Basics of research design
Designing research is similar in many ways to designing buildings in architecture. In architecture, you design an architectural project by creatively suggesting ways of how your project will look like when built. In research, you design a research project by creatively and methodically suggesting ways of how your research project will proceed until it is completed or "built". Research design may be different from architectural design in that the former relies more on substantive information and less on artistic talents and skills.
Designing research is basically deciding on the structure of the exact steps or procedures that you need to take, in order to successfully carry out the research objectives.
In the "wheel" of the research process (see link), designing the research comes in second place following the problem identification stage. To be able to design a research, you need to have established a good research question or have formulated a good problem statement, and preferably have carried out the literature review. Although sometimes you need to draft the design of the research project to check the possibilities of conducting it before actually spending the effort on the literature review. However, conducting the literature review may assist you in revising your research design by overcoming obstacles occurring in the previous research for example.
You need to know and understand the basic components of research to correctly design your research project. These components include variables (types, definitions, relationships between variables, and measuring variables), hypotheses (definition and formulation), and sampling (types and procedures).
Research Variables
A variable is "a name for something that is thought to influence a particular state of being in something else .... a special kind of concept that contains within it a notion of degree or differentiation" (Hoover, 1980, pp.22). Examples of variables are heat, pressure, temperature, age, etc.
Types of variables
Variables come in different types depending upon their role in the research.
The independent variable is the variable that influences another variable (the dependent variable).
The dependent variable is the one influenced by another variable (the independent variable).
Studying the effect of sunshine heat on heat gain in buildings: heat sunshine is the ID.V., heat gain in buildings is the D.V.
Studying the effect of building style on user preferences: building style is the ID.V., user preference is the D.V.
An independent variable in one study can become the dependent variable in another study. Could you think of examples?
Operationalizing variables
Variables need to be clearly defined and in a way that permits some kind of measurement. The operational definition of a variable must retain the meaning of the original variable and allow possible measurement using the available resources. Any variable can be operationalized in different ways depending upon the hypotheses and focus of the study.
Measurement of variables
Variables vary, but they may vary in different ways. Some vary in terms of differences of degree such as temperature or height, others vary in terms of differences in rank or order, others vary in terms of differences in variety such as nationality or religion, others vary in simply whether they exist or not. These different ways of variation represent what is known as "levels of measurements" for variables. The following table summarizes the four levels of measurements (adapted from Hoover, 1980, pp.96).
| Level of measurement | Allowable variable properties | Examples |
| 1. Nominal | classification | ethnicity, race, religion, gender, marital status, occupation, group affiliation |
| 2. Ordinal | classification & order | social class, socioeconomic standing, formal education |
| 3. Interval | classification & order & setting standard units of distance (zero is arbitrary) | Biblical time, temperature |
| 4. Ratio | classification & order & setting standard units of distance & locate absolute (true) zero | income, age, weight, distance |
Research Hypotheses
"A hypothesis proposes a relationship between two or more variables" (Hoover, 1980, pp.31). A hypothesis includes the variables of the study and the expected type of relationship that exists between them. To correctly establish hypotheses for your study, you need to clearly define measurable variables (operationalization of variables), and precisely establish a measurable relationship between the variables.
Hoover (1980, pp.83) summarizes the types of relationships between variables that could occur in a hypothesis.
| Relationship | Meaning |
| 1. Null | No relationship is presumed to exist |
| 2. Inferential / Correlative | A relationship is presumed, but it is a relationship that deals with degrees of influence of one variable on another |
| 3. Direct / Inverse | A specific correlative relationship is presumed in which one variable has a predictable association with another--either one variable increases as the other increases (direct) or one increases while the other decreases (inverse) |
| 4. Causal | Changes in one variable are presumed to result from variations in another |
Sampling Procedures
Sampling is the process of selecting a sample from the population to conduct the study upon. The ultimate goal of any sampling procedure is to ensure that the sample is the best representation of the population therefore avoiding any bias or negative influence on the findings of the research. Random selection is the best way to ensure an unbiased representative sample. Random selection is the basis for probability sampling procedures. When random sampling is not possible, non-probability sampling procedures may suffice. Read more on sampling types. For more on probability sampling check the link. For more on non-probability sampling check the link.
Proposal writing
Once the design of the research is complete, it is often the case that the researcher submits a research proposal that describes the research problem and research design. The purpose of the proposal could be to send it to a funding agency or to an academic institution for approval. To guarantee funds for research or to get approval for continuation on a graduate program for example, the research proposal has to be thorough and complete. Although the exact contents of a proposal may vary according to the requirements of one institution to the other, it often includes the following basic sections:
Title and research team
Abstract
Literature Review (a brief account)
Research objectives (including research problem or questions)
Methods (research methods suggested to be used in the research)
Budget and resources
Time schedule
References
The following items are topics for discussion during this lecture. Students should prepare their thoughts and ideas around these topics.
Designing a research
Hypotheses
Variables: Causal and correlational relationships
Types of data
Qualitative and quantitative research
Zeisel, J. (1984). Inquiry by Design: Tools for Environment-Behavior Research. Cambridge: Cambridge University Press. (Chapter 2, pp.18-31, Research: Concepts, hypotheses, tests).
Humoud (Sec 01 second term 2000-2001) found an internet site about a research methods course on the internet. Pages from the site are actually used in our course but it was a good effort by Humoud to actually locate the site.
The site is easy to read with lots of summary points for many of the issues discussed in our course. So take a look at it and enjoy. Click on the link.
Back to top | Previous lecture | Next lecture
Copyright © 2000 by Hisham S. Gabr. All rights reserved.
This page last revised: 09/05/03
