Scribd
Upload
3 views3 pages

Analyzing Effort in ITIL Processes

The document summarizes a statistical analysis of effort data from an IT organization. The response variable is the logarithm of effort. There are three explanatory variables: complexity (with three levels), ITIL process (with three levels), and service domain type (with two levels). Estimates were calculated for each level of the explanatory variables based on the average response value. For example, complexity had estimates of 0.718, 0.29, and -0.191. Finally, maximum likelihood estimates found that the error term follows a normal distribution with a mean very close to 0 and a standard deviation of 0.90, validating the assumption of normality.

Uploaded by

Arjun Seerams
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
3 views3 pages

Analyzing Effort in ITIL Processes

The document summarizes a statistical analysis of effort data from an IT organization. The response variable is the logarithm of effort. There are three explanatory variables: complexity (with three levels), ITIL process (with three levels), and service domain type (with two levels). Estimates were calculated for each level of the explanatory variables based on the average response value. For example, complexity had estimates of 0.718, 0.29, and -0.191. Finally, maximum likelihood estimates found that the error term follows a normal distribution with a mean very close to 0 and a standard deviation of 0.90, validating the assumption of normality.

Uploaded by

Arjun Seerams
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Question:2

Response variable: Effort

Explanatory variables: i) Complexity (3 levels-Ordinal) ii) ITIL Process (3 levels) iii) Service Domain
Type (2 levels)

Data generation process:

Y=µ+β +Ƭ +ϒ +ɛ
i j k
Here, µ is the unconditional mean,

β,Ƭ,ϒ corresponds to complexity, ITIL process & service domain type respectively.

The error term ɛ follows normal distribution~ N(0,σ2).

Since data generation process assumes, Property of Additivity and Homoscedasticity, lets look at the
distribution of response variable (using histogram)

As we can see, it is highly right skewed with a skewness of ~2.62

In order for our assumptions to be good enough, we can reduce the skewness by transforming the
response variable.

New response variable: log(Effort)


As we can see the skewness reduced drastically to ~0.06 which is near normal.

Estimating the equation:

i) Equation when none of the explanatory variables are considered:


Y =Log(Effort) = µ + ɛ
The unbiased estimate for µ is the sample mean = 3.991
Y = 3.991 + ɛ
ii) Equation when Complexity is taken as an Explanatory variable:
Y =Log(Effort) = µ + βi +ɛ

β1 = when complexity is complex


β2 = when complexity is medium
β3 = when complexity is simple

Complexity Average of Y
Complex 4.708984712
Medium 4.281909931
Simple 3.800847456
Grand Total 3.990564428

Best estimate of (µ+ β1) = 4.709


Hence best estimate for β1 = 4.709– 3.991 = 0.718
Similarly, β2 = 0.29 & β3 = -0.191

iii) Equation when Service Domain type is taken as an Explanatory variable:


Y =Log(Effort) = µ + Ƭj+ ɛ

Ƭ1 = Service Domain Type is App support


Ƭ2 = Service Domain Type is Database Support

Service Domain Type Average of Y


App support 4.146598593
Database Support 3.899544498
Grand Total 3.990564428

Best estimate of (µ+ Ƭ1) = 4.147


Hence best estimate for Ƭ1 = 4.147– 3.991 = 0.156
Similarly, Ƭ2 = -0.101

iv) Equation when ITIL Process is taken as an Explanatory variable:


Y =Log(Effort) = µ +ϒk+ ɛ

ϒ1 = ITIL Process is Incident Management


ϒ2 = ITIL Process Type is Release Management
ϒ3 = ITIL Process Type is Service Request
ITIL Process Average of Y
Incident Management 3.830001681
Release Management 4.350334249
Service Request 4.103310331
Grand Total 3.990564428

Best estimate of (µ+ ϒ1) = 3.83


Hence best estimate for ϒ1 = 3.83– 3.991 = -0.161
Similarly, ϒ2 = 0.359 & ϒ3 = 0.11

So by additivity property,

log(Effort) = 3.991 + β + Ƭ + ϒ + ɛ
i j k
Impacting
variables Estimates
β1 0.718
β2 0.29
β3 -0.191
T1 0.156
T2 -0.101
ϒ1 -0.161
ϒ2 0.359
ϒ3 0.11

Error term distribution,

Maximum likelihood estimates if we assume a normal distribution are as follows,


mean sd
-0.000754922 0.902375401
( 0.011227188) ( 0.007938821)
Which is makes our assumption of N(0,σ2) valid

You might also like