Transcript Methods and Procedures - University of Hawaii at Manoa

Methods and Procedures
Chapter 9
Distinctions between Methods and
Methodology
Methods - tools or techniques applied in the
research process
Procedures – a way we put the tools and
techniques together in sequence or
combinations to achieve objectives
For methodology, the appropriate methods and
procedures need to be carefully selected and
applied to achieve the objectives and produce
reliable knowledge
Treatment of Methods and Procedures
• Perspective on planning the project and the
written proposal
• Proposal contains planned methods and
procedures
• Final report contains the actual methods and
procedures
• The proposal contained planned methods and
procedures and there may be reason to alter
the plans as the research proceeds.
Organization of the Chapter
1. Purposes of Methods and Procedures
2. A Historical Perspective on Empirical
Methods
3. Models in Economic Research
4. Types of Empirical Methods
5. Data Considerations
6. Procedural Suggestions
7. Summary
1. Purposes of Methods and
Procedures
• To provide the plan, its description and how
the objectives will be achieved
• It is the what, why and how of the research
project
• How: order, steps, model specification,
analytical methods, data collection and
administration and input, how tested, how
results being interpreted etc..
• How does the methods relate to the activities
Methods and Procedures
• Levels of details – depends on the type of work
(extension, industry) and expectations of your
committee
• To specify the approach for testing hypothesis
– Can be both qualitative or quantitative
• Directly addresses the objectives
• Research objectives are derived from the research
problem
• The research methods and procedures are driven by
the problem and objective not the other way around
• But if the objective is to test a new technique then it is
different
Methods and Procedures
• Choice of methods and procedures important
Focus on :
• Identifying meaningful researchable problems
• Specifying appropriate objectives
• Developing appropriate methods and
procedures to achieve those objective
Sample Problem
Clearly, if 1.3 billion Chinese people continue to
use plastic bags on a regular basis, there will
be dire consequences on China’s environment,
as there already is in the major cities of
Beijing, Shanghai, Guangzhou and Tianjin.
Actions must be taken to reverse this
destructive trend, before its impacts become
irreversible.
Sample Objectives
• The objective of this study is to determine
consumer preferences for shopping bags
made from alternative materials and to
determine the tradeoffs among the important
purchasing attributes for the purchaser of
these alternative-material bags.
Specific Objectives
• (1) to evaluate the attributes of shopping bags
which are important to consumers, (2) to
determine the socio-economic demographics
which might affect their buying preferences,
and, (3) to discuss the results and marketing
implications.
Sample Methods And Procedure
• (1) developed a conjoint choice experiment
survey to collect data on consumer
preferences, (2) conducted the survey and
collected data from several markets in Tianjin,
(3) analyzed the data with latent class method
and, (4) made conclusions and examine the
implications.
• http://www.ctahr.hawaii.edu/nrem/staff/dow
nloads/20091006_Formatted.pdf
Historical Perspective on Empirical
Methods
• The experimental method by Mills
• Two experimental techniques
– Method of agreement
– Method of difference
Method of Agreement
• Two classifications– Positive Canon of Agreement -PCA
– Negative Canon of Agreement
• PCA - When multiple occurrences of a given
phenomenon have one condition in common
that condition is regarded as the cause of the
phenomenon
• Relationship established between condition C
and phenomenon X
Negative Canon of Agreement
• Reasons from the absence of both the
conditions and phenomenon
• The experiment without C results in the
absence of phenomenon X
• Provides stronger relationship is the canon of
agreement
Method of Difference
• Combination of the Positive and Negative
Canons of Agreement
• Control case – without C , no X
• Experimental case – with C, presence of X
This is much stronger evidence of relationship
between C and X than the previous Canons
Causation vs Relationship
• Experimental methods do not have the
capability to establish causation
• Our techniques be they observational or
statistical, can establish only associations
• They help us determine, often
probabilistically, whether things are related to
one another
• Evidence of causation is derived by first developing
hypotheses of direction of causation from conceptual
reasoning (theory)
• Then examine for evidence of the expected
relationship.
• If empirical evidence exists to support the
relationships, then it supports the hypothesis of
causation.
• The causative implications come from the conceptual
reasoning (theory) rather than the empirical evidence
• Example: Higher avocado production is hypothesized to
be influenced by higher price
• Research is concerned with isolating and
quantifying effects of individual conditions
• In economics, we try to understand and/or
quantify the effects of the different variables on a
particular economic phenomenon independent
of the effects of the other relevant variable.
• E.g. higher price negatively effects on
consumption – ceteris parabis
• Understanding each is important to understand
the collective influence
• Control of variable in economics is difficult because of
the complexity of systems and phenomena studied
• Necessitate the adoption of statistics to control for the
effects of variables in economics
• Through techniques such as regression analysis, a
means for statistical ‘control’ of other forces was
provided
• The field of econometrics grew from this emphasis
which greatly helped with the advent of high speed
computers and technology
• Econometrics – technique based on economic theory
and applying statistics
Models in Economic Research
• Economic models are abstractions from reality,
developed in whole or part from theory, often
expressed in mathematical format to provide:
– Explanations and predictions
– Discovery
– Description and illustration
• They could be use with or without data
• When the models are constructed with the intent
of estimating structure or parameters, the model
constitutes a form of hypothesis
Purpose of a Model
• Explain relationships or system works
• To identify factors or forces that drive a
phenomenon
• Explain with specificity how those forces act and
interact to cause the phenomenon
• Adaptation of theory to a set of phenomenon
forms a model
• The model can use to predict the direction of
change and how policy instruments can affect
change
Model Specification
P (i) = f (C, M, T, D, A, GE, HI, ED, CO)
P (i) = Probability of choosing product profile A vs. B,
C = Shopping bag cost, taking values of 0.3 CNY, 1.5 CNY, or
3.0 CNY.
M = Types of materials, biodegradable plastics, degradable plastics, paper, and cloth.
T
= Number of reuse times, taking values of 1, 5 and 30.
D = Time it takes for the material to naturally degrade, taking the values of 1.5 month,
3 months, and 100 years.
A = Age group: 16 to 18, 19 to 29, 30 to 39, 40 to 49, 50 and above.
GE = Gender: Male or Female.
HI = Household income group (per month) : <3,000 CNY, 3,000 to 5,000 CNY, and >
5,000 CNY.
ED = Educational attainment group: elementary school diploma, junior high school
diploma, high school diploma, bachelor degree and above.
CO = Plastic bag consumption per week, per household: <10, 10 to 20, and >20.
EXPLAIN RELATIONSHIPS
P (i) = f (C, M, T, D, A, GE, HI, ED, CO)
- ? + - ? +/-
Predict Direction and Policy
Models in Economic Research
• When we merge data and theory in a model
with a particular intent, we build an empirical
model.
• Empirical models can be classified as
econometric, optimization, or simulation
Classification of Empirical Models
• Econometric – stochastic (some probability of
error and positive – from actual data)
• Optimization – normative (from desired
objectives, stochastic (probability distributions
are derived from outside the model or nonstochastic)
• Simulation –positive and non-stochastic
Notes and Caveats
• Optimization and simulation models typically involve no
statistical hypothesis testing but are often the most
effective in cases in which there is no way to observe the
phenomena being studies.
• They are not mutually exclusive. Econometric models can
be used for simulation.
• Models have limitations – means to an end
• “Employment of mathematical methods does not
substitute for rigorous theoretical formulations. Rather, it
itself requires more precise understanding of the constructs
of which the discipline is constituted” Breimyer 1991.
• Economists often mistaken statistical significance for
scientific significance.
Methods
• The Descriptive Method
• Statistical and Econometric Tools
• Simple statistical estimations and determinations of fit
and distribution (t-test, F-test)
• Single equation multivariate statistical analysis
(regression, probit, logit estimated with OLS etc.)
• Structural econometric models (Systems of
simultaneous equations, economy model, 2SLS)
• Time-series methods/models (concentrates on the
behavior of (economic) variables or systems of
variables through time)
Simple Regression
Multiple Regression
Y = a + b1*X1 + b2*X2 + ... + bp*Xp
http://videolectures.net/ssmt09_kittel_mra/
Model Specification
P (i) = f (C, M, T, D, A, GE, HI, ED, CO)
P (i) = Probability of choosing product profile A vs. B,
C = Shopping bag cost, taking values of 0.3 CNY, 1.5 CNY, or
3.0 CNY.
M = Types of materials, biodegradable plastics, degradable plastics, paper, and cloth.
T
= Number of reuse times, taking values of 1, 5 and 30.
D = Time it takes for the material to naturally degrade, taking the values of 1.5 month,
3 months, and 100 years.
A = Age group: 16 to 18, 19 to 29, 30 to 39, 40 to 49, 50 and above.
GE = Gender: Male or Female.
HI = Household income group (per month) : <3,000 CNY, 3,000 to 5,000 CNY, and >
5,000 CNY.
ED = Educational attainment group: elementary school diploma, junior high school
diploma, high school diploma, bachelor degree and above.
CO = Plastic bag consumption per week, per household: <10, 10 to 20, and >20.
EXPLAIN RELATIONSHIPS
P (i) = f (C, M, T, D, A, GE, HI, ED, CO)
- ? + - ? +/-
Parameter Estimates
Results
•
•
•
•
•
•
Class 1 are degradable plastics (+ sign) and non-degradable plastics (-), cost (-) and
degradation period (-). Therefore, Class 1 respondents prefer bags made of
degradable material, lower cost, and less time for the material to degrade
naturally. These signs are expected and significant at the 0.05 or 0.01 levels.
For Class 2, the significant attributes found in this group are degradable (+) and
non-degradable plastics (-), reuse times (+), and degradation period (-). Again, the
signs are expected and they are all significant at the 0.01 level. Cost has the
expected negative correlation in this class, but was not significant. Class 2
respondents prefer degradable plastics and bags that can be used many times, and
do not prefer non-degradable plastic bags that take a long time to degrade. Cost
and bags made of either cloth or paper are not important for this group.
For Class 3, the significant attributes are cloth (+), non-degradable plastics (-),
paper (+), and time it takes to degrade (-). These parameters are all significant at
the 0.05 level.
In Class 4, all parameters except for degradable plastics and paper are significant
and have the expected signs.
Class 5 respondents do not prefer high cost (-). They prefer paper (+), and higher
number of times the bag can be reused (+). These parameters are significant at the
0.05 or 0.01 level.
Operations Research Tools
• Optimization Techniques
– Linear Programming –use for research and
management applications; min or max obj. ftns s.t
some constraints.
– Nonlinear Programming – obj. ftns non-linear
• Simulation Techniques –positivistic, nonstochastic
– Budgeting –enterprise budget (B/C)
– Mathematical Simulation –I/O, CGE to simulate an
outcome
– Probabilistic Simulation- use information on
probabilities as the basis of their simulations.
Financial Analysis
Enterprise Budget
Cash Flow Analysis
Sensitivity Analysis
Enterprise Budget Analysis
Political Risk Insurance = $200 farm/month
Enterprise Budget
The enterprise budget for Humpback grouper shows
positive returns only when the live fish are
transported and sold in Hong Kong at average
wholesale prices
Revenue = $157,500 per cycle/farm
Total cost = $81,588
Net Return = $44, 410 cycle/farm
Net Return to Enterprise = $532,920 cycle
Financial Analysis
Enterprise Budget
Cash Flow Analysis
Sensitivity Analysis
Cash Flow Analysis
Cash Flow Analysis
Minimum Investments:
$140,000/farm
$2.1 million/enterprise (10-year)
IRR of baseline model ($60/kg)
for humpback grouper
=38%
Methods for Financial Analysis
Enterprise Budget
Cash Flow Analysis
Sensitivity Analysis
Sensitivity Analysis:
Market Prices vs. Fish Survival Rates
Baseline farmgate price of $25/kg
Low: $15 & High $35
Baseline wholesale price of $60/kg
Low: $40/kg
& High: $80/kg
Survival Rates between 50% and 90%
Sensitivity Analysis:
Market Prices vs. Fish Survival Rates
Sensitivity Analysis:
Market Prices vs. Fish Survival Rates
Baseline Financial Model
Sensitivity Analysis:
Market Prices vs. Fish Survival Rates
Optimistic!
Sensitivity Analysis:
Market Prices vs. Fish Survival Rates
Investor Confidence
(Marketing Scenario 2)
Data Considerations
• Secondary data – from many different
sources, census, industry, state. Results are no
more reliable than the data with which we
work.
• Primary data – when readily available data
does not exist, the option is to forgo the
research or collect the data via surveys.
– Care is needed to the design and questioning to
avoid biasing responses.
Essential Points of Surveying
• Constrain to what you are researching on
• Point, intent and potential value of the survey
should be made clear to those being surveyed
• Ask in value-free language: factual information
and opinion without connotation of
goodness/badness or right/wrong
• Pretest the survey
Sample Population for Plastic Bag
Study
Two hundred and five surveys were completed
during 11 days from June 10th to June 20th,
2008. Every fifth person was selected to
conduct the face-to-face interview. As almost
everyone has experience using a shopping
bag and has basic knowledge of the different
bag materials, it was not difficult to explain
our experiment and administer the survey.
Procedure Suggestions
• Recognize that they are planned and procedures
• Approach the methods and procedures in both
general and specific terms
• Be as specific and detailed as possible in the
specifications and explanations, but recognize
that all specific details cannot be worked out
before the analytical work begins
• Recognize the audience for which the methods
and procedures are written
Summary
• Distinguish between methods and procedures
• Addresses what needs to be done, how it is
done and why
• Addresses data concerns, use, and availability
• Identifies analytical technique and how data
and various analytical tools will combined to
achieve objectives
• Concern with testing hypotheses
Summary cont.
• Types of empirical methods
• Role and function of models
• Data collection, accuracy and sources