Applied Quantitative Methods MBA course Montenegro Peter Balogh PhD

Applied Quantitative Methods
MBA course Montenegro
Peter Balogh
PhD
[email protected]
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
BASIC DATA OF THE SUBJECT
Name of the subject: Applied Quantitative Methods
Course status: obligatory
Language: English
Subject educator
Name: Dr. Peter Balogh
Title: Associate Professor
Affiliation: University Debrecen
Period: September 2011
Prerequisite: none
Objective of the training: The students became familiar with the use of
quantitative methods in business
Contact education:
Consultation:
Individual assignment:
Total:
Credit:
20 hours
0 hours
85 hours
105 hours
ECTS : 8
• Description of the individual assignment:
– Prepare and present a case study using the quantitative methods
within a working group.
•
•
•
•
•
•
•
Examinations requirements:
Oral: Presentation
Written: Prepare a case study using the quantitative methods
Compulsory literature:
Jon Curwin and Roger Slater: Quantitative Methods for Business
Decisions, Fifth edition,
• Cengage Learning Business Press, ISBN-13: 978-1861525314
•
• Recommended literature:
• David R. Anderson, Dennis J. Sweeney, Thomas A. Williams, Jeffry
D. Camm, Kipp Martin: Quantitative Methods for Business,
Cengage Learning Business Press, (2010) ISBN-13: 978-0-32465175-1
Course Design
• Part lecture, part skills development
– Usually one major topic per day
– Some time devoted to working with statistical
software packages (excel and SPSS)
Course Reading
Jon Curwin and Roger Slater:
Quantitative Methods for
Business Decisions
Statistical Software
• All course examples will use EXCEL
• You can download
the excel files of the course book:
http://www.agr.unideb.hu/~baloghp/Montenegro
Software and Computers
Bring your laptop to class if
applicable.
We will devote class time in many
sessions to working with statistical
software.
I encourage you to sit with anyone
who knows the MS EXCEL
software package when we begin
to use it in class.
Overall Course Goals
• You will have good knowledge of common research
methods used in quantitative research (surveys,
experiments)
• You will understand basic univariate statistics,
bivariate statistics, linear regression and time series
analysis
• You will be able to use the MS EXCEL to conduct
statistical analyses
Description of the contact education I.
•
•
•
•
•
•
•
•
•
Hour 1-4: Quantitative information
The quantitative approach
Managing data
Survey methods
Presentation of data
Hour 5-8: Descriptive statistics
Measures of location
Measures of dispersion
Index numbers
Description of the contact education II.
• Hour 9-11: Measuring uncertainty
• Probability
• Discrete probability distributions
• The normal distribution
•Statistical inference
• Confidence intervals
• Significance testing
• Non-parametric tests
Description of the contact education III.
• Hour 12-15: Relating variables and predicting
outcomes
• Correlation
• Regression
• Multiple regression and correlation
• Time series
1. The quantitative approach
• Quantitative information:
• We can get data quickly, but we need to be sure
that we are working on the right problem and
that the data is valid.
• Data means
– a few recording
– an extensive national or international survey
• An item of data becomes information when it
informs the user.
1. The quantitative approach
• Quantitative information:
• Internet has transformed the flow and
availability of data.
• The ability to manage data, produce information
and work with problems are all seen as and
important business competencies.
1. The quantitative approach
• Quantitative information:
• Desk research:
– First you need checking what work has already been
done.
– Provide information or identify techniques.
– It is always helpful to find a questionnaire that has
been used previous study and may only require
some modification.
1. The quantitative approach
• Quantitative information:
• Managing numbers is an important part of
understanding and solving problems.
• The collecting together of numbers, and other
facts and opinions provides data.
• This data only becomes information when it
informs the user!!
• The quantitative approach is more than just
‘doing sums’.
• It is about making sense of numbers within a
1. The quantitative approach
•
•
•
•
•
1.1 Problem solving
1.2 Methodology
1.3 Models
1.4 Measurement
1.5 Scoring models
1.1 Problem solving
• To understand problems within a context, it can
be useful to work through a number of stages:
• defining (and redefining) the problem,
• searching for information,
• problem description (and again redefinition if
necessary),
• idea generation,
• solution finding and finaly,
• acceptance and implementation.
Problem solving
Make a start
(problem
sensitivity)
Solution finding
Acceptance and
implementation
Define the problem
(should clarify the
problem,
Idea generation
gap)
(brainstorming)
Search for the
information
Problem description
and redefiniation if
necessary
what we have and what we want!!!
1.2 Methodology
•
•
•
•
•
•
Old methods
New methods
Reliability and validity of findings (conclusions)
Was the purpose of the research clear?
Was this research necessary? (desk research)
Was the means of data collection appropriate?
What can we infer?
(-inductive approach
generalization
-deductive approach)
1.3 Models
• 1.3.1 Model abstraction
• 1.3.2 The development of a mathematical
model
• 1.3.3 Models of uncertainty
• 1.3.4 Computer-based modelling
Modelling
Inputs
Transformation
process
Assumptions
Outcomes
1.3 Models
• A model is a representation of real objects or
situations
– A good understanding of the object or situation
– The recognition of all relevant variables
– The understanding of relationships
– The ability to undertake analysis
1.3.1 Model abstraction
• Physical (or iconic)
• Schematic
(organization charts, flowcharts)
• Analogue
(colours on a map: water, forest)
• Symbolic (or mathematical)
(numbers, letters, special
characters, symbols)
Least abstract
Most abstract
1.3.2 The development of a mathematical
model
• A variable is a quantity or characteristic of
interest that is allowed to change within a
particular problem (students’ mathematics
mark, travel time)
• A parameter is fixed for a particular problem.
• An assumption is something we accept to be
true for the model we are working on.
1.3.3 Models of uncertainty
• Deterministic
• Expected value
Stochastic (Probabilistic)
Mean
1.3.4 Computer-based modelling
• Computational
(spreadsheets, ‘what if’)
• Analytical
(mathematical techniques and
manipulation)
• Simulation
(equations and distributions)
• Expert systems
(advising on solution)
Least abstract
Most abstract
1.4 Measurement
• Measurement is about assigning a value or a score
to an observation.
• Measurement is the representation of
– type,
– size or
– quantity by numbers.
• How we work with data will depend on the level of
measurement achieved.
• Measurement can be categorized as:
nominal, ordinal, interval, ratio
1.4 Measurement
• Nominal (or categorical) level of measurement:
• If responses merely classified into a number of
distinct categories, where no order or value.
• The classification of survey respondents on the
basis of
– religious affinity,
– voting behaviour or
– car ownership.
• The numbers assigned give no measure of amount
or importance.
1.4 Measurement
• Nominal (or categorical) level of measurement:
• For data processing convenience, we may code
respondents 0 or 1 (e.g. YES or NO) or
1, 2, 3 (Party X, Party Y, Party Z), but these numbers
do not relate to meaningful origin or to a
meaningful distance.
• We cannot calculate statistics (mean, standard
deviation).
• We can make percentage comparisons (e.g. 30 %
will vote for party X), present data using bar charts
or use more statistical methods (non-parametric
tests).
1.4 Measurement
• Ordinal level of measurement:
• has been achieved when it is possible to rank order
all categories according to some criteria.
• The preferences indicated on a rating scale ranging
from ‘strongly agree’ to ‘strongly disagree’ or the
classification of respondents by social class
(occupational groupings A, B, C1, C2, D, E) are both
common examples where ranking is implied.
• Individuals are often ranked as a result of
performance in sporting events or business
appraisal.
1.4 Measurement
• Ordinal level of measurement:
• In these examples we can position a response or a
respondent but cannot give weight to numerical
differences.
• It is as meaningful to code a five point rating scale
7, 8, 12, 17, 21 as 1, 2, 3, 4, 5 though the latter is
generally expected.
• Only statistics based on order really apply.
1.4 Measurement
• Ordinal level of measurement:
• You will, however, find in market research and
other business applications that the obvious
codings are made (e.g. 1 to 5) and then a host
of computer-derived statistics calculated.
• Many of these statistics can be useful for
descriptive purposes, but you must always be
sure about the type of measurement achieved
and its statistical limitations.
!
1.4 Measurement
• Interval scale:
• is an ordered scale where the differences between
numerical values are meaningful.
• Temperature is a classic example of an interval scale, the
increase on the centigrade scale between 30 and 40 is the
same as the increase between 70 and 80.
• However, the heat cannot be measured in absolute terms
(0 oC does not mean no heat) and it is not possible to say
that 40 oC is twice as hot as 20 oC, but we can say it is
hotter.
• In practice there are few business-related measurements
where the subtlety of the interval scale is of consequence.
1.4 Measurement
• Ratio scale:
• The highest level of measurement,
- which has all the distance properties of the interval scale
and in addition,
- zero represents the abscence of the caracteristic being
measured.
• Distance and time are good examples.
• It is meaningful, for example, to refer to 0 time and 0
distance and refer to one journey taking twice as long as
another journey or
one distance as being twice as long as another distance.
1.4 Measurement
Nominal data
• Nina and Ravinder own cars, Kristain does not own car
Ordinal data
• In local talent contest, Nina came second, Ravinder
third and Kristain sixth
Interval data
• In a numeracy test Nina got 55, Ravinder got 45 and
Kristain got 90
Ratio data
• In a survey of travel time, Nina took 25 minutes,
Ravinder took 13 minutes and Kristain took 18 minutes
1.4 Measurement
• In summary, it is considered more powerful to
achieve measurement at higher level as this will
contain more discriminating information;
• it is more useful to know how many cigarettes a
respondent smokes on average (0 or more) than
just whether they smoke or not.
• The measurement sought will depend on the
purpose of the research.
1.4 Measurement
• Another useful system of classification is whether
measurement is discrete or continuous.
• Measurement is discrete if the numerical value is
the consequence of counting. (the number of
respondets, the number of companies)
• Continuous measurement can take any value within
a continuum, limited only by the precision of the
measurement instrument. (5 seconds or 5.17
seconds)
Quantitative and Qualitative Perspectives
• "There's no such thing as
qualitative data. Everything is
either 1 or 0“
– Fred Kerlinger
• "All research ultimately has
a qualitative grounding“
– Donald Campbell
Quantitative and Qualitative Perspectives
• First it is useful to distinguish between the quantitative
and qualitative approaches to problem solving.
• Essentially, the quantitative approach will describe and
resolve problems using numbers.
• Emphasis will be given to:
– the collection of numerical data,
– the summary of that data and
– the drawing of conclusions from data.
• Measurement is seen as important and factors that
cannot be easily measured, such as attitudes and
perceptions, are generally difficult to include in the
analysis.
Quantitative and Qualitative Perspectives
• Qualitative approaches describe the behaviour of
people individually, in groups or in organisations.
• Description is difficult in numerical terms and is likely to
use illustrative examples, generalization and case
studies.
• The qualitative approach can use a variety of methods
such as observation and the written response to
unstructured questions.
• Data may come in the form of script, for example,
transcripts of interviews or observations such as video
recordings.
1.5 Scoring models
• Scoring models provide a way of combining
such information and informing decisionmaking.
– Can provide a useful basis for thinking about the
problem
2. Managing data
2.1 Issues of data collection
2.2
Published
sources
„The truth is out
there somewhere”
2.4 A census or a
survey?
2.3 Internet
sources
2.5 Market research
2.1 Issues of data collection
• The five W’s and H technique:
– Who?, What?, Where?, When?, Why? and How?
• Who? is an important question in any problem. Data
will always relate to a particular group of people or set
of items in time and we use this concept to define the
population we will be working.
• The population is defined as those people or items of
interest.
– Given limited resources, including time, the identification
of the relevant population is essential.
Women
Man
2.1 Issues of data collection (cont.)
• Having decided who, we must then consider
whether we need information on all of them
or just a selection.
• A census is a complete enumeration of all
those people or items of interest (whereas a
sample is just a selection from all those people
or items).
2.1 Issues of data collection (cont.)
• What? data will depend on the purpose of the
research.
• A statistical enquiry may require the collection of
new data, referred to as primary data, or be able
to use existing data, referred to as secondary
data. (Combination of both sources.)
• Sources of primary data include observation,
group discussions and the use of questionnaires.
Collection for a specific project.
– Take a long time to collect,
– Be expensive
2.1 Issues of data collection (cont.)
• Secondary data has been collected for some
other purpose.
• Low cost but may be inadequate for purposes
of the inquiry.
• Example: The impact of a new shopping
center on the local community!!
• First step
Second step
2.1 Issues of data collection (cont.)
• Where? to find the right kind of data when you need it or where to
find the people of interest when you need them is an important
skill.
• In organizational research it is often useful to distinguish between
internal and externally generated data.
Recent sales volume, sales value, number of employees,
expenditure on advertising, expenditure on research
The data generated by national governments, local governments,
chambers of commerce, Internet
You still need to question its validity and reliability.
This stage of searching for data is often referred to desk research.
2.1 Issues of data collection (cont.)
• Why? is seen as part of a questioning
approach that should lead to greater
clarification and a justification of approach.
Useful technique called the why technique .
• Why did you say that?
• Why should that be the case?
• Why use that data?
2.1 Issues of data collection (cont.)
• Chapter 2 and 3 are particularly concerned with the
how?
• Having defined the population of interest and the
purpose of the research, a number of issues will need
to be addressed:
– Whether existing published sources provide sufficient
information
– Whether useful information can be found through an
Internet search
– What type of sampling should be used, if any
– How data should be collected
– How questions should be designed, if required
2.2 Published sources
Office for National Statistics (ONS)
• The Annual Abstract of Statistics
• The Monthly Digest of Statistics
• Financial Statistics
• Economic Trends
The Economic and Labour
Market Review
• Social Trends
• The most comprehensive source of
statistics in the UK, Annual Abstract of
Statistics is a statistical encyclopaedia
including over 10,000 series of data and
covering key aspects of the UK’s
economic, social and industrial life.
• It covers the following areas:
parliamentary elections; international
development; defence; population and
vital statistics; education; labour
market; personal income, expenditure
and wealth; health; social protection;
crime and justice; lifestyles;
environment, housing; transport and
communications; national accounts;
prices; government finance; external
trade and investment; research and
development; agriculture, fisheries and
food; production; banking and
insurance and service industry.
•
•
•
•
•
•
•
Monthly Digest of Statistics (MDS) provides the latest
monthly and quarterly statistics for UK businesses,
economy and society. An important reference work, it is
an indispensable source of statistics containing twenty
chapters of tables updated each month.
Did you know…?
In 2007 the average hourly earning (excluding overtime)
for men was £14.98 but for women it was £12.40.
February 2008 was the sunniest since 1929 for England,
Wales and the UK as a whole. But in NW Scotland some
weather stations recorded over 200 percent average
rainfall.
Wind farms supplied 3.79 terawatt hours of electricity in
2007, the equivalent of 0.33 million tonnes of oil.
The most popular destination of air passengers departing
from the UK between April and June 2008 was Spain at
over 8 million, followed by the USA at almost 5 million.
Almost half of adult internet users in the UK use internet
banking (49 per cent), although the most popular use for
the internet is for sending and receiving e-mails (97 per
cent).
• Financial Statistics is a
monetary compendium of
the UK's key financial and
monetary
statistics.
Published monthly, it
contains data on public
sector finance, central
government revenue and
expenditure,
money
supply and credit, banks
and building societies,
interest and exchange
rates, financial accounts,
capital issues, balance
sheets and balance of
payments.
Publication title: Economic Trends (discontinued)
This monthly compendium of statistics and articles on the UK economy was been
replaced by the Economic and Labour Market Review.
• Economic & Labour Market Review (ELMR) provides
an up-to-date summary of the UK economy and
labour market, bringing statistics to life through
news, reviews and features that highlight the most
recent trends and developments through impartial
commentary and analysis.
• ELMR provides readers with new analysis of major
economic measures including output, expenditure,
prices, income and welfare, employment and pay.
• The journal is an authority on how economic
measurement responds to a changing economy
and to new policy challenges, publishing the latestthinking on how economists and statisticians are
helping to shape international standards.
• ELMR assesses the impact of new statistical
methodologies and highlights the benefits to users
of newly released or improved official statistics.
The journal brings together a broad range of
statistics, commentary and analysis from which
users can draw a comprehensive picture of the UK
economy and labour market.
• An established reference source, Social
Trends draws together the most up-todate social and economic data from a
wide range of government
departments and other organisations.
Data is presented clearly in a
combination of tables, figures and text
providing the ideal tool for researching
life and lifestyles in the UK.
• Each chapter focuses on a different
social policy area: population,
households and families, education
and training, labour market, income
and wealth, expenditure, health, social
protection, crime and justice, housing,
environment, transport, lifestyles and
social participation.
European Statistical Sources
1. Non-governmental sources:
– Mintel Market Intelligence Report (monthly
reports on consumer products)
– Retail Business
– Marketing in Europe
– Kompass Directory (provides information on a
variety of companies)
– NEILSEN reports on markets and shopping
behaviour
– ADMAP (Warc's monthly magazine)
2. European Official Statistics
– ADMAP (Warc's monthly magazine)
Warc is home to thousands of effectiveness case studies
from across the world charting the success strategies of
winning brands in every sector and market.
„Extremely relevant articles by great practitioners in an easy
to assimilate format - just what the overstretched
marketer needs"
Michael Harvey, Global Head of Planning, Diageo
Diageo is the world's leading premium drinks business
with an outstanding collection of beverage alcohol brands
across spirits, beer and wine. These brands include
Johnnie Walker, Crown Royal, J&B, Windsor, Buchanan's
and Bushmills whiskies, Smirnoff, Ciroc and Ketel One
vodkas, Baileys, Captain Morgan, Jose Cuervo, Tanqueray
and Guinness.
European Official Statistics
• Europe in figures - Eurostat yearbook 2010
• Basic figures on the EU
• European Business: Facts and figures - 2009
edition
• Key figures on European business - with a special
feature on SMEs
• Key figures on Europe - 2011 edition
• Forestry in the EU and the world
• Science, technology and innovation in Europe
• Food: from farm to fork statistics
•
Europe in figures – Eurostat yearbook
2010 – presents a comprehensive
selection of statistical data on Europe.
With just over 450 statistical tables,
graphs and maps, the yearbook is a
definitive collection of statistical
information on the European Union.
Most data cover the period 1998-2008
for the European Union and its
Member States, while some indicators
are provided for other countries, such
as candidate countries to the
European Union, members of EFTA,
Japan or the United States. The
yearbook treats the following areas:
the economy; population; health;
education; the labour market; living
conditions and welfare; industry and
services; agriculture, forestry and
fisheries;
trade;
transport;
environment and energy; science and
technology; and Europe’s regions. This
edition’s spotlight chapter covers
national accounts statistics – with a
particular focus on the economic
downturn
observed
during
2008/2009.
• The quarterly series “Basic
figures on the EU” presents
the freshest Eurostat data
on a small number of key
indicators in the economic
and social fields. Each issue
is released during the
second month of each
quarter.
• This publication gives a
comprehensive picture of the
structure, development and
characteristics of European
business and its different
activities: from energy and the
extractive
industries
to
communications, information
services and media. It presents
the latest available statistics
from a wide selection of
statistical sources describing for
each activity: production and
employment;
country
specialisation and regional
distribution; productivity and
profitability; the importance of
small and medium sized
enterprises (SMEs); work-force
characteristics; external trade
etc.
This publication summarises
the main features of
European business and its
different activities in a
concise and simple manner.
The publication is intended
to function as a showcase for
and introduction to the data
available in this field. This
edition includes a special
feature section on SMEs,
which presents an analysis of
the different characteristics
of micro, small, medium and
large enterprises
• Key figures on Europe presents a
selection of statistical data on
Europe. Most data cover the
European Union and its Member
States, while some indicators are
provided for other countries, such
as members of EFTA, candidate
countries to the European Union,
Japan or the United States. The
pocketbook treats the following
areas: economy and finance;
population; health; education and
training; the labour market; living
conditions and social protection;
industry, trade and services;
agriculture, forestry and fisheries;
international trade; transport; the
environment; energy; and science
and technology.
• The International Year of
Forests 2011 is a UN
initiative reinforcing the
message that forests are
vital to the survival and
well-being of mankind. This
Eurostat
publication
supports the UN initiative
by statistically depicting
forests in their various
dimensions.
• This publication presents
information for the EU and
its Member States, as well
as
comparisons
with
countries
that
have
considerable
forest
resources.
• This
publication
draws
a
comprehensive picture of the
Science, Technology and Innovation
activities in the European Union as
carried out by its people,
enterprises and governments. It
provides the reader with statistical
information to appreciate the
evolution and composition of
science and technology in Europe
and its position with regard to its
partners. The pocketbook is divided
into seven chapters among which:
Government budget appropriations
or outlays on Research and
Development
(GBAORD),
R&D
Expenditure,
R&D
Personnel,
Human Resources in Science and
Technology, Innovation, Patents,
High-technology.
• This pocketbook provides the
reader with information on
how the food chain evolves in
Europe; it presents a range of
statistical indicators for each
step of this chain from the
farm to the fork, passing from
production on the farm,
through food processing, to
logistical activities such as
importing, transporting and
distributing, before reaching
the end consumer either
through purchases made in
retail outlets or through the
consumption of food and
drink in cafés, bars and
restaurants.
• Secondary data will often provide a useful
overall description (e.g. economic or social
trend) and inform the collection of primary
data.
• Primary data will add specific detail,
particularly current attitudes and opinions.
2.3 Internet sources
• Office for National Statistics (ONS)
http://www.statistics.gov.uk/hub/index.html
http://www.ons.gov.uk/ons/index.html
• Surveys:
http://www.ons.gov.uk/ons/guide-method/surveys/list-ofsurveys/index.html
• http://www.roughguides.com/
• www.dis.strath.ac.uk/business/
• General Global Marketing Informations:
http://webpages.dcu.ie/~gannonm/Websites%20General%20Global
%20Marketing%20Information.html
EUROSTAT:
http://epp.eurostat.ec.europa.eu/portal/page/portal/eurostat/home
• Essentially, we need to be able to evaluate such
information, and be ready to reject any that is
suspect
• To check that data from the web is
– appropriate
– complete
– without bias
• Big advantage of the Internet:
– the scale of the information, which cannot be
matched by any imaginable traditional library
• Disadvantage:
– the lack of any quality control
2.4 A census or a survey?
• We need to decide whether to include
– all people (item)
– or take a selection
undertake a census
sample
• Census:
– Representative!
Mean, Standard Deviation
– prohibitively expensive,
– prohibitively time consuming,
– limit the possible depth of the enquiry
• Designed survey can provide
– the acceptable quality of results
– at lower costs and
– greater speed
2.4 A census or a survey?
• A census will attempt to include everyone
• providing maximum numbers for analysis
• avoiding conserns about sampling or selection
bias
• Provide a benchmark for research activities
(age, gender, ethnic origin, …..)
• Population census
2.4 A census or a survey?
• Full censuses have taken place in the United
Kingdom every ten years since 1801 with the
exceptions of 1941 (during the Second World War).
• In addition to providing a wealth of interesting
information about aspects of the make-up of the
country, the results of the census play an important
part in the calculation of resource allocation to regional
and local service providers, by governments in the
United Kingdom and European Union levels.
• The last:
27 March 2011
• The next census in United Kingdom will take place in
March 2021
2.4 A census or a survey?
• The term ‘census’ is usually associated with a
government count of the population, but a
census can be any complete count
• An identified population is small
• Local or sub-group level
• A census can be taken of
– all the suppliers to a particular company
– all the schools in the UK
– all the sports shops in Leeds
2.4 A census or a survey?
• A survey is likely to be preferred when:
– it is known as a methodology that works
– cost constraints exit
– time constraints exit
• Most commercial and most governmental
research will based on survey methodology
• If the selected sample is representative and
sufficiently large, then the results will be good
enough for purpose.
2.4 A census or a survey?
• The procedure used to select the sample is particularly
important and this is described by the sample design.
The sample needs to represent the population in such
a way that results from the survey can be used to make
generalizations about the population. We talk about
making an inference from the sample to the
population.
• The concepts of inclusion and exclusion are also
important in sample design.
• General election result
– If you were to ask the next five (10, 100) people you see
how they are likely to vote at the next general election, it is
very unlikely that the answers given would be a guide to
general election result.
2.4. 1 How should we decide sample size?
• The size of the sample required will depend
on the following factors:
– the accuracy required
– the variability of the population
– the detailed required in analysis
2.4.1 How should we decide sample
size?
• If an accuracy of ± 1 % is required rather than
±5% for example, then a larger sample will be
necessary.
• If the average weekly household expenditure on a
particular item is only required to an accuracy of
±£5.00 rather than ±£0.50 then a smaller sample
should be sufficient.
• The important point here is that the user or client
needs to be able to specify such levels of
accuracy.
2.4. 1 How should we decide sample size?
• The variability of the population will also be a
determining factor in the sample size
required.
• In the extreme case where everyone held
exactly the same opinion (no variability
existed) we would only need to ask one
person to make an inference to the population
as a whole.
2.4. 1 How should we decide sample size?
• As views become variable, larger samples are
required.
• If accuracy is also required by subgroup, e.g.
female smokers under 25 years of age, then
we would need to ensure that the sample was
sufficiently large to provide the necessary
number in each of the subgroups.
2.4. 1 How should we decide sample size?
• Since most surveys are not designed to find
out a single piece of information, but the
answers to a whole range of questions, the
determination of sample size can become
extremely complex.
• It has been found that samples of about 1000
give results that are acceptable when
sampling the general population.
2.4. 1 How should we decide sample size?
• 'Gallup and other major organizations use sample sizes
of between 1000 and 1500 because they provide a
solid balance of accuracy against the increased
economic cost of larger and larger samples.
• If Gallup were to - quite expensively - use a sample of
4000 randomly selected adults each time it did its poll,
the increase in accuracy over and beyond a well-done
sample of 1000 would be minimal, and generally
speaking, would not justify the increase in cost.‘
Source: www.gallup.com
2.5 Market research
• Market research is seen as a major industry.
• Data is collected on behalf of a range of organizations,
much of it for business use.
• Data can have considerable commercial value and
access can be limited.
• A variety of methods are used to collect data including
face-to-face interviewing, telephone interviewing, and
group discussions.
• Market research provides information on people’s
preferences, attitudes, likes and dislikes, and can help
companies understand what consumers want.
• National and local government use market research to
provide the data to inform policies on everything from
planning local transport to the provision of efficient
health and social services.
2.5 Market research
• Market research can be directly concerned with a market
(which will need definition) and can provide information
on market size, market trends, market share by brand,
customer characteristics and other factors.
• Aspects of market research include advertising and
promotional research, product research and distribution.
• Market research companies also sell a range of services,
and will frequently undertake research for government,
both national and local, academic projects and not-forprofit organizations.
• The Market Research Society provide a range of useful
information on their website: http://www.mrs.org.uk
2.5 Market research
2.6 Conclusion
• Obtaining and using data as information is an
important part of understanding and solving any
problem.
• There is little doubt about the volume of data
now available, and any search of the Internet can
easily produce reams of computer printout.
• As with all problem solving we need to work
within boundaries that ensure the problem
remains manageable and yet does not exclude
new avenues of enquiry.
• Given the diversity of possible data sources we
need to check that data is appropriate, adequate
and without bias.
2.6 Conclusion
• As discussed, the choice is rarely between
secondary data (existing data) or primary data
(new data that needs to be collected for the
specific purpose).
• Secondary data will help describe and define
the existing problem.
• The examination of secondary data can also
provide guidance on what research methods
work and which don't.
• Primary data will generally be needed to add
specific detail.
2.6 Conclusion
• The purpose of any statistical investigation needs to be
clear.
• A statement that we wish to investigate the
management of change within the organization will
mean different things to different people.
• In this case, we need to be clear about our meaning of
change or changes, 'management' and the general
context.
• Decisions will need to be made on who to include and
who to exclude.
• In all statistical work the definition of population (all
those people or items of interest) is particularly
important.
• If we refer to the workforce, for example, do we mean
only full-time employees, those at a particular location
or those doing a particular job?
2.6 Conclusion
• It is a frequently reported experience that
'desk research' yields some of the information
required but also yields other data of interest
and a wealth of new ideas.
• It is also worth considering how much
research is genuinely original!
• If the purpose of the statistical investigation
requires the collection of original data, then
the sample survey is probably the most widely
used method in business and economics.
2.6 Conclusion
• Once collected, data needs to be collated and
presented (see Chapter 4).
• Available computer hardware and software now
allows data to be stored, manipulated and analysed
with relative ease.
• Many types of computer software are available for
dealing with survey data.
• You could use a standard spreadsheet, such as, Excel
or Lotus-123 to record the answers (in a coded form),
or you could use more specialized software such as
SPSS.
• The choice that you make will depend on the size of
the survey, the resources available and the
sophistication of the analysis necessary.