The Influence of Employee Status On Salary Prediction for IT Firms: Case of Grant Technologies, Inc.

Posted: August 27th, 2021

Student’s Name

Instructor’s Name

Course

Date

The Influence of Employee Status On Salary Prediction for IT Firms: Case of Grant Technologies, Inc.

Introduction/Problem Statement

Proper recruitment of employees is critical in advancing a business strategy. Currently, recruitment processes are drastically changing and becoming highly complex tasks that require intensive interviews and evaluation. Salary is an important aspect when it comes to employment(Frost, n.d). It determines whether a given firm can attract qualified employees. Thus, it influences the quality of employees that a firm would higher.

Among software engineering firms, challenges of predicting employee salary are prominent. A prediction refers to an assumption regarding the future based on existing knowledge and experience(Frost, n.d). It is essential in helping firms plan their future. Thus, this paper aims at predicting the salary of employees for software engineering firms based on different factors.  Some of these factors include test scores, age, years of experience, and gender. 

Methodology

Collection of Data

The regression analysis in this paper seeks to understand if years of experience, test score, age of the employee, and gender can predict employee salary among software engineering firms. In an attempt to implement the project, data was collected from a sample of 30 web developers from Grant Technologies, Inc. An aptitude test was utilized in collected the test scores for each web developer. Thus, this study seeks to conduct a multiple regression to assess the validity of the model and statistical significance of independent variables (test score, years of experience, age, gender) in influencing the dependent variable’s behavior (salary).  The study defines employees in terms of age, job experience, test scores, and gender.

Regression Model

The following regression model was used in achieving the objectives of the paper;

(i)

Therefore, the above regression model was used to assess how age, age, years of experience, and test score affect employee salary for software engineering firms. Gender was denoted by 1 (if male), 2 (if female) and 0 (undefined).

Results

The following shows the results of regression analysis displayed in the regression table;

Table 1: Regression Analysis

	Coefficients	Standard Error	t Stat	P-value
Intercept	334.75	104.86	3.19	0.00
Gender	9.35	23.66	0.40	0.70
Age (years)	1.68	2.48	0.68	0.51
Test Score	-0.45	0.96	-0.47	0.64
Years of Experience	12.37	4.48	2.76	0.01

As shown in Table 1, the intercept score is 334.75 with a p-value = 0.00. The gender, age (years), test-score, and years of experience coefficients scores 9.35, 1.68, -0.45, and 12.37. As such, this yields the following regression model;

(ii)

p-value = 0.01

The ANOVA F-test results are as displayed in the following table;

Table 2: ANOVA F-test Results

ANOVA
	df	SS	MS	F	Significance F
Regression	4	205043.2269	51260.8067	8.838	0.000136667
Residual	25	144993.0305	5799.72122
Total	29	350036.2574

p-value = 0.01

Interpretation

Regression analysis model (ANOVA F-test)

The regression analysis model (ii) shows that if all factors are held constant, web developers at Grant Technologies, Inc. will earn $334.75. Employee’s age increases employee salary by 9.35, gender by 1.68, and years of experience by 12.37. However, test scores have a negative influence on employee salary as it reduces it by 0.45 units. As illustrated by the R-squared value, these independent variables can predict 58.6% of all the employee salary changes. Thus, the rest is predicted by variables outside the model. The F–test results show that the calculated value (F-test Calculated) is 0.00014 against a p-value of 0.01 or 10% significance level. Since F-Calculated < F-Critical (3.49) or (F-calculated <3.49 at alpha = 0.01), it implies that we fail to reject the null hypothesis. Thus, this implies that the model is statistically significant at 10%. The independent variables can predict 90% of the changes in employee salary.

Implications

The study indicates that age, experience, gender, and test scores have a significant influence on employee salary. Test scores negatively affect employee salary while the rest of the variables have a positive influence. Thus, employers must look at these factors when evaluating the salary of their employees.

Short Comings

Although the model scores a high R-squared value of 58.6%, it leaves out a significant score of 41.4%. Thus, the current independent variables cannot thoroughly explain all the changes in the model. There are other factors outside the model that should be considered that the study did not incorporate. These may include inflation rate and economic stability, among others.

Works Cited

Frost, J. “Regression Tutorial with Analysis Examples.” Statistics by Jim, 13 June 2019, statisticsbyjim.com/regression/regression-tutorial-analysis-examples/.

Appendix

Definition of variables

Age– refers to the age of the employee

Test score– refers to aptitude tests obtained from employee responses

Years of experience– refers to time employee has worked within the same profession

Gender – the particular gender of employee (Male =1, Female = 2, Undefined = 0)

Salary – the amount the firm pays to an employee

Table 1. Employee Data

NO	Salary (US $)	Gender	Age (years)	Test Score	Years of Experience
1	300.7	1	25	76	4
2	330.1	1	18	67	2
3	480.0	1	30	80	4
4	500.0	1	40	10	5
5	500.0	2	43	80	10
6	300.0	1	41	76	10
7	454.7	2	23	76	9
8	469.8	2	28	76	8
9	485.0	1	28	73	9
10	500.1	1	29	70	9
11	515.2	1	29	80	10
12	530.4	0	36	100	12
13	545.5	1	38	100	8
14	560.7	1	38	76	14
15	575.8	1	38	73	14
16	591.0	0	56	85	17
17	606.1	2	56	80	19
18	621.2	1	45	82	16
19	636.4	2	45	81	18
20	651.5	2	49	83	17
21	666.7	2	48	82	19
22	700.0	2	41	86	15
23	450.0	2	38	90	15
24	350.0	1	38	93	12
25	400.0	1	18	93	1
26	332.2	1	21	79	1
27	480.0	1	33	84	3
28	415.1	2	33	88	4
29	419.3	2	33	82	4
30	423.6	2	36	82	6

Source: Grant Technologies, Inc. Database. www.granttechnologies.com

Final Regression and Significance Test

Summary Output

SUMMARY OUTPUT
Regression Statistics
Multiple R	0.765
R Square	0.586
Adjusted R Square	0.520
Standard Error	76.156
Observations	30
ANOVA
	df	SS	MS	F	Significance F
Regression	4	205043.2269	51260.8067	8.838495	0.000136667
Residual	25	144993.0305	5799.72122
Total	29	350036.2574
	Coefficients	Standard Error	t Stat	P-value	Lower 95%
Intercept	334.75	104.86	3.19	0.00	118.79
Gender	9.35	23.66	0.40	0.70	-39.37
Age (years)	1.68	2.48	0.68	0.51	-3.44
Test Score	-0.45	0.96	-0.47	0.64	-2.43
Years of Experience	12.37	4.48	2.76	0.01	3.15

Residual Output

RESIDUAL OUTPUT
Observation	Predicted Y	Residuals
1	401.187	-100.487
2	368.766	-38.666
3	407.771	72.229
4	468.573	31.427
5	513.161	-13.161
6	502.262	-202.262
7	469.028	-14.361
8	465.051	4.758
9	469.427	15.525
10	472.462	27.633
11	480.311	34.927
12	498.408	31.973
13	461.637	83.887
14	546.705	13.962
15	548.061	27.748
16	600.609	-9.657
17	646.309	-40.214
18	580.482	40.756
19	615.023	21.358
20	608.464	43.060
21	631.977	34.690
22	568.940	131.060
23	562.096	-112.096
24	514.280	-164.280
25	344.642	55.358
26	356.007	-23.762
27	398.629	81.371
28	418.540	-3.418
29	421.252	-1.906
30	451.028	-27.456

Residual Plots – Line Check