Download The Research Process

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
Transcript
The Research Process
Problem Recognition
Problem Structuring
Research Design
Data Collection (Surveys,
Requirements Elicitation, experiments, focus groups etc.)
(PILOT STUDY followed the FULL SCALE study)
Data Analysis, generating, interpreting results
Writing up results and recommendations
Implementation
Problem Recognition/Selecting
the Research Topic
• Personal Interest
• Suggested by Research/Practitioner
Literature
• Emergence of a new technology
• Perceptions of discrepancy between desired
and actual state
• Management Directives and Policies
• Social Concerns/Popular Issues
Conceptual Framewrok
•
•
•
•
Identify Key Concepts
Define the Key Concepts
Operationalise the Concepts
Explore systematic relationship between the
concepts.
Specific Research Questions
Main Considerations:
-Specificity and answerability– can the
questions be answered through research?
- Scale and Scope in relation to needs,
available resources.
- Resource Adequacy in Relation to available
time.
Research Strategy and Design
• Data gathering methods
- Type of method to be used.
- Type of data to be gathered.
- Pilot Study
• Data analysis methods
• Budget and timetable
• Reporting the results
Employee Self-Service (ESS)
Module of PeopleSoft ERP
system (Univ. of Sydney)
• System Development and Testing
completed.
• Need to decide on university-wide roll out
and a strategy doing this.
Reducing Cycle Time for New
Product Development at Bosch
• Average cycle time for new product
development/product redesign was 18
months – need to compress it to 9-12
months
3G Wireless Applications for the
Univ. of Sydney
• 3G wireless technology emerging as the
foundation for mobile applications in a range of
domains.
• The Major Projects Group at the university wants
to:
- Make an assessment of the feasibility and viability
of the technology and the applications it can offer
- Identify potential applications that the uni might
benefit from.
- Develop business cases for these applications
Decision Support System
application for Johnson &
Johnson
• Need to decide on how much to spend on a
variety of special promotions at large retail
outlets of J&J such as Woolworths, Coles.
• Prefer a system solution to the problem.
Primary Data
• Data gathered and assembled for the
specific research project at hand.
• Primary data gathered through observations,
focus groups, experiments, field studies etc.
• Format could be numeric, text, image,
video, sound recordings.
• Source may be internal or external to an
organisation.
Secondary Data
• Secondary data are data collected and assembled
for a purpose other than the project at hand, but
may be useful for the project.
• Source may be internal or external to an
organisation.
• Typical sources include:
Australian Bureau of Statistics, Australian Stock
Exchange, Reserve Bank of Australia, OECD, UN,
National Archives, AC Nielsen (UPC scanner
data), Austrade etc.
Primary Data
Research Methods for collecting Primary Data
• Exploratory: Focus Groups, Pilot Studies.
• Sample surveys
• Experimental studies
Definitions
• Respondent: the person who answers an
interviewer’s questions or the person who
provides answers to written/printed questions in
self-administered surveys.
• Sample survey: indicates that the purpose of
contacting the respondents is to obtain a
representative sample of a target
population;method of data collection based on
responses from a representative sample of
individuals from a population of interest.
Types of Errors in Survey Data
• Random Sampling Error
• Systematic Error (Bias) – arising from some
imperfect aspect of the research design or
errors in the execution of the research.
Systematic Error
• Non-response error
• Self-selection bias
• Response Bias
- Deliberate Falsification
- Unconscious Misrepresentation
- Acquiescence Bias
- Interviewer Bias
- Social Desirability Bias
Types of surveys
• Cross sectional
• Longitudinal
Advantages of Secondary Data
• In some situations, useful for clarification and to
define a research problem more sharply –
exploratory research
• Lower cost of research
• Time saving- data readily available
Disadvantages:
• Data may be outdated
• Units of analysis and measures may not be
appropriate.
• Difficulties in combining multiple sec. Data
sources
• Lack of information to verify the accuracy of data.
Uses of Secondary Data
•
•
•
•
Fact finding
Trends in the economy, markets etc.
Exploratory analyses
Building and testing analytical
(mathematical, econometric, forecasting
etc.) models
Types of Secondary Data
• Internal – generated by the organisation’s
accounting systems
• External, Proprietary – commercial
organisations like IDC, Dow Jones,
Standard and Poors etc. routinely gather
data which can be purchased.
• Other external – Government and other
public agencies
Types of measurement scales
• Nominal data: are measurements that
simply classify the units being measured (
of a sample or the population) into
categories.
Eg. Gender in census data, post code of
residential units, political party affiliation of
individuals, industrial classification code of
businesses.
Types of measurement scales
(contd.)
Ordinal data are measures that enable the
units to be ordered (ranked) with respect to
the variable of interest; no indication of how
much.
Eg. A wine taster’s ranking of 10 wines
Ranking of candidates from a job interview
Types of measurement scales
(contd.)
Interval Data: Measurements that enable the
determination of how much (greater or lesser) the
characteristic being measured is possessed by the
unit than another;
Interval scale subsumes ordinal scale but it also tells
us how far apart the units are with respect to the
characteristic (or attribute) of interest.
Always numerical but there is no knowledge of a
zero point (origin) on the measurement
continuum.
Interval Scale
Examples:
-Measurement of temperatures (in celsius) at
which sample of 30 pieces of heatresistant plastic begins to melt.
- Scores of high school students in a
standardised test
Ratio Scale
• Ratio scale data are data are measurements that
enable the determination of how many times the
attribute or characteristic being measured is
possessed by the unit:
Eg. Sales revenues of 50 firms, bonus payments to
managers, unemployment rates for the past 60
months etc.
Always numerical and the zero point is defined.