Research Design Overview 1
This presentation covers material found in Chapters 7 and 8
Sources of Knowledge
Trial and Error
Logical Reasoning (Inductive vs. Deductive)
Sources of Questions
Reading the literature can generate ideas
Applications Professional practice or day to day operations indicate a need or suggest a
Purpose of the research
The purpose of a research endeavor is often stated in one or more Specific Aims and sub-
aims of a ‘research’ proposal.
The most important aspect is to be as specific and clear as possible about the purpose. All
other components of the design will be referred back to the purpose.
Remember the frequentist approach of specifying all aspects of the analysis prior to
collecting the data. Essentially have a detailed plan.
A variable is any characteristic of the items or individuals in the population which varies
among the individuals.
EX. Weight, Height, Religious affiliation, Favorite food, or level of immune response, etc.
Values and Levels
Variables have a typology which relates to their characteristics.
For example length can take on almost any value greater than zero, and is referred to as
A variable like Gender has two discrete values: male or female, and referred to as discrete.
In some situations, particularly for discrete variables the values are sometimes referred to
as levels. Cause Effect
The basis for much research is to uncover cause-effect relationships, or to look for
associations among variables.
The variable that is considered the cause is termed the Independent variable.
The variable which is the effect is called the dependent variable.
There may not be an clearly identified role for variables as cause or effect, leading to a
study of associations among variables.
The essential idea is that changes in one variable are associated with changes in another
variable, but there is not necessarily a causal link.
Association is related to the idea of co-variation. The variables change together in some
The research questions are usually presented as hypotheses. These statements are called
Research Hypotheses. The research hypothesis is a conjecture regarding the
relationships among two or more variables.
Typically a research hypothesis is stated that expresses a specific cause-effect relationship,
or a type of association to be studied.
The Independent Variable can also be referred to as a Cause, Exposure, Risk Factor or
The scientific belief captured in the hypothesis is that changes in the level or value of the
Independent variable will lead to (or cause) changes in the value of the dependent variable.
The Response Variable is the observation relevant to the phenomenon being studied.
The response variable is also called the Dependent Variable or Outcome, or criterion
An effect is a consistent change in the value of the response variable brought about by
changes in value of another variable. The consistency aspect includes magnitude and
the direction of the change.
More on Levels
The levels of the independent variable are referred to as treatments in some designs.
The levels often define the groups to be compared in the research.
Note that changes in the values or level of the Independent variable supposedly cause
changes in the value of the dependent. A research design must include at least two levels of the independent variable to
discern a change in the dependent variable.
Different Study Designs are used to study cause effect vs. association.
That is, some designs like experiments can strongly support cause and effect.
Observational designs support claims of association.
Immunotherapy in Cancer patients
Choice of independent
How to stimulate the immune system?
viral, biochemical or other?
Choice of dependent
biological or clinical?
Internal Validity is the degr