SOLUTION TO ASSIGNMENT ON QUANTITATIVE METHODS

QUESTION1

HYPOTHESIS;
HO; μA =μB = μA there are significant differences in the mean sales calls made across the three
branches.
H1; μA ≠μB ≠ μA there are no significant differences in the mean sales calls made across the three
branches.

LEVEL OF SIGIFICANCE;
α = 5% =0.05

TEST STATISTIC;
Because we are comparing means of more than two groups, use the F statistic

Where, F = MST
MSE
DECISION RULE;
Reject H0 if Fcalculated > Ftabulated
Fcalculated > Fa,k-1,n-k
Fcalculated > F0.05,(3-1),(18-3)
Fcalculated > F0.05,(2,15)
Fcalculated > 3.68

COMPUTATIONAL PROCEDURE FOR TABULATED;

Branch
Sales Person Branch A Branch B Branch C total
1 36 41 53 130 Where XG =
2 45 47 45 137 826/18 =
3 51 38 57 146 45.88889
4 48 39 61 148
5 45 52 51 148
(X 6 41 37 39 117 –XG)
Column total 266 254 306 826
-9.88889
n 6 4.8888896 -7.11111
6 18
0.888889
mean 44.3 -1.11111 42.3 0.888889
51 45.9
-5.11111 7.888889 -11.1111
-2.11111 6.888889 -15.1111
0.888889 -6.11111 -5.11111
4.888889 8.888889 6.888889

(X –XG)2

97.79012 23.90123 50.5679

0.790123 1.234568 0.790123
26.12346 62.23457 123.4568
4.45679 47.45679 228.3457
0.790123 37.34568 26.12346
23.90123 79.01235 47.45679

Therefore, sum of square total = ∑(X –XG)2 = 881.78

(X –XC)

-8.33333 -1.33333 2
0.666667 4.666667 -6
6.666667 -4.33333 6
3.666667 -3.33333 10
0.666667 9.666667 0
-3.33333 -5.33333 -12

(X – XC)2

69.44444 1.777778 4
0.444444 21.77778 36
44.44444 18.77778 36
13.44444 11.11111 100
0.444444 93.44444 0
11.11111 28.44444 144

Therefore, sum of square error = SSE = ∑(X –XC)2 = 634.67

Sum of square treatment = SST = SS total – SSE = 881.78 – 634.67 = 247.11

ANOVA TABLE
Source of variation Sum of square Degree of freedom Mean square F
Treatment 247.11 3-1 = 2 123.556 2.92
Error 634.67 18-3= 15 42.311
Total 881.78 18-1 = 17

The computed value of F is 2.92, which is less than the critical value of 3.68, so the null hypothesis is
accepted.
Conclusion: The population means are all equal. The mean sales calls are the same for the three
branches.

QUESTION 2
Subject to the number of observation involved the correlation coefficients will be computed using a
statistical package called Mega stat for excel.
The summary from the statistical package is shown table below;
 Retention rank Post-secondary Post-secondary Academic
participation rank achievement rank achievement rank
Retention rank 1.000      
Post-secondary .349 1.000    
participation rank
Post-secondary .118 .074 1.000  
achievement rank
Academic .073 .033 .604 1.000
achievement rank

70 sample size

Summary of the table above based on the questions;

(1) The Spearman’s rank correlation for Retention rank and Post-Secondary Participation rank is 0.349.
(i.e r = 0.349). This shows a weak positive correlation between retention rank and post-secondary
participation.

(2)The Spearman’s rank correlation for Post-Secondary Achievement rank and Academic Achievement
rank on exams is 0.604. (i.e r = 0.604). This shows a strong positive correlation between post
secondary achievement and academic achievement.

(3)The Spearman’s rank correlation for Post-Secondary Participation rank and Post-Secondary
Achievement rank is 0.074. (i.e r = 0.074). This shows a weak positive correlation between post
secondary participation and post secondary achievement.

(4)Comment on your results.

The correlation between the variables under consideration shows that there exit a positive association
between:
(1) Retention rank and Post-Secondary Participation rank is 0.349.
(i.e r = 0.349). This shows a weak positive correlation between retention rank and post-secondary
participation.
(2) Post-Secondary Achievement rank and Academic Achievement rank on exams is 0.604. (i.e r =
0.604). This shows a strong positive correlation between post secondary achievement and academic
achievement.
(3) Post-Secondary Participation rank and Post-Secondary Achievement rank is 0.074. (i.e r = 0.074).
This shows a weak positive correlation between post secondary participation and post secondary
achievement.

QUESTION 3

Individual Age(X) Amount X2 Y2 XY

spent(Y)
1 25 68 625 4624 1700
2 47 46 2209 2116 2162
3 58 31 3364 961 1798
4 37 56 1369 3136 2072
5 59 35 3481 1225 2065
6 21 81 441 6561 1701
7 55 45 3025 2025 2475
8 43 38 1849 1444 1634
9 47 42 2209 1764 1974
10 24 68 576 4624 1632
11 61 28 3721 784 1708
12 55 18 3025 324 990
13 29 65 841 4225 1885
14 50 31 2500 961 1550
15 39 55 1521 3025 2145
650 707 30756 37799 27491

(1)Pearson’s correlation coefficient between age and amount spent on this particular product.

r= n∑XY -∑X∑Y = 15(27491) – (650Χ707) = (412365 – 459550) =

√ (n∑X2 – (∑X)2)(n(∑Y2 –(∑Y)2 √(15(30756) – 6502)(15(37799) - 7072) √(38840Χ67136)

r = -47185 = -0.92
51064.3
There exist a strong negative correlation between age and amount spent on this particular product.

(2) Test this correlation for statistical significance (at a 5% significance level).

Hypothesis;
H0: r = 0 (the correlation in the population is 0)
H1: r ≠ 0 (the correlation in the population is not 0)

LEVEL OF SIGNIFICANCE;
α = 5% = 0.05.

Decision rule;
Reject H0 if:
tcalculated > ttabulated.
tcalculated > ta/2,n-2
tcalculated > t0.025,15-2
t calculated> 2.16

TEST STATISTIC;

t = r√(n-2) = -0.92√(15-2) = -3.32 = -8.4694

√(1- r2) √(1-(-0.92)2 0.392

|t| = 8.4694

The computed t (8.4694) is within the rejection region; therefore, we will reject H0. This means the
correlation in the population is not zero. From a practical standpoint, it indicates that there is correlation
between age and amount spent on this product.
(3) Estimate a linear regression to examine the relationship between age and amount spent on this
product.

The linear regression equation is given as:

Y = a + bx
Where: b = n∑xy - ∑x∑y = 15(27491) – (650Χ707) = (412365 – 459550) = -47185 = -1.2149
n∑x2 – (∑x)2 (15(30756) – 6502) 38840 38840

and a = y – bx = ∑y/n – b(∑x/n) = 707/15 –(-1.2149Χ650/15) = 99.7771

Therefore, the estimate of the linear regression trend line is given as;
y = 99.7771 – 1.2149x

QUESTION 5

t y t2 ty Trend value
1 48 1 48 64.978
2 51 4 102 63.381
3 52 9 156 61.784
4 54 16 216 60.187
5 68 25 340 58.590
6 70 36 420 56.993
7 71 49 497 55.396
8 75 64 600 53.799
9 56 81 504 52.201
10 57 100 570 50.604
11 59 121 649 49.007
12 61 144 732 47.410
13 31 169 403 45.813
14 30 196 420 44.216
15 32 225 480 42.619
16 33 256 528 41.022
136 848 1496 6665

The linear regression equation is given as:

Y = a + bt

Where: b = n∑ty - ∑t∑y = 16(6665) – (136Χ848) = (106640 – 115328) = -8688 = -1.597

n∑t2 – (∑t)2 (16(1496) – 1362) 5440 5440

and a = y – b t = ∑y/n – b(∑t/n) = 848/16 –(-1.597Χ136/16) = 66.575

Therefore, the estimate of the linear regression is given as;
y = 66.575 – 1.597t

(2)Based on these results, would you describe turnover as increasing or decreasing?

The turnover is decreasing (downward) as shown in the trend values obtained.

(3) Using this data and incorporating seasonal effects, forecast quarterly turnover for 2010.
YEAR Q1 Q2 Q2 Q4

2006 0.738712 0.804658 0.851642 0.897204

2007 1.160608 1.228221 1.281681 1.394078

2008 1.072776 1.126393 1.20391 1.286648

2009 0.676664 0.678487 0.750839 0.804446

TOTAL
3.648759 3.837759 4.088072 4.382376
AVERAGES
0.91219 0.95944 1.022018 1.095594 3.989242
ADJUSTED 0.9146 0.9620 1.0248 1.0986 4.0000
INDEX

Correction factor for adjusted quarterly means;

Correction factor = 4.00 = 4.00 =1.0027
Total of four means 3.989242
Forecast of quarterly turnover for 2010 are;

t Trend=66.575 –1.597t Forecast=Y=seasonal index Χ Trend

17 39.4 0.9146Χ39.4 = 36.03524
(4) 18 37.8 0.9620Χ37.8 = 36.3636 Relative
impact of 19 36.2 1.0248Χ36.2 = 37.09776 the trend
and 20 34.6 1.0986Χ34.6 = 38.01156 seasonal
effects on this
forecast:
(a)The trend helps in determining the long pattern of the series. It reveals if the series has an upward or
downward pattern.
(b)Understanding seasonal fluctuations help plan for sufficient goods and materials on hand to meet
varying seasonal demand.
(c)Analysis of seasonal fluctuations over a period of years helps in evaluating current sales and it
enables forecasting for a future period.

QUESTION 6
Subject to the number of observation involved the correlation coefficients and regression will be
computed using a statistical package called Mega stat for excel.
The summary from the statistical package is shown table below

(1) Pearson’s correlation coefficient using hours lost absence and overtime payments.

Overtime
hours
absent
23.64 36.55
140.4 22.02
837 102.86
464.16 145.49
145.8 94.65
1406.16 7.73
1118.76 22.5
349.8 167.52
381.6 94.89
147.6 66.79
1673.4 7.06
166.08 7.56
70.2 65.84
298.44 254.18
294 52.02
231 160.98
945 43.1
81.6 95.124
52.56 73.1
41.28 15.03
368.4 37.53
690.36 0
1113.84 7.59
178.2 205.25
397.8 182.76
59.04 22.03
102 15.83
165.84 58.81
1897.56 65.18
2967.36 43.17
99.48 161.25
1289.64 131.91
1323.48 6.88
1518.48 115.73
361.8 50.84
20.4 58.82
530.4 81.32
56.4 132.68
1185.84 145.25
78.96 81.14
165 15.28
42.96 123.55
128.64 36.39
186 146.97
78.6 58.81
608.04 80.18
607.08 59.88
172.56 52.26
270 44.24
283.56 37.58
1500.84 0
248.4 81.51
 610.56 58.13
230.04 14.56
57.6 132.64
834.36 29.52
605.16 0
412.56 22.69
371.28 37.75
2177.52 0

Correlation Matrix

Overtime hours absent

Overtime 1.000  
hours absent -.228 1.000

60 sample size
The above table is an extract from the summary of the outcome of the statistical package (megastat).
It can be deduced that there exist a weak negative correlation between overtime payments and hours
lost absent.
Test for significance of the correlation coefficient;

Hypothesis;
H0: r = 0 (the correlation in the population is 0)
H1: r ≠ 0 (the correlation in the population is not 0)

LEVEL OF SIGNIFICANCE;
Α = 5% = 0.05.

Decision rule;
Reject H0 if:
tcalculated > ttabulated.
tcalculated > ta/2,n-2
tcalculated > t0.025,60-2
t calculated> 2.00

TEST STATISTIC;

t = r√(n-2) = -0.228√(60-2) = -1.7363 = - 1.7832

√(1- r2) √(1-(-0.228)2 0.9737

|t| = 1.7832

The computed t (1.7832) is outside the rejection region; therefore, we have no evidence against H0.
This means the correlation in the population is not significant. From a practical standpoint, it indicates
that there is no correlation between overtime payments and hours lost absent.

(2)Multiple regression predicting hours lost due to absence using mega stat for excel;
Regression Analysis

R² 0.368
Adjusted R² 0.269 n 60
R 0.607 k 8
Std. Error 49.733 Dep. Var. hours absent

ANOVA
table
Source SS df MS F p-value
Regression 73,408.8644 8 9,176.1080 3.71 .0018

Residual 126,139.5996 51 2,473.3255

Total 199,548.4640 59      

Regression output confidence interval

std. p- 95% 95%
variables coefficients error t (df=51) value lower upper
-
Intercept -55.6845 74.1495 -0.751 .4561 204.5460 93.1769
Rate -14.4599 8.0017 -1.807 .0766 -30.5240 1.6042
Overtime -0.0225 0.0120 -1.877 .0662 -0.0466 0.0016
grade 20.2851 13.0631 1.553 .1266 -5.9401 46.5103
Bank quals. -0.6962 17.9625 -0.039 .9692 -36.7574 35.3649
Age 3.9011 1.6273 2.397 .0202 0.6343 7.1680
Tenure -3.7204 1.7250 -2.157 .0358 -7.1836 -0.2572
Sex 57.0187 16.6656 3.421 .0012 23.5610 90.4764
Married -29.7375 16.9210 -1.757 .0848 -63.7078 4.2327

From the above summary table the multiple regression predicting hours absent is presented below;
Y = a0 +b1x1 +b2x2 + b3x3 +b4x4 + b5x5 + b6x6 + b7x7 + b8x8
Y = -55.6845-14.4599X1-0.0225X2+20.2851X3-0.6962X4+3.9011X5-3.7204X6+57.0187X7-29.7375X8
(3) Describe these results:
For question (1),
 It can be deduced that there exist a weak negative correlation between overtime
payments and hours lost absent.
 The computed t (1.7832) is outside the rejection region; therefore, we have no evidence
against H0. This means the correlation in the population is not significant. From a
practical standpoint, it indicates that there is no correlation between overtime payments
and hours lost absent.
For question (2),
 Rate is negatively related to hours absent.
 Overtime is negatively related to hours absent.
 Grade is positively related to hours absent.
 Bank qualification is negatively related to hours absent.
 Age is positively related to hours absent.
 Tenure is negatively related to hours absent.
 Sex is positively related to hours absent.
 Married is negatively related to hours absent.

(4)If an individual were to be promoted by one grade, by how much would you expect
absenteeism to change?
 For an individual promoted by one grade, we expect the hours absent to increase by
approximately 21.
(5) Multiple regression predicting overtime payments using mega stat for excel:

Regression Analysis

R² 0.279
Adjusted R² 0.166 n 60
R 0.529 k 8
Std. Error 561.959 Dep. Var. Overtime

ANOVA table
Source SS df MS F p-value
Regression 6,247,065.1512 8 780,883.1439 2.47 .0239
Residual 16,105,693.1638 51 315,797.9052
Total 22,352,758.3150 59      

Regression output confidence interval

p-
variables coefficients std. error t (df=51) value 95% lower 95% upper
Intercept -1,461.4162
Rate 52.1041 92.9804 0.560 .5777 -134.5618 238.7700
grade 443.2581 137.7160 3.219 .0022 166.7817 719.7346
Bank quals. -396.5531 195.2288 -2.031 .0475 -788.4913 -4.6150
Age 15.0009 19.2819 0.778 .4402 -23.7090 53.7109
Tenure -26.5093 20.0206 -1.324 .1914 -66.7024 13.6838
Sex -46.0477 208.7117 -0.221 .8263 -465.0539 372.9585
Married -353.7644 190.5722 -1.856 .0692 -736.3541 28.8252
hours absent -2.8726 1.5303 -1.877 .0662 -5.9447 0.1996
From the above summary table the multiple regression predicting overtime is presented below;
Y = a0 +b1x1 +b2x2 + b3x3 +b4x4 + b5x5 + b6x6 + b7x7 + b8x8

Y = -1461.162 + 52.1041X1 + 443.2581X2 - 396.5531X3 + 15.0009X4 - 26.5093X5 - 46.0477X6 - 353.7644X7 - 2.8726X8

(6) Describe these results.

 Rate is positively (directly) related to overtime.

 Grade is positively (directly) related to overtime.
 Bank qualification is negatively related to overtime.
 Age is positively (directly) related to overtime.
 Tenure is negatively (inversely) related to overtime.
 Sex is negatively (inversely) related to overtime.
 Married is negatively (inversely) related to overtime.
 Hours absent is negatively (inversely) related to overtime.

(7)If an individual were to be promoted by one grade, by how much would you expect overtime
to increase?
For an individual promoted by one grade, we expect the overtime to increase by approximately
$443.26.

(8)What advice can you give to the HR manager?

Considering the severity in the anticipated days of absence and outrageous anticipated
overtime that will be incurred, if a upper level clerk is promoted by one grade, it is advised that
the HR officer should abstain from promoting the upper level clerical employee to minimize the
possible occurrence of this predicted outcomes.