Flyingoverseas 版 (精华区)
发信人: kingzhc (朔漠孤鹰,傲立风雪), 信区: Flyingoverseas
标 题: 关于research proposal
发信站: 哈工大紫丁香 (2002年12月26日16:29:46 星期四), 站内信件
一门叫做Research Methodology的课程的复习笔记
过来之后上的一门课
没有专门针对写RP修改,但相信会有一些指导作用
不是八股文,千万别套模版
切之切之
Problem Formulation and Hypothesis
1. The importance of problem definition
2. Identification of research topic
Problem formulation begins with the identification of a research problem tha
t is initially formulated in broad terms and ultimately in the form of a car
efully written research question.
3. Conducting an exploratory study
The purpose of an exploratory study is to gather, refine or test preliminary
ideas,
Types of exploratory studies: experience survey; secondary data research
At the end of exploratory study, you will have a better idea on the likely t
opic or data requirement.
4. Preliminary literature review
The aim is to narrow the broad topic into a researchable problem.
5. Identifying the scope of the research
The aim is to transform the topic into something manageable in terms of the
student’s ability as well as the resource constraints.(In terms of attribut
es, dimensions, time and place etc.)
We need to define the scope of the dependent variable not the independent va
riables.
6. Identifying the research question(RQ) and objectives
The research question(RQ) is the research problem stated in the form of a qu
estion. It focuses on the dependent variables or what we seek to explain.(Ex
amples are shown in chapter 2.6)
The research objectives set out what the study seeks to achieve, often state
d as follows:
to explore, to describe, to determine whether, to recommend
7. Development of hypothesis
7.1 Frameworks and hypothesis
A hypothesis should explicitly state and explain the expected relation betwe
en variable.
7.2 Elements of a theory
Theory links concepts
7.3 Types of theories
7.4 Elements of a hypothesis
The empirical manifestation of a concept is a variable
Hypothesis links variables, a hypothesis is the testable portion of a theory
Types of linkage terms:
Correlation: “is associated with” “is related to” “depends on”
Causality: “affect”“Cause”
7.5 The literature review
At this stage, the review directs attention to hypothesis rather than merely
a broad sweep of the literature. The hypothesis is essentially a summary of
the literature review.
One good method is starting from the dependent variable and the independent
variables
7.6 Criteria for evaluating theories
7.7 Potential errors in hypothesis
Reductionism: A complex event is explained in terms of a small number of fac
tors.
Fallacy of common factors: To falsely attribute the cause to the common fact
or
Ecological fallacy: The aggregate data are used to infer individual behavior
.
Spurious relations: Causality is incorrectly attributed to two correlated va
riables.
Tautology: Circular reasoning
Teleology: The belief that events occurs because they have a particular purp
ose in reaching a final state.
8. Operationalization
Translate the key concepts into measurable variables.
9. Other sources of measurement errors
10. The research proposal
10.1 Title
10.2 Research objectives
10.3 Statement of theoretical and practical implications
10.4 Preliminary literature review
10.5 Research design
10.6 Data collection method
10.7 Data collection plan
10.8 Data processing plan
10.9 Data analysis plan
10.10 Schedule of costs
10.11 Schedule of deadline
10.12 Background of researcher involved
Research Design
1. What is research design
The plan for getting from the research question to the conclusion; The strat
egy for testing the hypothesis to rule out alternative explainations.
The blueprint for data collection and interpretation.
Deal with the logic of scientific inquiry, a strategy for testing the hypoth
esis
A good research design should rule our alternative explanantions
2. Types of research design
Case study: Appropriate for exploratory and interpretive work or where conce
pts are difficult
to operationalize and the study is qualitive.
Surveys: Suitable for descriptive or correlational studies
Experiments: Used for causal research, only if the number of variables is sm
all and
controllable.
Correlational research: In the social science many variables are not within
the control of the
researcher, correlation or regression is an appropriate
technique.
Causal-comparative research: Seeks to explain differences that have occurred
between or
more groups.
Historical research: Explain the past to understand or draw lessons for the
present and future.
3. Case studies
A research strategy involving empirical investigation of a particular phenom
enon.
The purpose is to probe intensively using a small number of cases.
The weakness: Investigation of a single case does not allow for valid genera
lization.
4. Surveys
A systematic method of collecting primary data based on a sample.
The purpose is to capture the main characteristics of the population at any
instant or monitor changes over time.
4.1 Types of surveys
Cross-sectional study(Control for time): Sample study; Census
Longitudinal study(Allow temporal changes to be monitored):
Trend study: Monitor the movement of some characteristic of the popu
lation
Cohort study: (Cohort means A generational group as defined in demog
raphics,
statistics, or market research:) Examine the changes in
more specific
sub-populations.
Panel study: A longitudinal study that uses the same set of individu
als.
4.2 Sampling
Basic sampling concept:
Population: The set of all elements of interest.
Sampling frame: The actual list of elements from which sampling takes pla
ce
Sample: A selected subset of the sampling frame
Element: A member of the population
4.3 Types of samples
Non-probability sample: Not representative of the population and techniqu
e of statistics
inference can not be used.
Further divided to four methods:
Convenience sampling: The criterion for selecting sample element is co
nvenience to
the sampler
Purposive sampling: Draw a sample using judgmental selection procedure
s rather
than the law of probability.
Quota sampling: Similar to a stratified sample except that chance sele
ction is not
used in arriving at the quota.
Snowball sampling: Begin with a few respondents who provide referrals
for
additional
respondents
Probability sample: The elements are selected by chance and there is a kn
own probability
of each element being selected.
Further divided into four methods:
Simple random sampling: Each element has an equal chance of being sele
cted
Systematic sampling: Every kth element is selected from the sampling f
rame after a
random start.
Stratified sampling: Divide the population into homogeneous strata and
selects within
these strata random samples
Cluster sampling: Does away with the need for a large sampling frame b
y selecting
groups rather than element.
4.4 Determination of sample size
4.5 Error in surveys
Non-sampling errors:
Random sampling errors:
5. Experiments
Experimental group
Treatment
Control group
5.1 Classical experimental design
Experimental Group Control group and Treatment all exist
5.2 Quasi-experimental designs
Maybe control group does not exist
Types of Quasi-experimental design:
One-shot posttest only study
One-group pretest-posttest design
Time series design
Intervention analysis
Non-equivalent group design
5.3 Repeat measures
Each individual in a group serve as its own control
5.4 Factorial designs
The experiment design when more than one independent variable is involved
The means and variances of each sample are computed and analyzed using an
alysis of
variance ANOVA
5.5 Threats to validity
6. Correlational research
Examine associations between two or more variables. No attempt is made to co
ntrol or manipulate the variables.
Two or more variables may be related in five ways:
6.1 Correlated X-Y
6.2 Depends on Y=f(X)
6.3 Interdependent Y=f(X) and X=f(Y)
6.4 Cause X?Y
6.5 Mutual Cause X??Y
The correlation or regression by themselves do not imply causation
7. Causal-comparative research
Seek to explain difference between two groups that have occurred.
8. Historical research
Methods of Data Collection
1. Data collection methods
2. Questionnaires and interviews
2.1 Methods of communication
2.2 Questionnaire design
2.2.1 Decide what should be asked
2.2.2 Phrase the questions correctly
2.2.3 Decide on the sequence in which the questions appear
2.2.4 Decide the layout
2.3 Scaling and measurement
Scales are used to assess orders of magnitude and preferences
Types of variable Scales
Discrete
Categorical
Non-metric Nominal
Ordinal
Continuous
Metric Interval
Ratio
2.4 Dealing with a poor response rate
2.5 Example
3. Observation techniques
4. Analysis of past documents
5. Simulation
The Collection and Processing of Data
1. Considerations in data collection
2. Selection of interviewers and observers
3. Selection of equipment
4. Training
5. Scheduling and transport plan
6. Supervision
7. Data Processing
Edit or screen the data to ensure completeness consistency and readability
Code data
8. Exploratory data analysis(EDA)
The numerical, counting and graphical detective work in search of patterns b
efore more complex statistical tools are used for confirmatory data analysis
.
Types of display of data:
Simple frequency distribution table
Cross tabulations
Pie chart
Bar chart
Histogram
Frequency polygon
Ogive
Frequency curve(The Bera-Jarque test for normality---to compare sample coeff
icients of
excess skewness and excess kurtosis with that of a normal dis
tribution)
Graphs (displaying trends and cycles)
Stem and leaf plots
Box plot( Mean , the upper and lower quartiles and the minimum and maximum v
alues)
Level curve( for equation like Z=f(X,Y ))
Isoquant
Standard scores
PARTII
Conclusion and Research Report
1. How to write the conclusion
Highlight the major findings of the study.
A brief summary on why the research was conducted and what aspects of the pr
oblem were investigated.
A statement and full explanation of the outcome of the study.
Comment on the significance of the outcome.
Recommendations relating to the outcome and further study..
2. The research report
The summary or abstract section: Summarize the research problem, objectives,
Hypothesis, Methodology, Principal results and implications.
The first chapter of the body: Introduction giving the background and resear
ch problem.
Including research problem, objective,scope, organization of the report.
The second chapter of the body: Literature review; Hypothesis and Operationa
lization
The third chapter of the body: Research methodology(See Page 370)
The fourth chapter of the body: Data analysis
The fifth chapter of the body: Conclusion and recommendations.
Appendix
3. The writing process
4. Writing form and style
5. The examination process
6. Research ethics
7. Final comments
校粒遥浴。桑桑?
The Research Process
1. Formulation of the research problem
A preliminary literature review to provide the background
Identify the research question
Work out research objectives and proposal
A thorough literature review and formulate the hypothesis
Operationalize relevant concepts for empirical test
2. Determination of research design to test the hypothesis
Common types of research design:
Case studies
Surveys
Experiments
Correlational research (regression)
--
--
世贸中心整塌了,五角大楼喀喳了.白宫门脸挂花了,布什这下抓瞎了.
老窝让人家整惨了,中央情报局现眼了.素质过硬不手软了,拉登大叔露脸了.
巴勒斯坦兴高采烈,纳斯达克一路狂泻.伊斯兰人小样挺倔:"谁再整事叫他毁灭!"
布什该总结经验了,别在地球到处捣乱了.树敌太多你不好办了,以后别惹伊斯兰好汉了.
美国那边闹的挺欢,咱也别老是冷眼旁观.一人抄起一块板砖,跨过海峡解放台湾
※ 来源:·哈工大紫丁香 bbs.hit.edu.cn·[FROM: 202.118.237.58]
Powered by KBS BBS 2.0 (http://dev.kcn.cn)
页面执行时间:206.639毫秒