Transcript Empirical Work - Architecture and Landscape. UoG.

Types of Research
Aims, Objectives & Hypotheses
Introduction
In this presentation we will:
 Discuss empirical work;
 Consider the role of experience;
 Examine qualitative and quantitative
methods;
 Review the role of hypotheses;
 Examine means of testing hypotheses.

Getting Started
It is all to common for people to believe
that research hasn’t started until data
collection has begun.
 That is incorrect!
 All to often data collection is begun
prematurely - before all relevant theory
and literature has been reviewed.
 This leads to two major problems.

Getting Started
Data is collected which is not relevant to
the research.
 Data that is essential is not collected.
 It is difficult to collect data once, having
to collect a second supplementary set
compounds the difficulty (and takes
time).
 Don’t rush into data collection.

Empirical Work
In its strictest sense empirical work is
concerned with experience gained from
experimentation.
 Experience involves observation,
evaluation, memory and recall.
 Unfortunately experience is not totally
reliable.

Empirical Work
What is important for research is that
the problems of comprehensiveness
and accuracy in relying on experiences
are understood and their implications
noted.
 It is essential therefore that ALL data is
recorded accurately and speedily
(memory plays tricks).

Qualitative Approaches
Tesch (1991) identifies three categories
for the analysis of qualitative data:
 Language based;
 Descriptive/Interpretive;
 Theory-building.

Language based
These methods focus on how the
language is used and what it means.
 Techniques such as discourse analysis
and symbolic interactionism are often
employed.
 Communication flows with
organisations.

Descriptive/Interpretive
Attempts to develop a coherent,
comprehensive view of the subject
material from the perspective of those
who are being researched.
 The Delphi technique is one approach
in which repeated questionnaires are
used to obtain consensus amongst a
group of ‘experts’.

Theory-building
Seeks to develop theory out of data
collected during the study.
 Grounded theory is the best example of
this approach.
 Such ‘open-ended’ studies require
meticulous recording; and constant
scrutiny of data to aid recognition of
themes.

Oakley (1994)
Suggest that ‘qualitative’ is used to
describe research which emerges from
observation of participants.
 Sociological:

– Researcher is part of the population being
studied.
– They can use their knowledge to isolate
themes and develop frameworks.
Oakley (1994) cont....

Anthropological:
– Researcher is not part of the population
being studied;
– No initial understanding or knowledge of
the population;
– Capture all data to enable hypotheses to
emerge.
Qualitative Studies
In summary, much qualitative research
concerns the generation of concepts
through the researcher getting
‘immersed’ in the data collection.
 To aid the analysis of the vast amount
of data generated many advocate the
use of topologies or taxonomies.
 Much analysis is carried out in the field.

Quantitative Approaches
Essentially concern making
measurements by collecting data.
 The approach is built upon previous
work which has established:

– Principles;
– Laws;
– Theories.

Which indicate data requirements.
Quantitative Approaches
The major question to be answered is
HOW should the measurements be
made:
 Examine the hypothesised relationships
in the research model.
 Which in turn are based on a thorough
review theory and literature.

Quantitative Approaches
Throughout quantitative studies the
objective is that the research is
unaffected by the beliefs and values of
the researcher.
 It is OBJECTIVE
 There are three main approaches to
quantitative studies.

Questionnaires/Interviews

Much research in management involves
obtaining answers to questions.
– Questionnaires;
– Interviews;
– Case Studies.

The techniques used depend upon the
scope and depth required for the
research.
Questionnaires/Interviews
Depth
Breadth
Questionnaire
Interview
Case Study
Breadth v Depth in Question-based studies
Experimental studies
Seek to measure the change in one
variable (dependent) caused by small
incremental changes in the other
(independent) variable between a
control group and an experimental
group.
 The cause and effect in direction and
magnitude are measured.

Desk-based studies
Using data collected by others can be
problematic as the data, sampling etc
have not been tailored to the specific
project.
 However such studies are cheap.
 In some cases (macro-economics) it is
the only way of obtaining the data
required.

Triangulation
Many researchers seek to use
triangulation (applying more than one
research methodology) to substantiate
the results of their research.
 However, triangulation is not a
substitute for rigour.

Aims & Objectives
The aim of a research project is a
statement of what the research will do.
 What is to be investigated (qualitative
research).
 What impact(s) the main independent
variables have on the dependent
variable (quantitative research).

Aims & Objectives
The aim is a strategic statement of what
the researcher would like to do if no
constraints exist.
 The aim provides an indication of the
context of the research project.

Aims & Objectives
The objectives are statement within the
aims.
 They can be considered as tactical
statements that, in the light of
constraints, translate the aims into interrelated yet self-sufficient coherent
statements of what the researcher
hopes to achieve.

Aims & Objectives
Objectives specify what will be known at
the end of the project that isn’t known at
its beginning AND has been revealed by
the research.
 In qualitative studies the objectives may
also concern what is being studied.

Hypotheses

Not all research project have
hypotheses.
– Fundamental research;
– Where there is little previous research.

However, for the vast majority of
research it is possible and important to
draw on theory and literature to
formulate hypotheses.
Hypotheses

Where hypotheses are formulated they
act as:
– A focus for the work;
– Describe the relationships to be tested;
– Define the boundary of the project.
Hypotheses

Remember; an hypothesis is a
statement of conjecture which can be
tested rigourously to remove as much
Of the supposition (uncertainty) and
replace it with knowledge (probability)
as possible.
Hypotheses
In testing an hypothesis you should
seek to provide evidence, through the
results of testing, to support or refute
the statement at an appropriate level of
probability (confidence).
 The specification of confidence normally
involves the use of statistics.

Testing Hypotheses
Testing an hypothesis usually involves
collecting data and performing analyses
on a representative sample of a
population.
 Sampling involves the application of
statistics to show that the sample is a
sub-group of the population such that
their means and distributions are the
same.

Testing Hypotheses
In essence samples should reflect the
structure of the population.
 Most populations and samples are
assumed to have a normal distribution
(Bell shaped) and as such exhibit
certain standard properties.
 It is about this assumption that most
(simple) statistical tests are based.

Confidence Levels
From a knowledge of the mean and
standard deviation the area under a
normal distribution can be found.
 Thus for a particular combination of
values we can estimate the level of
confidence which can be placed on the
variability of results.

Confidence Limits

These are known as confidence limits:
– 5% (significant);
– 1% (highly significant);
– 0.1% (very significant).

These confidence limits really come into
play when we are trying to establish
whether there is a difference between 2
distributions.
The Null Hypothesis

The null hypothesis speculates that
there is no difference between the
observations made from two groups, i.e.
that any apparent differences in the
mean and standard deviation can be
explained by sampling error.
Student t Test
The null hypothesis is normally tested
by applying a student t test to the data.
 The t statistic is compared with tables of
the t distribution and level of
significance established.
 In essence the t test establishes
whether the sample means are the
same as the population mean.

Statistics

Those of you who are not familiar with
statistics should consult any standard
text.
Summary

In this presentation we have:
– Examined empirical work;
– Qualitative and quantitative methods;
– Hypotheses;
– Testing hypotheses.

Next Research Methodologies.........