Are workers with a disability less productive or less understood? An empirical investigation from an entrepreneurial business planning perspective

Full text

Deakin Research Online
Deakin University’s institutional research repository
DDeakin Research Online
Research Online
This is the authors’ final peer reviewed version of the item
published as:
Hindle, Kevin, Noble, Jock and Phillips, Brian 1999, Are workers with a disability less
productive or less understood? An empirical investigation from an entrepreneurial
business planning perspective, in ANZAM 1999 : Proceedings of the 1999 Australian and
New Zealand Academy of Management conference, ANZAM, [Lindfield, N.S.W.], pp. 1-
37.
Available from Deakin Research Online:
http://hdl.handle.net/10536/DRO/DU:30029671
Reproduced with kind permission of the copyright owner.
Copyright : 1999, the authors

- 1 -
-
The Productivity of Workers with a Disability:
Evidence Dispels Past Myth; Entrepreneurship Plans
Future Reality.
Kevin Hindle
Division of Entrepreneurship Research
Swinburne Graduate School of Management
Jock Noble
Research Fellow
Swinburne Graduate School of Management
Chief Executive Officer
Central Marketing Services
Brian Phillips
School of Mathematical Sciences
Swinburne University of Technology
CONTACT
Kevin Hindle
Wk 9214 8732
Hm 9853 7059

- 1 -
-
Are Workers with a Disability Less Productive or Less
Understood? An Empirical Investigation from an
Entrepreneurial Business Planning Perspective.
ABSTRACT
This study investigated selected work-performance data of a large call
centre using the entrepreneurial business planning paradigm as a
theoretical framework and tested the hypothesis that levels of
productivity would be different for each group between workers with a
disability and workers without a disability. On five measures of
productivity, no significant differences were discernible but on a sixth
measure, length of employment, it was found that disability workers
remained in employment significantly longer. These results strongly
refute the ‘intuitive wisdom’ that workers with a disability are less
productive. The results support a growing body of corporate experience
and descriptive research indicating that workers with a disability perform
as well as or better than their non-disability colleagues. Yet workers with
a disability remain disproportionately under-employed. The key to
translating the growing evidence of this research into higher levels of
employment of workers with disabilities will depend upon employers
adopting an entrepreneurial approach to the planning of human resource
management.
Key Words: Disability, workers, productivity, entrepreneurship,
entrepreneurial business planning.

- 2 -
-
INTRODUCTION
The Research Problem
The purpose of this study was to investigate selected work-performance
data of a large call centre to detect any significant productivity
differences between workers with a disability and workers without a
disability. The research was conducted from an entrepreneurial business
planning perspective in order to enhance both ability to understand and
capacity to utilise a possibly misunderstood human resource. The null
hypothesis was that there are no differences in average productivity
between the two groups of workers.
Much of society in general and many employers in particular hold it as
axiomatic that workers with a disability are less productive than workers
without a disability. If this seemingly self-evident proposition is both
strongly believed and demonstrably wrong, workers with a disability are
an under-utilised human resource. This under-utilisation might give rise
to an entrepreneurial opportunity. Organisations understanding the true
value of workers with a disability might plan and derive resource benefits
not available to competitors operating on false assumptions and averse to
the alleged risks of utilising people with disabilities to perform important
tasks. So, this study had its origins in the contextual development of one
simple question: is the general assumption that workers with a disability
are less productive true or false? Guided by the developing theory of
entrepreneurial business planning (EBP), a focused empirical
investigation emerged. Theoretical context was provided by the EBP
paradigm (Hindle 1997; Legge and Hindle 1997). The empirical context
was provided by data randomly sampled from the employment-
performance records of a call centre run by Telstra, Australia’s largest
telecommunications company.

- 3 -
-
Definition of Key Terms and Problem Orientation
A call centre is defined as a ‘managed environment where telephone is
used systematically to provide value added contact with customers and
suppliers ‘(Hallis: ?). The definition of disability is ‘a condition caused
either by accident, trauma, genetics or disease that may restrict a person’s
mental, sensory or mobility functions to undertake or perform a job in the
same way as a person who does not have a disability’ (ref ?).
Entrepreneurship is ‘the creation and management of a new organisation
designed to pursue a unique, innovative opportunity and achieve rapid,
profitable growth’ (Hindle 1999: ?).
For the purposes of this investigation, the essential characteristic of
entrepreneurship is its emphasis upon what might be called ‘opportunity-
driven management’. This is a managerial approach not constrained by
resources currently controlled but where efficient and effective
deployment of the minimum feasible set of required resources is an
essential component of planning and operating an enterprise (see
Stevenson, Roberts and Grousbeck: 8-11 and 22-28). An entrepreneurial
business plan is the key tool for securing this minimum feasible set of
required resources (Hindle 1997: xx). Detailed definitions of the
entrepreneurial business plan as an output, entrepreneurial business
planning as a process, and the EBP paradigm as an operational
framework are provided in a following sub-section of the paper.

- 4 -
-
The Disability-Productivity Literature: Alleged Productive Capacity
Unmeasured and Unused
American and Australian figures suffice to demonstrate that workers with
a disability constitute a heavily under-utilised resource in developed
countries.
In the USA 11.4 million people with work disabilities do not participate
in the labour force and in addition there are 723,000 who are actively
looking for work. This number alone represents a 13.4% unemployment
rate among jobseekers with a disability - more than twice as high as the
5.6% unemployment rate for people without disabilities (LaPlante et al.
1998). In all, only 27.8% of working age people with a disability have
jobs (4.7 out of 16.9 million) compared to 76.3% of those without
disabilities. Among working age unemployed people with a disability
79% say they would like to have a job. Australia also has an opportunity
gap. The Australian Bureau of Statistics (1993) reports that 3,176,700 or
18% of the Australian population have a disability. The workforce
participation rate of people with a disability is 46.5% compared to the
participation rate of 76.9% of people without a disability. These under-
utilisation rates occur in an environment where, as Lankard says:
…
‘The cultural, educational, economic and societal diversity
among members of the workforce will continue to force
organisations to look to unique staffing, scheduling, and
training policies and practices that will attract qualified
workers and meet their personal as well as professional
needs.’ (Lankard 1993: 1)

- 5 -
-
Such corporate and academic research as has been done indicates both
favourable attitudes and positive performance when workers with a
disability are given a chance.
Concern that persons with disabilities are often eliminated from
consideration for jobs because of erroneous stereotyping led Smith and
others to a study the issue. They found, unsurprisingly, that both attitudes
towards and performance appraisals of workers with a disability were
more positive from employers who had previously worked with or
supervised people with a disability. (Smith et, al, 1985: 39-41). Wolfe
found that managers at DuPont who had worked with people with a
disability rated their performance at levels comparable to employees
without disabilities (Wolfe 1973). In a qualitative study, Johnson et al.
(1988) found that despite many employers perceiving employees with
physical disabilities as having positive work performance capacity, there
was still reluctance to employ them because of worries about mental,
emotional and communication stability.
Some corporations have transcended these concerns, expressing and
acting on strong belief in the capacities of workers with a disability.
Sears, Roebuck and Company has been running affirmative action
programs for people with a disability since 1947, and now employs
people with a disability at all levels from repair technicians to attorneys.
Sears also monitors their progress to ensure they are being promoted on
the basis of their performance. IBM started teaching typing and key-
punching to the blind as well as persons with cerebral palsy in the 1940’s.
As long ago as 1972 IBM began a major program to train and place
people with severe physical disabilities as entry-level computer
programmers. This initiative now embraces 16 centers throughout the
United States. IBM’s efforts have also included developing and

- 6 -
-
marketing new products for people with a disability, such as computers
that can print Braille.
One major and continuing corporate attempt to assess, systematically, the
contribution of people with a disability to a workforce has been
maintained by the DuPont Corporation in the United States. A series of
studies under the heading, Equal to the Task - A Survey of Employment
of the Handicapped, have been conducted and internally published in
1958, 1973, 1981 and 1990. The studies do not use inferential statistics
but report simple percentages of those surveyed. DuPont found that using
the data collected in 1973 from 1,452 employees with a disability and in
1981 from 2,745 employees with a disability, the diversity of their
impairments did not adversely effect safety, job duties or attendance.
What stood out, was the uniformity of their performance irrespective of
particular disabilities (DuPont 1981: 5-8). DuPont stated that ‘the
significance of the 1981 and earlier surveys is the picture that emerges of
workers with disabilities as an important human resource’. DuPont lists a
number of individual case studies that demonstrate how employees with a
disability have mastered a broad range of occupations and how many,
through uncommon ingenuity, have overcome serious limitations in order
to pursue their professions. (DuPont 1981: 10 –16 and 1990: 8 -19).
A small volume of mainly descriptive academic research supports the
faith of pro-disability corporations.
Based on a study of the responses of 65 supervisors in human service
agencies and 27 employers, Reisman and Reisman (1993) found that
people with a disability compare favorably to the general population in
terms of some basic work habits. Zemans, and Voelckers (1994) argued
that there are long term benefits to employing the disabled. Rusch,
Wilson, Hughes and Heal (1994) found that interactions between workers

- 7 -
-
with disabilities and workers without disabilities are remarkably similar.
Zivolich (1997) reported that Pizza Hut’s laudable initiatives for persons
with severe disabilities over 10 years, have in the company’s belief,
resulted in substantial financial benefits. Levy, Jones, Jessop and Levy
(1992) studied 1,140 CEOs of Fortune 500 industrial and service
corporations seeking to measure their attitudes to employing people with
severe disabilities. 16 predictor variables were chosen as representing
respondent characteristics, and the mean scores recorded on attitude
scales by corporate executives responsible for hiring decisions were
analysed. The findings suggest, once again, that it is positive contact
around work itself that determines the attitudes towards employability
and that attitudes towards people with a disability are also similarly
effected by such employment contact.
Outside corporations, the American population is favourably disposed. A
1991 Harris poll showed that American people recognise the potential
contribution of workers with a disability. Eight out of ten people agreed
that people with a disability have under-used potential ‘to contribute by
working and producing’ and only one out of ten disagreed. (Brown 1993:
60-62).
So, despite the perceptions and good intentions of pro-active corporations
and the general public, why is it that workers with a disability constitute a
heavily under-utilised resource in developed countries?
A clear theme emerging from the productivity/disability literature is that
‘you have to try it to appreciate it’: that ability to judge the capacities and
productivity potential of workers with a disability is a function of
experiencing their performance in the work environment. One absent
theme in the literature is evidence of any willingness on the part of small,
entrepreneurial and early-stage businesses to try using workers with a

- 8 -
-
disability. All the descriptive studies come from large corporations. This
is dangerous because it is well-established by research too numerous to
cite that the majority of job growth in an economy will come from
entrepreneurship: high-growth-potential new ventures. It is a reasonable
inference that entrepreneurs may be making the judgement that, with so
many risk factors already militating against new venture success, it would
be foolhardy to add the risk of lower employee productivity. These two
themes converge on one problem.
The problem lies with employers’ – and especially entrepreneurial
employers’ - preparedness to take the alleged risk of employing people
with a disability in the first place. And here, research to date has not
provided truly hard evidence of the desirability of ‘taking the plunge’.
Existing research does not encourage the entrepreneur to be
entrepreneurial. Entrepreneurs are attracted to challenges not risks. No
study until this one has formally tested, in a dispassionate, empirical,
quantitative manner, the hypothesis that there is no difference between
the average productivity of workers whether or not they have a disability.
An empirical test of this ‘no productivity difference’ hypothesis will be
valuable. If there were a clear, empirical demonstration that the alleged
risks of employing the disabled are not risks at all, this might enhance the
likelihood that the most important constituency of all employers – the
entrepreneurs of potentially high-growth new ventures – will increase
their willingness to employ people with a disability. The consequences
for employment could be profound.
The EBP Paradigm as a Theoretical Framework
Any examination of the productivity of workers with a disability runs the
risk of being embroiled in some highly emotive issues. ‘Disability’ at any

- 9 -
-
level of debate is a philosophically, socially, morally and anecedotally
complex concept. On the other hand, ‘productivity’ at its most general
conceptual level, is relatively simple. It is a description or measure of
economic performance: not social, moral or any other kind of value. At
its crudest, productivity is a ratio: a number obtained from dividing the
yield of outputs by the cost of inputs. Of course, no sensible social
scientific approach to the productivity of human beings – with or without
a disability - can operate in either a social or a moral vacuum. However, a
concentration on quantitative performance comparisons between people
with and without disabilities can be helpful if it is recognised as merely a
first step to an integrated approach. This is especially so if a structured
empirical investigation can help to dispel a myth. The balancing trick to
is to find a theoretical framework where focus on measuring productivity
is sharp but context is still rich enough to support discussion of extended
implications.
The theory of entrepreneurial business planning provides a useful balance
between the requirement for hard, empirical numbers and an ability to
interpret those numbers as meaningful social science. Hindle (1997)
sought to put EBP on a sound theoretical basis. He provided the
following core definitions (12).
'Entrepreneurial Business Planning (EBP) is the process of
convincing investors of the desirability of investing in a new venture
by articulating and programming the economic consequences of a
strategy which determines relevant antecedent variables, expresses
them in holistic relationship and subjects them to sensitivity analysis
in order to maximise the probability of a desired change of state.'

- 10 -
-
'An Entrepreneurial Business Plan is the formal argument used to
secure, from prospective investors, resources required for a proposed
entrepreneurial process.'
These definitions clarify the research problem. The task of getting an
employer to make that first crucial decision to ‘give a worker with a
disability a chance’ is an example of soliciting an investment decision in
an entrepreneurial context. In this case, the ‘plan’ is very abstractly
conceived as an argument to convince employers to make an initial
investment in employing a worker with a disability. The EBP paradigm is
appropriate.
In general, the entrepreneurial business planning paradigm at any given
time will be the accepted mixture of theoretical and practical principles
which best describes, explains and justifies EBP as a distinct field of
human endeavour and posits rules for the creation of a successful
entrepreneurial business plan. To clarify the many issues involved in
understanding the EBP paradigm, Hindle developed the ‘interrogative
matrix’ illustrated in table 1, below. The four column headings -
boundaries, laws, success rules and instrumentation requirements - are the
four essential elements of any paradigm. The three row headings -
communications, control and simulation - represent the three roles of a
plan, as defined by Mintzberg (1994, passim) and others.
Table One
The Questions Involved in The EBP Paradigm

- 11 -
-
At any given time, the ‘state of the art’ of the EBP paradigm will be
defined by the answers to questions contained in the various cells of the
matrix. With increased knowledge and research, the paradigm will
develop by providing gradually changing answers to the key questions
contained in this framework. The current state of the EBP paradigm – i.e.
the replacement of all the question marks in figure 1 by answers - is
described in detail in Legge and Hindle (68-94) and a summary in matrix
format is attached as appendix 1. The prevailing EBP paradigm is a
complex network of research-derived relationships involving seven
boundary conditions, twelve laws, six success rules and two primary
instrumentation requirements.
However, what is most important for this study is not every item of
paradigm detail: the current specific answers to any of the key EBP
questions in the framework of figure 1. It is the overall managerial
perspective that EBP provides on the issue of resource evaluation. At its
essence, EBP is focused on assessing, acquiring and managing resources
– including human resources - in the context of opportunity management.
This is an appropriate perspective for investigating the possibility that
workers with a disability may be a misunderstood, opportunity-laden
resource for organisations capable of an entrepreneurial approach to
management. The paper demonstrates that the EBP perspective proved
What are the defining elements? How does one obtain success?
PARADIGM PARADIGM PARADIGM INSTRUMENTATION
BOUNDARIES LAWS SUCCESS RULES REQUIREMENTS
COMMUNICATIONS ????
CONTROL ????
SIMULATION ????
THEORETICAL JUSTIFICATION. Why does this paradigm contain these prescriptions?

- 12 -
-
useful both for designing and interpreting the empirical aspects of this
research.
Measuring productivity in an EBP context
An initial sharp focus for this study was established by taking a subset
(three laws) of the prevailing EBP paradigm and using it to develop an
empirical approach to the disability/productivity question. Subsequent
ability to discuss and interpret results broadly was guided by the totality
of the prevailing EBP paradigm.
Three laws of the prevailing EBP paradigm (Hindle 1997: 122) are set
out below.
Law 4. Identify all major plan objectives, primarily as financial targets.
Law 5. Define the investment offer(s) as an expected return on
investment.
Law 7. Provide comprehensive statements of opportunities and risks.
Any decision to employ a worker with a disability is primarily concerned
with the balance between opportunity and risk. To assess that balance
requires the ability to specify financial objectives and calculate an
expected return on investment. An ROI calculation relies on financial
targets. Financial targets for resources are dependent upon productivity.
Thus, from an EBP perspective, any measured decision about the value of
workers with a disability must start from establishment of some measures
of productivity.
The entrepreneurial business planner’s expectation of employee
productivity is a construct composed of measures in five distinct
categories. For ease of illustrating the distinction between these five

- 13 -
-
measurement categories, the simple analogy is made of the worker being
a male ‘digger’ whose task is ‘digging holes for fence posts’ (illustrations
are in brackets). The measurement categories are:
1. Measures of Experience. (How long have you been a digger?).
2. Measures of Attendance. (OK, you’re experienced but I won’t get
the benefit of it if you are often absent from work.)
3. Measures of Task Engagement. (The ratio of digging time to paid
time – digging versus tea breaks etc.)
4. Measures of Task Efficiency. (OK you’re digging – not having
cups of tea – but is the hole getting deeper or are you just resting
on your spade?)
5. Measures of Task Effectiveness. (OK, you’re making holes but
are they deep enough, in the right place, well-constructed etc?)
The method section of the paper (below) shows how these measurement
categories were represented by specific operational variables.
The Investigative Context: Anatomy of the Call Centre Industry
Increasingly, organisations are gaining efficiencies by establishing or
subcontracting centralised telephone call centres. The first commercial
call centre began in America in 1968 when a US federal court judge
ordered the Ford Motor Company to establish free phone lines to
facilitate recall of a faulty car. Today, the telemarketing and call centre
industry is one of the fastest growing industries worldwide and in
Australia. For many organisations, call centres provide the entry point for
customers to an organisation: they are the focus of the organisation’s
service provision. Call centres cost companies about half what it costs
them to communicate in writing and play an increasingly important role
in improving customer retention rates, assisting with customer acquisition

- 14 -
-
and generating increased sales (Hepworth and Co: ?)1 In Australia, in
1998, the industry was reported to be worth $1.8 billion a year, growing
at an annual rate of over 20% and employing over 40,000 people(ref). 2
Ibis Research has forecast that currently in Australia, 3 of every 1,000
new jobs created are in a call centre. Ibis predicts that, by no later than
2003, the figure will have increased tenfold to 3 out of every 100 new
jobs created (ref).3
The call centre environment was chosen for three reasons.
1. Call centres constitute one of the fastest growing industries in the
world and the features of the work environment are similar
irrespective of nation, language or culture. Thus, any findings might
have a higher illustrative value than findings from a less cosmopolitan
environment.
2. A call centre is a highly monitored environment where accurate
record keeping of work-performance and productivity data is highly
automated. This reduces the chance for bias and human error in the
data recording process. Each individual employee's performance is
closely monitored and recorded electronically in the same way, for
every hour of paid work and so meaningful comparisons between
people with a disability and people without a disability can be
conducted.
3. People with a disability make up 14 % of Australia's workforce and
50% of these are currently unemployed (ref). Many call centres have
up to 33% of their workforces composed of people with a disability
and are large places of employment. This makes them better providers
of sample sizes conducive to effective quantitative research.
1 'Canadian Society of Consumer Affairs Professionals Tollfree Number Study' 1996
Hepworth & Company Ltd
2 call centre Staff Salary Survey. Hallis May 1998

- 15 -
-
METHOD
General Research Design
The empirical study was a simple comparison of means design using t-
tests, supported by the non-parametric Mann-Whitney U test as a
precaution against the distribution of results not coming from a normally
distributed population. The mean performance of two samples (one
representing the call centre’s population of workers with a disability and
one the population of workers without a disability) were compared on six
variables. The analytical aim was the same for each performance
variable: to gather evidence against the null hypothesis that there was no
difference between the mean performance of workers with a disability
and workers without a disability.
Data Source, Sampling and Collection
The research reported in this paper was conducted using Burwood call
centre (hereafter ‘Burwood’) as a sample frame. Burwood is operated by
Telstra, Australia’s largest telecommunications company. In July 1998,
Burwood employed approximately 250 people, with and without
disability. Australiawide, Telstra has over x call centres ranging in size
form more than 800 operators to less than 10. In metropolitan Melbourne
(a city of over 3 million people) Telstra has x call centres ranging in size
from over x operators to less than y operators with an average employee
complement of z. After agreeing to provide productivity data to the
researchers, Telstra nominated Burwood as the sample frame, because it
was a ‘typical, large, metropolitan call centre’. Of course, this does not
3 Ibis Research 1997

- 16 -
-
qualify the choice of Burwood as a random selection. However, from this
point on, random sampling was employed.
Within the Burwood call centre, three samples were drawn at random.
Taking three samples rather than one was a decision motivated by ethical
considerations on the part of Telstra management to help ensure
respondents’ anonymity. First a sample of 200 names was drawn to
supply data for the first variable – ‘length of service’ (see next section for
details of all variables). Second a separate sample of 200 names was
drawn to supply data for the second variable – ‘absent days’. Finally, a
sample of 65 employees was used to provide data for the final four
‘efficiency and effectiveness’ variables.
Once the samples were drawn by Burwood’s centre manager, recorded
performance data were ordered from Telstra’s central records department.
The records covered all operators employed in Burwood as at the 8th of
July 1998, ensuring that neither Telstra nor the researchers were
infringing privacy laws, or any ethical standards required of responsible
social research. Privacy protocols were established in conjunction with
Telstra’s national call centre manager, Mr. Robert Holland, who
consulted with Telstra’s legal team.
Assumptions
1. Currently 68% of call centres intend to base the future remuneration
of their staff on workers individual performance6. This study assumed
individual performance would increasingly be a factor in the
employment of staff.

- 17 -
-
2. Workers are employed and assessed on the basis of their capacity to
undertake call centre telephone operations, and not their capacity to
move into other staff roles in the call centre or the larger organisation.
3. The skill shortages currently faced in the industry will continue,
leading to the recruitment of new employees into the sector.
4. Advances in technology will not significantly change the basic
competencies or performance required for call centre operators.
5. The environment at Telstra's Burwood call centre is generally
representative of call centre technology and staff performance.
6. No environmental change or other influencing factor significantly
affected established patterns and trends in performance during the
data collection period covered by the research.
7. There are no significant differences in performance of permanent staff
compared with part-time staff on a shift-for-shift basis.
Limitations
• This was a study with high internal validity and limited external
validity. There is no reason to think other than that Burwood represents a
typical Telstra call centre. However, Burwood was a convenience choice
mandated to the researchers by Telstra. So, this study limits its inferential
arguments to the claim that the three samples drawn from within
Burwood were representative of the populations of disability and non-
disability workers at this particular call centre – not beyond it.
Accordingly, Burwood provides a quantitatively analysed case study
rather than a quantitative basis for extrapolation to any larger populations

- 18 -
-
of disability/non-disability workers. Further research from a larger
sampling frame would be required for claims of greater generality to be
valid.
• People classified as ’with a disability’ were not sub-divided by
degree or type of disability.
Measurement Framework – Six Variables
There was one variable of experience - length of service. This was
literally a measure of the amount of time the respondent had been
employed at Burwood, irrespective of the category of employment (part
or full time). Sample drawn was 200. After data cleaning 192 usable
cases remained.
There was one variable of attendance – absentee days. This was the
number of absentee days the respondent had logged in the calendar year
preceding July 8, 1988. Absenteeism was defined as the failure of
operators to report for work when they are scheduled for work. This did
not include operators who were away on recognised holiday, vacation, or
approved leave of absences. Unplanned absences in the context of this
research refer to absences that are recorded as unplanned by Telstra, these
were consistent with the above definition. The most common
explanations for absenteeism were sick leave with or without a doctor’s
certificate and sick leave taken for a family member but excluding
maternity leave. Sample drawn was 200. After data cleaning x usable
cases remained.
The final sample was of 65 cases, comprised a ‘work section’ (which in
turn consisted of four ‘work groups’) at Burwood, chosen at random from
the four work sections which comprised the total Burwood workforce.

- 19 -
-
Raw data for these cases covered the months of May and June 1998 and
included: ‘paid hours’, ‘logged hours’, ‘contacts made’, ‘upgrades per
100 calls’ and ‘new sales per 100calls’. From this data, four composite
variables were constructed.
There was one variable of task engagement - logon ratio. This was the
subject's total hours spent logged on (i.e. actually making phone calls) for
the months of May and June as a percentage of the total paid hours of
every worker in all four groups.
There was one variable of efficiency - contact efficiency. This was the
subject’s percentage of total customer contact hours for the period.
There were two variables of effectiveness. - upgrade effectiveness index
and newsale effectiveness index. An ‘upgrade’ was defined as the sale of
additional features of a service to a client already subscribing to that
service at a more basic level. A ‘newsale’ was defined as the sale of a
completely new service or product to someone not currently using it.
Each index consisted of the subject’s sales-per-100-calls in May and
June, averaged them, and then divided them by the averaged total of sales
for May and June of the whole group.
Hypothesis
For each of the six variables the competing hypotheses were identical.
The null hypothesis, H0, was  :
On this productivity measure, the mean scores of the two populations are
the same: there is no difference in the mean scores of workers with a
disability and workers without a disability.

- 20 -
-
The alternative hypothesis, H1, was :
On this productivity measure, the mean scores of the two populations are
not equal: there is a difference in the mean scores of workers with a
disability and workers without a disability.
A significance level of .05 was set, indicating the null hypothesis would
be rejected if significance testing yielded p-values less than .05. The
alternative hypothesis was deliberately made non-directional since no
assumptions about the nature of any differences were hypothesised.
Analytical regime
The t-test for independent samples was chosen as the principal analysis
method because it can effectively measure whether the differences in the
means of two groups are significant, even when the sample sizes are
relatively small, providing certain assumptions are met. These include:
• that the subjects used are randomly drawn from the population of
interest;
• that the data are normally distributed within each group;
• that the variances of the two groups are equal.
It is also desirable to have similar size samples in each group since this
leads to less risk of making incorrect conclusions with any small
violations of the assumptions.
To support the analysis, the non-parametric Mann-Whitney U test was
also applied to each variable as a precaution against any possible
violation of the t-test assumptions. Finally, the four variables of
efficiency and effectiveness were analysed in combination using a
multivariate test.

- 21 -
-
RESULTS
The data were analysed using both the SPSS (version 9) and the Stata
(version 5) packages. The non-directional independent samples t-test
using pooled variance estimates was used to test for differences in means.
The non-parametric Mann-Whitney U test for independent samples was
also used in all cases. The tests for normality were based firstly on
skewness then on kurtosis, and finally the two tests combined into an
overall test statistic of normality (Stata Reference Manual 1997: 223-4
and D'Agostino,Balanger and D'Agostino, Jr. 1990). Levene's test for
equality of variances was used.
Length of Service
The mean length of service for operators without a disability was 3.20
years, standard deviation = 1.400, n = 166 while for those with a
disability the mean was 4.11 years with standard deviation 0.927, n = 30.
The results from the two groups are summarised in the boxplots in Figure
1.
30196N =
Disability Status
With DisabilityNo Disability
Length of Service
7
6
5
4
3
2
1
0
-1
203
207212225
17141169165
16
13642118659031137749318341514477
14015052180628723145121156988
157
662539
14
26
Figure 1. Length of service for groups with and
without a disability

- 22 -
-
The means were significantly different, t = 3.442 on 224 df, p = 0.0007.
Testing for normality we see, for skewness p < .0005, Kurtosis p = .087
with a combined index 2 = 21.34, p < 0.00005. The test for equality of
variances indicates a significant difference between the variances of the
two groups, F-ratio variances of 2.6503 p=.0105. Hence, as the both the
normality and equal variance assumptions of the t-test have been violated,
a non-parametric test was carried out. The non-parametric results,
supported the t-test findings U = 1,773.5, z = 3.516 p = 0.00044
This suggested a significant difference in lengths of service. Operators
with a disability are likely to stay significantly longer than operators
without a disability.
Note: Initial analysis identified two outliers from the operators without a
disability (10.6 and 15.9 years) which was longer than the call centre had
been established. These were removed from the final analysis of this
variable.
Days Absent
The mean number of days absent for operators without a disability was
19.24 days, standard deviation = 18.365 , n = 158 while for those with a
disability the mean was 11.8 years with standard deviation 7.536, n = 30.
The results from the two groups are summarised in the boxplots in Figure
2.

- 23 -
-
30158N =
BODY
With disabilityNo disability
Days absent
120
100
80
60
40
20
0
-20
159
178
180181179
182
183
184
185186
187
188
Figure 2. Number of days absent for groups with
and without a disability
The means were significantly different, t = 2.181 on 186 df, p = 0.0305.
Testing for normality we see, for skewness p < 0.0005, Kurtosis p <
0.0005 with a combined index 2 = 62.31, p < 0.0005. The test for
equality of variances indicates a significant difference between the
variances of the two groups, F-ratio variances of 11.000 p=.0.0011.
Hence as the both the normality as equal variance assumptions of the t-
test have been violated, a non-parametric test was carried out.
The non-parametric results, do not support the t-test findings U = 1940.5,
z = 1.573 p = 0.116. This different result may have been due to the t-test
assumptions not being satisfied and/or the large differences in sample
sizes, 158 and 30. The disagreement in test results means that days absent
results are inconclusive. More extensive studies are needed to be
statistically rigorous, but from the sample results obtained it is suggested
that operators with a disability are likely to have less days absent than
operators without a disability.
More conservatively, it is very safe to infer no difference in average
performance.

- 24 -
-
Logon Ratio
The mean logon ratio for operators without a disability was 0.00784,
standard deviation = 0.00303, n = 43 while for those with a disability the
mean was 0.00688 with standard deviation 0.00324, n = 21. The results
from the two groups are summarised in the boxplots in Figure 3.
2144N =
Disability Status
With DisabilityNo Disability
Logon Ratio
.016
.014
.012
.010
.008
.006
.004
.002
0.000
-.002
23
60
48
25
Figure 3. Logon ratio for groups with and without
a disability
The means were not significantly different, t = 1.164 on 62 df, p = 0.249.
Testing for normality we see for skewness p = 0.015, kurtosis p = 0.477
with a combined index 2 = 6.07, p = 0.0481 indicating this variable is
close to satisfying the normality assumption. The test for equality of
variances indicates there is no significant difference between the
variances of the two groups, F= 0.697 p=.0.407. Hence both the
normality as equal variance assumptions of the t-test have been satisfied
for this variable. However for consistency with the other analyses, a non-
parametric test was also carried out. Its results support the t-test findings
U = 389, z = 0.894, p = 0.372.

- 25 -
-
This suggests there is no significant differences in the mean logon ratios
for operators with a disability compared to operators without a disability.
.
Contact Efficiency
The mean contact efficiency index for operators without a disability was
0.0142, standard deviation = 0.0054, n = 43 while for those with a
disability the mean was 0.0141 with standard deviation 0.00696, n = 21.
The results from the two groups are summarised in the boxplots in Figure
4.
2143N =
Disability Status
With DisabilityNo Disability
Contact Efficiency
.04
.03
.02
.01
0.00
-.01
52
4839
9
Figure 4. Contact Index for groups with and
without a disability
The means were not significantly different, t = 0.0664 on 62 df, p =
0.947. Testing for normality we see for skewness p = 0.516, kurtosis p =
0.042 with a combined index 2 = 4.58 , p = 0.101, indicating this
variable satisfies the normality assumption. The test for equality of
variances indicates a no significant difference between the variances of
the two groups, F = 0.697, p= 0.416.

- 26 -
-
Hence both the normality and equal variance assumptions of the t-test
have been met for this variable. However for consistency with the other
analyses a non-parametric test was also carried out. Its results supports
the t-test findings U = 446, z = 0.0787, p = 0.937.
This suggests there is no significant differences in the mean contact
efficiency index for operators with a disability compared to operators
without a disability.
.
Upgrade Sales Effectiveness
The mean Upgrade Sales Productivity Index for operators without a
disability was 0.0150, standard deviation = 0.00544, n = 43 while for
those with a disability the mean was 0.0155 with standard deviation
0.00353, n = 21. The results from the two groups are summarised in the
boxplots in Figure 5.
2143N =
Disability Status
With DisabilityNo Disability
Upgrade Effectiveness Index
.04
.03
.02
.01
0.00
38
48
Figure 5. Upgrade Index for groups with and
without a disability
The means were not significantly different, t = 0.376 on 62 df, p =
0.7086. Testing for normality we see for skewness p = 0.004, kurtosis p <
0.0005 with a combined index 2 = 16.03 , p = 0.0003, indicating this

- 27 -
-
variable does not satisfy the normality assumption. The test for equality
of variances indicates no significant difference between the variances of
the two groups, F= 2.781, p=0.1004 Hence the normality assumption of
the t-test has not been met for this variable and a non-parametric test was
also carried out. Its results supports the t-test findings U = 427.5, z =
0.3432, p = 0.731.
This suggests there is no significant differences in the mean Upgrade
Sales Productivity Service for operators with a disability compared to
operators without a disability.
.
New Sales Effectiveness
The mean new sales index for operators without a disability was 0.0148,
standard deviation = 0.00755, n = 43 while for those with a disability the
mean was 0.01650 with standard deviation 0.00610, n = 21. The results
from the two groups are summarised in the boxplots in Figure 6.
2143N =
Disability Status
With DisabilityNo Disability
Newsale Effectiveness Index
.04
.03
.02
.01
0.00
-.01
Figure 6. New sale index for groups with and
without a disability
The means were not significantly different, t = 0.873 on 62 df, p =
0.3861. Testing for normality we see for skewness p = 0.814, kurtosis p =

- 28 -
-
0.154 with a combined index 2 = 2.17, p = 0.3373, indicating this
variable satisfies the normality assumption. The test for equality of
variances indicates a no significant difference between the variances of
the two groups, F= 1.899, p=.0.173 Hence the normality assumption of
the t-test have been met. However for consistency with the other analyses
a non-parametric test was also carried out. Its results supports the t-test
findings U = 378.5, z = 1.044, p = 0.297
This suggests there is no significant differences in the mean new sales
index for operators with a disability compared to operators without a
disability.
Multivariate test
On grounds of prudence and completeness it was considered useful to
look at the four efficiency-effectiveness variables in combination using a
multivariate technique. Before doing this the inter-correlations between
the four performance variables were examined. See Figures 7 and 8.
Logon Ratio
Contact Efficiency
Upgrade Effectivenes
Newsale Effectivenes
Figure 7. Plots of the four performance variables
Pearson Correlation

- 29 -
-
Sig. (2-tailed)
N
LOGON CONTACT UPGRADE
NEWSALE
LOGON 1.000
.
64
CONTACT .264 1.000
.035 .
64 64
UPGRADE -.191 -.321 1.000
.130 .010 .
64 64 64
NEWSALE -.059 .099 .558 1.000
.641 .438 .000 .
64 64 64 64
Figure 8. Intercorrelations between the four performance variables
The only significant relationships found were a weak correlation between
Contact and Logon, r = 0.264, p = .035 and a moderate correlation
between Newsale and Upgrade, r = .558, p < 0.0005.
Using four t-tests of performance may have increased (though only
slightly) the chance of making a Type 1 error, So. Hotelling's
multivariate test was also carried out. This test aims to compare the

- 30 -
-
four performance variables in combination across the two groups.
The results were H = 035, F= .514, (df = 4, 59), p= 0.726. The result
added further strong support to the proposition that there was no
difference in the performance of the two groups.Furthermore, the
SPSS output from multivariate test provided results from each
dependent variable used in the test. These also supported the earlier
results that none of the indices were significantly different between
the groups (Logon p=0.249, Contact p=0.947, Upgrade p = 0.709,
Newsale p= 0.386).
In summary, there was no difference between the measured
productivity of workers in efficiency, effectiveness or absenteeism.
Workers with a disability provided significantly longer length of
service.
DISCUSSION
What is a typical call centre?
The current picture of a typical Australian call centre emerges from a
1997 survey of 100 call centre operations(ref) 2:
 75% have operated over 4 years but less than 6 years
 On average, call centre’s had 67 work stations
 Full-time call centre agents took an average of 72 call s per day
with an average talk time of 2.7 minutes.
 The average span of control for team leaders to agents is 1:11
 Average length of full-time employment is 3.3 years. Some firms
have estimated that turnover rates in the first 12 months can be as
high as 30%.
 Absenteeism averages 7 days per year

- 31 -
-
 Each ‘agent’ (as an operator is called) is given an average of 21
days training in the first year and 9 days per year in the second
and third years.
 The average remuneration for a call centre operator is $29,779.
 Two thirds of the average call centre’s operating budget is labour.
 The most important measure of performance in call centres is
productivity as measured by revenue generation as a percentage of
cost and service delivery.
Length of service
This is an extremely significant cost to call centres. Anecdotal
information suggests that in some call centres staff turnover can be as
high as 300% per annum. The all industry benchmark for average
turnover for Australian call centres is 27%4. This contrasts with
Australian Human Resources Institute data that has calculated the average
turnover for 'all occupational groupings' is 19%. This indicates that the
average turnover in the call centre Industry is over 40% higher than the
average turnover of staff.
According to the Hallis 'Staff Turnover in call centre’s Study', the
average cost of turnover per separation is between $10510 and $12046.
These figures include Separation Costs, Replacement Costs and Training
Costs however they do not allow for losses of intellectual capital, the cost
of maintaining training facilities or additional costs incurred prior to
replacement, such as overtime for other staff or opportunity costs.
Variable 2: Absent Days
4 1998 Staff Turnover in call centres Study; Hallis August 1998

- 32 -
-
Unplanned absences are a significant cost to any business. In addition to
the direct cost of lost hours, productivity suffers indirectly as attending
operators must carry extra workload or train or support replacement staff.
This can also result in poor staff moral and inferior customer service.
There may also be financial costs. Overtime may need to be paid to
existing call centre operators or additional operators hired. There are also
likely to be increased administrative costs relating to the hiring,
reassigning and maintaining records of absenteeism and payment.
It is common for organisations to set aside a budget of 3% for
absenteeism that equates to an average of about eight days per year per
employee. There are some difficulties in benchmarking call centre
absenteeism rates, because approximately 75% of call centres offer
permanent part time work to operators and 53% of organisations offer
casual hours. 5 Employing part time and casual operators may result in
benefits, however they create problems with rostering, training, and
information flows. Optus Communications 1997 report on call centre
benchmarks6, calculated the average absenteeism for full-time call centre
operators at 6.9 days per annum. The range was from 5.2 days for those
engaged in manufacturing call centres, through 6.8 for those in the
communications industry, and up to a to a maximum of 7.8 days for those
engaged in the personal service and finance areas.
REFERENCES
5 Hallis call centre Staff Survey 1998 pp12
6 Optus Communications Ltd., Australian call centres, Changing the Face of Business
July 1997

- 33 -
-
±1.96*Std. Err.
±1.00*Std. Err.
Mean
Figure 7b. Box & Whisker Plot Positioned Staff Occupancy
0.013
0.014
0.015
0.016
0.017
0.018
0.019
1. Disabled 2. Non-Disabled

- 34 -
-
APPENDIX ONE Copy of Exhibit 4-3 from Hindle (1997): The Enhanced Paradigm of Entrepreneurial Business Planning
WHERE IT APPLIES : Paradigm Boundaries WHAT MUST BE DONE? : Paradigm Laws HOW TO DO IT? :Paradigm success rules and
instrumentation
PARADIGM BOUNDARIES PARADIGM LAWS PARADIGM SUCCESS
RULES
INSTRUMENTATION
REQUIREMENTS
COMMUNICATIONS
Receivers in general (total audience).
Investors – defined as potential providers of the funds or
resources not currently controlled but needed to achieve
identified plan objectives.
Receivers in particular (sub-audiences).
A tailored version of the plan should be targeted to each
sub-audience distinct enough to warrant a separate
investment offer.
Definition of the sender (business plan writer).
6. An entrepreneurial individual or team seeking
resources required to overcome the factors impeding
growth.
(2) Sophisticated; ie. Having both depth and breadth of
generic business skills as well as all required venture
specific skills.
Encoding laws.
1. Codify the selected strategy as a multi-disciplinary continuum.
2. Integrate the codified strategy as a ‘base case’ scenario.
(Note, obeying this law is intimately linked with the simulation success
r
ule).
Message Content laws.
3. Nominate the intended audience.
4. Identify all major plan objectives, primarily as financial targets.
5. Define the investment offer(s) as an expected ROI.
6. Distinguish the venture’s business concept, distinctive
c
ompetencies and sustainable competitive advantages.
7. Provide comprehensive statements of opportunities and risks.
Feedback Law.
8. Seek and respond to feedback.
(Note, obeying this law is intimately linked with the simulation success
r
ule).
Fundamental Communications Success
Rules.
1. Adapt plan length and depth of
detail to the interest level and stage of
involvement of the target audience.
2. Empower th e plan reader.
3. Create investor confidence by
providing flexible credibility.
Fundamental Communications
Instrument.
A unique, purpose-designed
document – embodying high
standards of literacy and
numeracy – of the minimum
length appropriate to the subject
matter and the target audience’s
information needs.
CONTROL
The fundamental defining circumstance.
Impeded growth.
Entrepreneurship process boundaries.
The nine entrepreneurial process parameters (identified
by Bygrave and Hofer) must apply.
Defined limits of planning as a process.
Planning is strategic programming – not strategy
formulation – (Mintzberg’s definition).
Elaboration Law.
1. Elaborate the selected strategy as a set of sub-plans.
Conversion Law
2. Convert the selected strategy into a differentiated suite of
financial budgets.
3. Re-combine the differentiated budgets into an integrated suite of
financial projections.
Fundamental Control Success Rules
1.
A
nticipate and address the tar
g
et
audience’s due diligence
requirements.
2. Create a value-adding deal
structure
Fundamental Coordinating
and Control instrument.
A comprehensive financial
projection model capable of
enumerating the financial
implications of alternative
scenarios.

- 35 -
-
SIMULATION (Simulation possibilities are unbounded) Adaptive Capacity Law.
1. Be able to answer the audience’s ‘what if’ questions in financial
terms.
(Note, obeying this law is intimately linked with the simulation success
rule)
Fundamental Simulation Success Rule.
6. Employ simulation techniques to
obtain the most plausible ‘base case’
scenario which can withstand rigorous
due diligence investigation.
Fundamental Simulation
Instrument
The same financial projection
model.

- 1 -
-