Correlation Analysis
v In the process of estimating and controlling costs, management must evaluate whether or not the factor selected for
estimating cost behavior is suitable for that purpose.
v Costs may or may not react with changes in the factor selected for cost analysis
Correlation
Ø a connection or relationship between two or more facts, numbers, etc.(Cambridge dictionary)
v The degree of correlation between the level of activity and costs may be measured by the “coefficient of determination”
(goodness of fit), most frequently designated as r².
q Coefficient – any one or more numbers or symbols placed before another number or symbol serving as its multipliers
Ex. 6a, 3b, abcd;
coefficients of a, b and d are 6, 3, and abc
q Goodness of fit - a statistical hypothesis test used to see how closely observed data mirrors expected data
v Formula:
r² = 1 - (Estimated conditional standard deviation measured from the line of regression) ²
(Standard deviation measured from the average of all data)²
q Standard deviation – statistics that measures the tendency of the data to be spread out
q Line of regression - predicts the value of y for a given value of x.
ü If r² expressed as a percentage will be high, the correlation is good; shows that parts of the variables are strongly related
v Formulas:
* (Estimated conditional standard deviation measured from the line of regression)²
√ ∑(Y – Ȳ )² / n-2
where Y = actual cost; Ȳ = line of regression cost; n= # of items
* (Standard deviation measured from the average of all data)²
√ ∑(Y– Ȳ )² / n
where Y = actual cost; Ȳ = average cost; n= # of items
Illustration: The estimated conditional standard deviation measured from the line of regression
X Y Ȳ * Y – Ȳ (Y – Ȳ )²
(Hours) (Actual Cost) (Conditional Expected Average Cost) (Deviations) (Deviations Squared)
20 P50 [(P2.04 x 20) + P28.60] = P69.40 [P50 – P69.40] = (19.40) [19.40 x 19.40] = 376.36
40 110 [(P2.04 x 40) + P28.60] = 110.20 [P110 – P110.20] = (0.20) [(0.20 x (0.20)] = 0.04
60 150 151.00 (1.00) 1.00
20 70 69.40 0.60 0.36
30 80 89.80 (9.80) 96.04
40 100 110.20 (10.20) 104.04
50 150 130.60 19.40 376.36
10 60 49.00 11.00 121.00
30 110 89.80 20.20 408.04
50 120 130.60 (10.60) 112.36
1,000 1,595.60
*Based from other methods, FxC = P28.60; VC = (P2.04)X
v Solutions:
Estimated conditional standard deviation measured from the line of regression
√ 1,595.60 / 10-2 = P14.12
1
Standard deviation measured from the average of all data
Y Ȳ Y – Ȳ (Y – Ȳ )²
(Actual Cost) (Average Cost)* (Deviations) (Deviations Squared)
P50 P100 (50) 2,500
110 100 10 100
150 100 50 2,500
70 100 (30) 900
80 100 (20) 400
100 100 0 0
150 100 50 2,500
60 100 (40) 1,600
110 100 10 100
120 100 20 400
* P1,000/10 11,000
v Solution:
Standard deviation measured from the average of all data
√ 11,000 / 10 = P33.17
Therefore,
r² = (Estimated conditional standard deviation measured from the line of regression) ²____________
(Standard deviation measured from the average of all data)²
= 1 – (P14.12)² = 1- 199.37__ = 1 – 0.1812 = 81.88%
(P33.17)² 1,100
Conclusion: There is apparently a very high degree of correlation between hours of operation and supplies costs.
