The Research Process
I. EXPLORE:
Formulate the Research Question or Hypothesis to be Tested
- Identify the (dependent) variable of interest.
- State the research question or hypothesis. Generally, you will want to complete the following question: "What are the primary determinants of changes in ...insert the name of your dependent variable here...?" or "What variables cause...insert the name of your dependent variable here... to change?"
- You must legitimize the topic you have chosen by convincingly arguing that yours is an important question. Typically this is accomplished by referring to the number of people affected, the number of dollars involved, or some other indication of the scale of the problem or issue. Possible sources of information include popular media (newspapers, TV, magazines...), interviews with experts, and informal focus groups.
- Two ethical issues must be addressed at this juncture. Are you competent to conduct the research? Is the study ethically acceptable?
- Tentatively state the research question/hypothesis.
|
II. PROPOSE:
Obtain
authorization to do the research
- You must obtain
approval of your topic.
- Your proposal must include these parts:
- Statement of the research question or hypothesis
- Explanation of why the question is important
- Indication of the current state of knowledge about the question as surmised from a formal review of the relevant scholarly literature. Also, insights into the independent (explanatory) variables you intend to investigate.
- Your informal assessment of the feasibility of the project
- Bibliography
|
III. DESIGN:
Explain in detail the steps you intend to undertake in order to answer the research question, and confirm that the topic is doable in the time available to you. Resolve any ethical issues.
- What method of data collection will be used? Will you use published data? Will you design and perform an experiment? Will you simply observe your subjects? Or will you conduct a survey?
- What is the time dimension: Will your data be cross-sectional (data collected at approximately one point in time) or longitudinal (time series) data?
- Your research environment (field setting for a survey, laboratory setting for an experiment, or simply an office and the library) will depend upon which data collection method you adopt.
- If
yours is a statistical study, a strategy to minimize nonsampling error must be developed.
- If you intend to produce
your own data, you will need to explain how individual elements in
your sample will be identified (i.e., sampling design) and how follow-up of nonrespondents will be accomplished.
You must also pilot test and revise your experiment, or pilot test and revise
your survey instrument (questionnaire ).
- If you use
published data you will need to verify that the required data are in
fact available and that your sources are of high quality. You must
also familiarize yourself with the sampling and data collection
designs employed by the publishers of the data, as well as the
precise definitions of variables used by the original authors.
- Determine the sample size required, based on prospective power analysis, the planned precision of the study endpoints, or other methods to assure appropriate scope for statistical analysis.
- Complete a realistic timeline with expected completion dates of the logical research steps (a Gantt chart).
|
IV. CAPTURE AND PREPARE THE DATA
- Take steps to guarantee data quality standards, i.e., the accuracy, consistency,
uniformity and completeness of the data. Train interviewers or other individuals who will actually collect the data; test the
accuracy of data collection software such as that used in on-line surveys;
- Code the data (assign numbers or other symbols to answers so
that responses can be grouped into a limited number of categories). Bear in mind that different statistical techniques are used for analyzing qualitative versus quantitative variables, and nominal versus ordinal data. This editing may be done in the field.
- Record the data.
- Prepare the data for analysis--format it, enter it, verify it, and save it in a file. Adopt some convention for identifying the most current data set, and then make certain that each team member has access to it.
- Make back-up copies.
|
V. ANALYZE: Identify a subset of multiple regression equations that answer the
research question, and then choose the best equation from among the equations in
the subset.
- Justify your choice based on theoretical grounds. Does your equation include all the independent variables that theory suggests should be there?
- Justify your choice based on statistical grounds.
- When these two bases for selection (theory and statistics) are in conflict (point to different equations) explain and defend the compromises you had to make.
|
VI. REPORT: In writing, and orally.
Suggested major components for the written report are:
I. Title Page
II. Executive Summary
III. Table of Contents
IV. Introduction
V. Literature Review
VI. Methodology
VII. Findings
VIII. Conclusions
IX. Appendices
X. Bibliography
Suggested components for the oral report are:
I. Introduction of the team
II. Background and purpose of the research
III. The research question or hypothesis
IV. Current state of knowledge about the question
V. Description of the dependent variable (theoretical and its empirical counterpart)
VI. Description of the independent variables (theoretical and their empirical counterparts)
VII. Best equation, including shortcomings
VIII. Conclusions
IX. Question and answer period
|
|
|
|
