63

CHAPTER 3

DELPHI METHOD

3.1 INTRODUCTION

Research methodology is the systematic way to solve the research

problem. In this study, qualitative and quantitative aspects are discussed. The
qualitative research comprises of comprehensive literature study in which
content, objects and subject terms were identified for inclusion a retention
factor and the development of the questionnaire.

The discussions on quantitative aspects include a population to test

the efficiency of the questionnaire. High technology employees in an IT/ITES
sector organisation are used to determine whether the final questionnaire must
be adjusted or not. The top retention factors that induce organisational
commitment are identified. This is significant in any research issue because it
assures authority and authenticity of the research results.

3.2 JUSTIFICATION OF THREE PHASED DELPHI

APPROACH

This chapter explains the overall methodology used to collect the

data to address the research issues of this three phased study. The objective of
this study was to investigate and determine the current Human Resource
practices on the retention of Information Technology employees in Bangalore,
Karnataka state, Indian organizations. In order to achieve this aim, this study
was divided in three phases.
 64

3.2.1 The Delphi Technique

The Delphi technique is a group process used to survey and collect

the opinion of the experts on a particular subject. It was originally developed
in the year 1950s by the RAND Corporation in Santa Monica. Delphi is a
structured communication technique, originally developed as a systematic
interactive forecasting method which relies on the collective intelligence of a
panel of experts.

According to Linstone (1975), “Delphi is a method for structuring a

group communication process so that the process is effective in allowing a
group of individuals as a whole, to deal with a complex problem”.

Lynn Stuter (1996) has observed “The Delphi technique is based on

the Hegelian principle of achieving ‘Oneness of Mind’ through a three step
process of thesis, anti-thesis, all present their opinion or views on a given
subject, establishing views and opposing views.

The Delphi Technique is applied usually in three or more stages

seeking opinion of the experts, through questionnaire without revealing their
individual identity. The first round of questionnaire is designed to focus on
problems, opportunities, solutions or forecasts by seeking response of experts
to some open ended questions, in each subsequent stage, a questionnaire is
developed based on the results of the previous questionnaire. The process
stops when the research questions are analyzed and suggestion were given.

Skutsch & Hall (1973) identified the Delphi technique is a method

for gaining judgment on complex matters where precise information is
unavailable. Delphi technique has been widely used all over the world for
interactive decision-making, e-democracy as well as in the area of social
policy research and public health.
 65

3.2.2 Use of Delphi Study in the Current Research Work

The researcher had observed Information Technology being new

area it did not have enough grounding in India. The multinationals and global
organizations of today are notorious for their high attrition rates. Hence
retaining employees has become one of the chief tasks of modern Human
Resource Managers.

Human Resource Managers feel proud if they are successful at

retaining employees for over a year, to work for companies in which most of
the employees are over a year old.

As the first stage of doctoral research in this area, therefore the

researcher attempted to conduct some initial investigations in the form of
experience survey, literature review and pilot studies. Discussions with
knowledgeable individuals, who were willing to share their ideas and
experiences about important issues/ aspects of the research subject in the form
of an experience survey and small surveys conducted in the form of Pilot
studies in advance of a major investigation to identify issues of problem in the
organization of intended doctoral research study were considered highly
desirable.

Cooper & Schindler (2009) While publishing data and literature are
considered valuable for any research study, it is often recognized that only a
fraction of the existing knowledge in any field is put in to writing and hence
an experience survey along with a comprehensive literature review was
considered necessary in order to take a more complete view of the research
problem.

In second stage the researcher conducted a Delphi study to gather

the considered views of some known experts in the field of study. The output
 66

of first and second stages was then used for the design and documentation of
subsequent stages of the research study.

The Delphi study was considered suitable for this research work as
there was incomplete knowledge about the research problem. The aforesaid
method largely facilitated improving understanding of the research problem,
research opportunities and possible solutions to research questions.

The study was conducted in three rounds (DS1, DS2, DS3) seeking
opinion of the experts through questionnaire without revealing their
individual identity. The first round of Delphi questionnaire for this research
study consisted of ten open ended questions covering all aspects of research
problem. The objective of the first round questionnaire was to conduct a brain
storming exercise and generate as many ideas and opinion of the experts as
possible on every issue. The time allocated for the first round was 30 days.

All the experts who participated in the first round were also invited
to participate in the next two rounds. Additional experts were also chosen for
the second round. The responses that were mentioned by the experts in the
first round formed the basis of the questionnaire for the second round. The
Second round had sixteen Linkert Scale questions in carrier development
practice factors, Ten in Human Resource influencing factor and Ten in
Organisational factor. The questionnaires developed in each round were first
subjected to Pilot testing before being administered to the experts.

In the second round consensus was reached in Thirty six issues

presented in the questionnaire. All the participants of the second round were
invited to participate in the third and final round. A questionnaire, with
feedback on the second round results for each item, was presented to third –
round participants.
 67

In accordance with the Delphi methodology, experts were asked to

rerate their responses for objects on which consensus was reached at the end
of the second round. Objects on which consensus was reached in the second
round were excluded, resulting in a third round questionnaire of only three
major issues, finally the consensus was reached.

3.2.3 Experts Selection for the Delphi Study

Selecting appropriate experts is the most important step in the entire

Delphi process because it directly affects the quality of the result. Since the
Delphi techniques focuses on eliciting expert opinion over a short period of
time, the selection of Delphi experts is generally dependent upon the
disciplinary areas of expertise required by the specific issue.

For the current Delphi study twenty experts were carefully selected
based on their job knowledge, experience and wisdom. There were eight
professionals in the panel experts, were working in the Reputed Multinational
Companies in Bangalore. Twelve other members of the panel were CEO cum
HR Managers of various organizations.

3.2.4 Panel Selection

Dalkey & Helmer 1963: Delbecq & Ven (1971) the success of a
study is largely dependent on the quality of the participants reported specific
criteria for the selection of panel experts.

1. The first is to the expert’s exhibit a high degree of knowledge

of experience in the subject matter.

2. Another is that they exhibit ‘representativeness’ of the

profession, so that their suggestions may be adaptable or
transferable to the population.
 68

3. Other experts like, HR policy makers, HR consultants,

extensive theoretical knowledge, experience in the field of
Human Resource Management both in the industry and
academia and advanced degree in Management or a closely
related field.

The Potential participants for this study were identified through their
expertise in the areas of Human Resource Management in the Information
Technology. They included Academics, Human Resource Practitioners and
Industrial Psychologists.

The nomination of people, who would be appropriate ‘experts’ for

this study, was based on the following general criteria. Initially a personal
letter was sent to each of the nominees. The letter invited them to participate
in a three round Delphi study. In addition, the letter included an explanation
of the study and provided an estimate of the time commitment for
participation. In the introductory letter, nominees were informed that
participation was voluntary and confidential and the three rounds of responses
would be required.

Follow –up telephone calls were made and letters were sent to non –
respondents after two weeks. Nominees were advised that each round of the
study would require thirty minutes and that data collection would occur over a
two month time period.

3.2.4.1 Academics

1. Participants must have a minimum of ten years of teaching

experience in Management at Bangalore University.

2. Evidence of fairly extensive publications in Management.

3. Research interest in areas of Human Resource Management.

 69

3.2.4.2 Human resource practitioners

1. Currently working as an HR practitioner in an Information

Technology sector.

2. Minimum Five years working experience in Human Resource

Management.

3.2.4.3 Industrial psychologists

1. Minimum of Five years of experience as an Industrial

Psychologist.

2. Minimum three years of experience as a practitioner.

From the initial pool of nominations thirty respondents were

formally invited to participate and of these twenty agreed to complete the
required three rounds of the survey.

3.2.4.4 Panel size

Rowe & Wright (1999) as this study is a preliminary investigation,

the small number of participants was deemed by the researcher to be
acceptable for determining a meaningful outcome. The panel size of twenty
fits within the guidelines recommended for Delphi studies. Helmer & Dalkey
(1953) used a panel of seven experts in their original Delphi method
experiment. Helmer (1983) Linstone & Turoff (1975) suggests a panel size of
anywhere from ten to fifty participants.

The panel nominees were asked to express their expert opinions and
judgments on the current development of retention management in
Information Technology and to identify the key of HR factors influencing
retention in the workplace.
 70

3.2.4.5 Research methodology

For this research, the Delphi Technique was chosen as a suitable

preliminary research method because the results will offer a better – informed
look at the current and potential status of retention management carried out by
Information Technology sector, Based on the attitudes and beliefs of a
carefully selected group of expert respondents, the expected prospects for
reform in the areas of retention management and Human Resource practice
will also be captured.

A substantial literature review has identified some key of HR

practices that influence retention. However, there is a little consensus among
researchers with regard to precisely which HRM practice should be included.

Becker & Gerhart (1996); Cappelli (1999) due to these glaring

discrepancies in the prescriptions made by different scholars in this area, the
results of this Delphi study will be relevant and provide clarification as well
as substantially enhance the literature review on retention. In addition, it will
expose the current retention management practice adopted by IT
organizations. Previous studies using such as approach have typically used
thirty experts based on the finding that larger groups create few additional
ideas and limit the in- depth exploration of the ones generated.

Rowe & Wright 1999; van De Ven & Delbecq (1974) However for
this study, a small sample size was deemed acceptable due to the preliminary
exploratory role of the Delphi technique in the first stage of the research. It
was, however critical to secure the participation of the right kinds of experts,
who understand the issues, have a vision and represent a substantial variety of
viewpoints.
 71

In the second round, the responses suggested in the first round were
presented to each respondent in the form of survey statements and
accompanying response selections, each selection serving to complete the
initial statement. The respondents were asked to indicate the degree to which
they agreed with each completed statement on 1 to 5 with 1 indicating
strongly disagrees and 5 strongly agree (Linkert scale).

The responses that received the greatest support for each of the
questions were fed back to the experts during the round three. In the third and
final round, the respondents were asked to rank the responses that
accompanied each statement according to their perceived importance with 1
being strongly disagree and 5 strongly agree. This was done to help the
respondents further refine their opinions and assist in achieving consensus.

This preliminary questionnaire was generated from a review of

retention management literature. Cappelli 2000; Huselid 1995; Kraut &
Korman 1999; Lepakand snell 1999) subsequent questionnaires (Round 2 and
3 ) were modified and focused, based upon responses to the first questionnaire
was used to priorities research areas and to rank specific research issues
within these areas.

3.2.4.6 Delphi data collection and analysis

Data collection analysis were based on Schmidt’s method (1997), in

which the Delphi survey process is divided in to three rounds
The preliminary instrument (Round 1) was developed following an extensive
review of the literature of the relationship between retention and Human
Resource Policies. The questionnaire was designed and formulated based on
the four key research questions.
 72

The questionnaire consisted of two sections (A and B). In section A,

the style of questions chosen was open –ended in order to enable to
respondents to express their responses in any way they choose. This open
ended style was considered suitable as it provided minimum direction to
respondents. In section B of the instrument, participants were asked to
provide demographic data (gender, age, experience and professional status).

This type of question removes the need for the researcher to pre –
judge appropriate categories for response, allowing groupings of similar
responses to be constructed if necessary after the data have been collected.

Two major changes were made to the (Round 2) instrument. First,

the general demographics and open-ended questions were omitted. Second
responses contributed by the participants during Round I were imported
verbatim under the appropriate categories on the instrument. In round 2,
Participants were instructed to review the comments from round 1 and to rate
each comment (1-5), bearing it in mind that the researcher sought validation
based on the objects most critical to the retention process.

In the third round and final round survey, the experts were asked to
strongly disagree or strongly agree with the final wording of an item as well
as provide additional comment under the specified concept areas. This
procedure stopped a three questionnaires or round which seems fairly typical
of many studies. Consensus or trend towards consensus was documented at
the conclusion of Round 3.

The content analysis Denzin & Lincoln (1998) was conducted to

examine panelists’ responses to open – ended questions regarding HR factors
that influence retention; HR factors summarized panelists’ ratings of objects.
 73

Wright, Lawrence & Collopy (1996) means and standard deviations

were obtained for each item and each category following Round 1 and 2.

During the last decade, employee retention has become a serious

and perplexing problem for all types of organization. Managing retention and
keeping the turnover rate below target and industry norms is, one of the most
challenging issues facing business. From all indications, the issue will
compound in the future, even as economic condition changes. Employee
retention will continue to be an important issue for most job groups in the first
decade of the 21st century.

3.3 RESEARCH DESIGN

The present study is designed as a descriptive research design based

on primary data and secondary data. Descriptive research includes surveys
and facts finding enquiries of different kinds. The major purpose here is the
description of the state of affairs as it exists at present. In social science and
business research the term ex post- facto research is used for descriptive
research studies. This includes attempts by researchers to discover causes
even when they cannot control the variables. The methods utilized in this
research are survey methods of all kinds, including comparative and
correlation methods.

3.4 PILOT STUDY

Dane (1990) has stated that a Pilot study is “an abbreviated version
of research project in which the researcher practices or tests procedures to be
used in the subsequent full scale project”. Since the measures of the research
are either new or reconfigured from their original sources, a pilot study would
ensure psychometric cleaning of the objects, So that only appropriate objects
chosen through proper analysis would be used. A systematic Pilot study was
 74

carried out on 50 samples who didn’t participate in the final research. The
research instrument tested to ascertain the reliability and validity of the
instrument used. Recommendations found to be valid were incorporated into
the survey design to the actual research.

A pilot study has been conducted by choosing 50 Experts

(employees) from different levels of 15 IT Industries in Bangalore, Karnataka
State. The above study equipped the researcher with a strong base in the
development of the questionnaire. It also helped in proposing the statements
(attrition factors) which affect the employee retention in identifying the right
respondents from different IT sectors and in data analysis. Reliability analysis
has been done by taking a sample of 50 respondents. In all, the reliability of
the three statistics namely, Spearman-Brown, Guttmann and Cronbach’s alpha
showed that the reliability of scale constructed for the general assessment is
between 0.70 and 0.87, which makes the constructed scale fairly reliable.

3.5 THE QUESTIONNAIRE

According to Collis & Hussey (2003), questionnaires are associated

with both positivistic and phenomenological methodologies. They (Collis and
Hussey) describe a questionnaire as a list of carefully structured questions,
chosen after considerable testing, with a view to eliciting reliable responses
from a chosen sample. The aim is to find out what a selected group of
participants do, think or feel.

3.5.1 Testing the Questionnaire

Collis & Hussey (2003) Prior to distribution, the questionnaire was

tested on thirty respondents in a Pilot study. A Pilot study is an essential
element of testing the questionnaire prior to distribution. The questionnaire
was submitted to these respondents via e-mail. Suggest that the researcher can
 75

have colleagues or friends read through the questionnaire and play the role of
respondents, even if they know little about the subject.

3.5.2 Distribution

Distribution of the questionnaire was via electronic mail to 528

respondents. Although the research problem appears to cover a large
population, Bangalore Information Technology Industries in Karnataka, the
indication of the extent of the research in terms of focus is stated as within the
Departments. Therefore the results obtained will be for the benefit of all the
departments. The target audience for the purpose of the questionnaire in the
study is all personnel from all the level, who are working at the Department of
IT/ITES sectors of Bangalore. This constitutes of 500 potential respondents.

3.5.3 Validity and Reliability

Validity is concerned with the extent to which the research findings

represent what is happening in the situation (whether it is a true picture of
what is being studied). Leedy & Ormrod (2005) define the validity of a
measurement instrument as the extent to which the instrument measures what
it is supposed to measure.

According to Gill and Johnson (2002), there are three types of

validity:

a. Internal validity is the extent to which the conclusions

regarding cause.

b. Population validity is the extent to which conclusions might

be generalized to other people.
 76

c. Ecological validity is the extent to which conclusions might be

generalized to social contexts other than those in which data
have been collected.

Collis & Hussey (2003) purport that; reliability is concerned with

the findings of the research. The findings can be said to be reliable if the
researcher or anyone else repeated the research and obtained the same results.

3.6 DATA COLLECTION

A structured questionnaire is the main tool used for collecting

quantitative primary data. It enables quantitative data to be collected in a
standardized way so that the data are internally consistent and coherent for
analysis. Also the questionnaire ensures standardization and comparability of
the data across interviewers, increases speed and accuracy of recording, and
facilitates data processing.

The scaling techniques used in the development of questionnaire

used in the study are: nominal scale, Likert type scale under interval scale,
itemized rating and rank-order scale under ordinal scale and word association
under disguised structured scale.

Renis Likert type 5 point scale was used to convert qualitative

nature of the data into quantitative type. The respondents are asked to indicate
their degree of agreement by checking one of the five response categories.

The data are typically treated as interval scale. When using this
approach to determine the total score for each respondent on each store, it is
important to use a consistent scoring procedure so that a high (or low) score
consistently reflects a favourable response. This requires that the categories
assigned to the negative statements by the respondents be scored by reversing
 77

the scale. Note that for a negative statement, an agreement reflects

unfavorable response whereas for a positive statement, agreement represents a
favourable response. Accordingly, a strongly agree response to a favourable
statement and a strongly disagree response to an unfavorable statement would
both receive scores of five. Each respondent’s total score for each store is
calculated.

Rajiv Grover Marco Vriens (2006) a respondent will have the most
favourable attitude towards the store with the highest score. It is easy to
construct and administer this scale, and it is easy for the respondent to
understand.

3.6.1 Collection of Primary Data

The primary data required for the study have been collected from
selected employees working in IT organizations located in Bangalore city,
Karnataka State. The primary data collection is done in two stages. In the first
stage, a well structured questionnaire has been developed and pre-testing of
the questionnaire has been done by choosing 50 employees (respondents)
from different levels of 25 IT organizations on a random basis from
Bangalore city, Karnataka State.

Reliability analysis for the questionnaire used in the present study

has been done with a sample of 50 respondents and found that the reliability is
good. A final questionnaire (attached in the appendix-1) with 52 statements
(Human resource factor influencing retention of employees factors in IT
sector) has been prepared and was administered to the target respondents.

The designations of the employees covered under the study are:

Process Analyst (entry level), Senior Process Analyst, Team Leader,
Supervisor and Manager (Division Head). In the second stage, primary data
 78

has been collected by conducting an expert opinion survey using interview

schedule developed separately for the samples selected from Bangalore city,
Karnataka State. The respondents chosen includes experts and higher officials
of the department of Information and Technology and Industry Experts
involving senior HR Managers, Division Heads and HR specialists in various
IT companies. Also, personal discussions and deliberations with IT
employees starting from Process Analyst to Senior Managers’ level have been
conducted to collect the required data.

3.6.2 Secondary Data Collection

The secondary data related to the study are collected from different
sources including text books, articles published in journals, news papers,
periodicals National Association of Software and Service Companies
(NASSCOM) websites, Mckinsey study reports company websites,
Government’s IT department sites, doctoral research thesis and various other
related sites.

3.7 CONTENT VALIDITY AND RELIABILITY

Content validity is the degree to which elements of an assessment

instrument are relevant to and representative of the targeted construct for a
particular assessment purpose in psychological assessment. The importance of
content validity for the validation of the target construct varies depending on
how precisely the construct is defined and the degree to which “experts” agree
about the domain and facets of the construct.

Lawshe (1975) Content validation is a multi-method, quantitative

and qualitative process that is applicable to all elements of the assessment
instrument. During initial instrument purpose of content validation is to
minimize and to increase the probability of obtaining supportive construct
 79

validity indices in later studies. Because sources of error vary with the
targeted construct, the method of assessment, and the function of assessment,
the methods of content validation will also vary across these dimensions. In
this research, data was analysed using content validity technique.

Crocker et al (1986) the optimal number of judges will vary with the
element under consideration, the internal consistency of the ratings, and
practical considerations. In the present research, twenty panelists were asked
to indicate whether or not the measurement item was “essential” to the
operationalisation of the theoretical context. The panelist’s inputs were then
used to compute the CVR for each I Th candidate item in the questionnaire
(CVRi) as follows Equation (3.1).

ne - N / 2
CVR = (3.1)
N/2

where N = the total number of panelists

ne = the number of panelists indicating “essential”

It is inferred from the CVR equation that the content validity ratio
takes on values between -1.00 and +1.00 where CVR=0.00 means that
50percent of the panelists of size N believe that a measurement item is
“essential”. A CVR>0.00 would, therefore believe that a measurement item is
“essential” and thereby valid.

Lawshe (1975) has further established minimum CVR s for different

panel sizes based on a one-tailed test at the 0.05 significance level as 0.5; this
was used for the research too.

Nunnally (1978) Content validity ratios were thus calculated for the
questions which could measure WL and JP for the Content Validity Ratio
 80

Table. All the statements in QWL and JP were considered after the
calculation of content validity ratio. Only statements which have a score of
0.50 or above have been included in the survey instrument. After the content
validation and reliability check, the final version of the questionnaire was
arrived at. In order to evaluate the reliability level of the data, Cronbach alpha
test is conducted. Only elements with alpha value of 0.70 or above are
considered. For all the variables of factor analysis, alpha value is above 0.70
which shows the internal consistency of the scales (Cronbach 1981).

3.8 SAMPLE DESIGN

There are 1200 ITES organisations registered with the Registrar of

Companies in Bangalore, Karnataka State, India, of which, 303organisations
involve in process of survey. All 303 organizations were approached and all
303 organisations were willing to participate. The employees at the
Managerial level from these 303 organisations constitute the universe for the
research. Total of 721 employees are functioning at the Managerial level in
these 303 organisations. The judgmental sampling was adopted for selecting
employees from companies which had employee engagement activities.
Multinational IT companies and Indian IT companies were selected only if
they had Employee Engagement practices prevalent for the last three years.
Convenient sampling was adopted to select the respondents. The sample size
was primarily designed to be 0.76 percent that is 500 employees. In order to
tackle the non response, more than 500 employees were selected using non-
probability purposive sampling and questionnaires were issued. Out of which
524 employees responded to the questionnaire. After cases with missing data
were eliminated, the final sample consisted of 500 useable responses.
 81

3.9 DATA ANALYSIS TECHNIQUES

The statistical procedures used are selected based on the suitability

to examine the objectives of the research using SPSS 21 statistical package.
The various statistical tools adopted are presented as framework for analysis.

Data analysis involves converting a series of recorded observations

into descriptive statement and/ or inferences about relationships. For the
Statistical analysis of the data used in the present study, the major tools used
which includes Factor Analysis, Multiple Regression analysis, Analysis of
Variance (One-way ANOVA), Chi-Square test, ‘t’test, pie-charts and
averages, percentages, graphs, bar diagrams and SPSS 21 software packages.

Collis & Hussey (2003) A Likert rating scale as described was used
for section two of the questionnaire. The questions were turned into
statements and the respondents were asked to indicate their level of agreement
by checking the chosen box with an “x”.

Collis & Hussey (2003) With the exception of sub-section me, all
question statements were posed in a positive context. The benefit of this was
to discourage leading statements, i.e. leading the respondent into a negative
context. It is after all, the negative context that the researcher attempted to
invalidate, but if established, it is an indication of a problem area. Therefore,
if the answer is “Disagree”, then it is actually so. Propose that questions of a
sensitive nature should be avoided, or if asked, they should be towards the
end of the questionnaire; however they strongly advise against asking
negative questions.
 82

3.9.1 Factor Analysis and Other Statistical Tools

Factor analysis has been used here to identify and define the
underlying dimensions in the original variables and is used to reduce the
number of variables by eliminating redundancy. The general purpose of factor
analysis is to find a method of summarizing the information contained in a
number of original variables in to a smaller set of new composite dimensions
(Factors) with minimum loss of information. That is, the Factor Analysis tries
to identify and define the underlying dimensions in the original variables.

Factor analysis usually proceeds in four steps:

a. First, the correlation matrix for all variables is computed.

Variables that do not appear to be related to other variables
can be identified from the matrix. The appropriateness of the
factor model can also be calculated.

b. Factor extraction, the number of factors necessary to represent

the data and the method of calculating them must be
determined. At this step, how well the chosen model fits the
data is also ascertained.

c. Rotation focuses on transforming the factors to make them

more interpretable.

d. Scores for each factor can be computed for each case. These
scores are then used for further analysis.

The set of 52 objects included in the Employee Attrition Scale was

used to find the underlying factors in it. Identified 10 factors affecting high
employee through factor analysis and a 10 factors model was developed.
 83

Analysis of variance (One-way ANOVA), Chi-Square test, and t-

test are done on the primary data. Also relevant interpretations of the tables
and other data have been done using pie-charts, percentage diagrams and bar
diagrams. Analysis of the respondent’s personal and other relevant factors
namely gender, location of the respondents, global position (multination
/national), age, experience in the present organization, respondent’s salary,
designation, qualification, area of work, number of training programs
undergone have been done and percentage charts are prepared and
interpretation also is given for the above factors. Also ranking of reasons for
stress to IT employees is done with suitable charts. Multiple regression
analysis was applied to find the critical factors and the non-critical factors or
variables which might affect the attrition of the employees.

3.9.1.1 Principal component analysis using varimax rotation

The main applications of factor analytic techniques are:

a. To reduce the number of variables and

b. To detect structure in the relationships between variables, that

is to classify variables.

Therefore, factor analysis is applied as a data reduction or structure

detection method.

The Principal Component Analysis is appropriate when researcher

obtained measures on a number of observed variables and wishes to develop a
smaller number of artificial variables (called Principal Components) that will
account for most of the variance in the observed variables. The Principal
Components may then be used as predictor or criterion variables in
subsequent analyses.
 84

Nunnally & Berstein, (1994) The Principal Component Analysis

with Varimax Rotation for (Retention factors) was done to investigate the
underlying relationships of a large number of objects and to determine
whether they can be reduced to a smaller set of factors. This analysis has a
high potential to inflate the component loadings. Thus a higher rule-of-thumb,
a cut off value of 0.40 is adopted.

The Kaiser-Meyer-Olkin measure of Sampling Adequacy and

Bartlett’s test of Sphericity were used to determine the appropriateness of
factor analysis. The four dimensions of retention factors were identified and
labelled as Reward system, Managerial style, Organisational support and Job
itself. The factor loadings in the four factors range from 0.50 to 0.77. The
total variance of retention factors explained is 60.426percent. The Principal
Component Analysis is employed for extracting factors.

The Varimax rotation was used as it is centre on simplifying the

columns of the factor matrix. The latent root criterion is used for the
extraction of factors. As per the criterion, only factors having latent roots or
Eigen values greater than one are considered significant by means of item
analysis based on Pearson correlation.

3.9.2 Correlation Analysis

Correlations are useful because they can indicate a predictive

relationship that can be exploited in practice.

3.9.2.1 Pearson product moment correlation and canonical correlation

The present research attempts to find the relationship between

Retention factors using Pearson Product-Moment Correlation and Canonical
Correlation. The correlation based on the Pearson Product-Moment
 85

Correlation is performed to understand the relationship between all factors of

the research. The correlation coefficient r is a measure of the linear
relationship between two attributes or columns of data. The value of r can
range from -1 to +1 and is independent of the units of measurement. A value
of r near 0 indicates little correlation between attributes; a value near +1 or -1
indicates a high level of correlation. When two attributes have a positive
correlation coefficient, an increase in the value of one attribute indicates a
likely increase in the value of the second attribute.

3.9.2.2 Canonical correlation analysis

Canonical Correlation Analysis is one of the multivariate analysis

techniques that focus on the simultaneous analysis of relationships between
multiple independent and dependent variables. If there are two sets of
variables, and there are correlations among the variables, then Canonical
Correlation Analysis will enable to find linear combinations of the x's and the
y's which have maximum correlation with each other. A typical use for
Canonical correlation in the experimental context is to take two sets of
variables and see what is common amongst the two sets. Following Pedhazurs
(1982) recommendation, interpretation is based on variables correlating 0.30
or greater with the Canonical Variate. This technique is employed to
investigate the percent of variance of one set of variables explained by
Canonical Variate for the other set along the dimension represented by the
Canonical Correlation.

3.9.3 Multiple Regressions

Regression models are created to understand the impact of Retention

factors. Multiple Regressions is a statistical method used to examine the
relationship between one dependent variable Y and one or more independent
 86

variables Xi. The regression parameters or coefficients in the regression

equation are the coefficients of the independent variable.

Y = b0+b1X1+b2X2+b3X3……bkXk

The coefficient of determination R2 is the proportion of the

variation in the dependent variable explained by the regression model. The
enter method which is a default standard method in SPSS and also a
straightforward method is used.

Before proceeding with Regression Analysis the assumptions of

Regression analysis have to be fulfilled. None of the multivariate analysis
may yield reliable results if the assumptions are not satisfied.

3.10 HYPOTHESIS

Objective: To study the retention of employees in MNCs of Information

Technology sector in general.

1.1 There is no significant difference among the IT industries in

applying scientific concepts and theories to the solution of
problems.

1.2 There is no significant difference among tasks consists of carrying

out technical work connected with the application of concepts and
operational methods.

1.3 There is no significant difference among possess competencies that

are exclusive

1.4 There is no significant difference among supportive to the

organisation demands, culture, and vision &Mission for getting
success
 87

1.5 There is no significant difference among involvement and active

participation in the carry out the essential tasks.

1.6 There is no significant difference among provide skills and/or

knowledge to ensure the success of the organization, sequence plan

1.7 There is no significant difference among Attractive attitude, friendly

relationship, 24x7 support, flexi schedules, nightshifts, and work
from home.

1.8 There is no significant difference among Global exposure and

practice while working from India.

1.9 There is no significant difference among site experience, innovative

talent, and professional satisfaction.

1.10 There is no significant difference among fulfilling the resources to a

competitive edge to the organisation

1.11 There is no significant difference among contributes to the

organizational memory – their departure would drain the
organization’s knowledge and sill bank.

1.12 There is no significant difference among possesses skills,

knowledge and abilities that ate relatively rare and irreplaceable.

1.13 There is no significant difference among essential to an

organization’s productivity and well being.

1.14 There is no significant difference among key motivators, mentors or

role models to other staff.

1.15 There is no significant difference among knowledge, skills and

attributes that are closely aligned with the existing or possible future
operational direction of the business.
 88

1.16 There is no significant difference among display identification with,

and commitment, to the organisation.

Objective: To identify the factors of retention of employees in Information

technology of Bangalore city, India.

2.1 There is no significant difference among the IT industries in

effective selection & recruitment process, and job design.

2.2 There is no significant difference among the provision of effective

training & development practices

2.3 There is no significant difference among Innovative fair and

equitable performance evaluation practices.

2.4 There is no significant difference among career development, and

promotional practices, employee assistance programs.

2.5 There is no significant difference among challenging employment

structures and opportunities.

2.6 There is no significant difference among equity of compensation,

social benefits and security of tenure.

2.7 There is no significant difference among reward and recognition of

employee value performance education

2.8 There is no significant difference among Grievance handling

systems

2.9 There is no significant difference among employee satisfaction and

Motivation
 89

2.10 There is no significant difference among employee retention

methods and employee commitment

Objective: To identify and analysis the perceptions of the employees with

regard to the career development practices being adopted by Information
technology

3.1 There is no significant difference among the IT industries company

policies, vision, mission and culture

3.2 there is no significant difference among job knowledge

communication and coordination

3.3 there is no significant difference among job satisfactory and

flexibility working environment

3.4 There is no significant difference among decision making and

employment of right employee.

3.5 There is no significant difference among provision of high –tech

performance equipment

3.6 There are no significant difference among Influential and sensitive

leadership style.

3.7 there is no significant difference among effective integration and

working relationship

3.8 There is no significant difference among effective performance

appraisal announcement

3.9 There is no significant difference among organizational loyalty and

pride.
 90

3.10 There is no significant difference among reward and suitability of

feedback to the employee.

3.11 CONCLUSION

This study explores the impact of Human Resource practices on

retention factors, as such the data consists of more than one variable the above
statistical techniques have been used to analyse the data for interpretation.
The IT profession is indeed a stressful profession. IT professionals have to
cope with long working hours, unending user demands, unmet deadlines and
skill obsolescence. If IT professionals face tremendous job stress which
cannot be contained, burnout will happen in them and this will lead to a
variety of health problems.

According to a recent study done in IT professionals are found to be

highly stressed. There is clearly a need for organizations and IT professionals
to pay attention to and manage the stress of IT professionals. Organizations
that fail to do so are likely to have dissatisfied employees and high staff
turnover. IT professionals who fail to effectively manage their stress may end
up with a variety of health problems.