service and age are not significant factor for predicting the salary of the employee.
