Transcript Understanding Research Design

Understanding Research Design


Can have confusing terms
Research Methodology


The entire process from question to analysis
Research Design


Clearly defined structures within which the study
is implemented
Is a large blueprint, but must be tailored to the
study and then mapped out in detail
Quantitative Design Concerns







Primary purpose (check question)
Is there a treatment (intervention)
Will the treatment be controlled
Is there a control group (untreated)
Is there a pre or post test (or both)
Is the sample a random sample
Will the sample be a single group or
divided into several groups
Quantitative Design Concerns-2







How many groups will there be
What is the size of each group
Will groups be randomly assigned
Will there be repeated measurements
Will the data be collected cross-sectionally
or over time
Have extraneous variables been identified
and controlled for
What strategies are being used for
comparison of variables or groups
Study Validity

Def: It is an examination of the
approximation of truth or falsity of the
propositions




Statistical Validity
Internal Validity
Construct Validity
External Validity

(Cook and Campbell, 1979)
Statistical Validity Errors

Violate assumptions about the data



Type I and Type II errors
Need for Power Analysis




Nominal, ordinal, interval, ratio data
Predicts the necessary N value
Inappropriate use of certain statistics for the
various types of data
Random irrelevancies in setting
Random heterogeneity of respondents
Numbers and Use of Numbers

Nominal (qualitative)


Ordinal (qualitative)


A scale that is subjective but shows a direction, e.g. pain
scale, cancer staging
Interval (quantitative)


A Named category given a number for convenience, e.g.
males are 1 and females are 2
Numbers where the interval between them is meaningful,
e. g. a temperature
Ratio (quantitative)

Numbers where the ratio to each other has meaning, e. g. a
pulse, heart rate.
Statistical Conclusion Validity
Type I and Type II Errors
Accept the Null Hypothesis Reject the Null Hypothesis
Reality is:
Type I Error
No
Wanted
There is no difference
difference
caused by fishing
Reality is:
There is a
Difference
Type II Error, there is
difference often caused
by a low N value
Wanted
Internal Validity

Definition:
*It is the extent to which the effects
detected in the study are a true
reflection of reality rather than the
result of extraneous variables;
* The independent variable did have an
impact on the dependent variable
Threats to Internal Validity








History: Natural events over time impacting
the subjects
Maturation: A person’s growth in any area
impacting his/her response
Testing effect caused by subjects
remembering previous testing
Instrument reliability of treatment
Selection process (randomized)
Mortality threat
Interaction with subjects
No equalization of treatment
External Validity

Definition:
To provide development of the design
that allows it to be generalized
beyond the sample used in the
study.
Most serious threat is that it can only
be said of the group being studied
Threats to External Validity



Small N
No randomization when it is needed
Special probability sampling





Simple
Stratified
Clustered
Systematic
Single vs. replicated
Quantitative Designs





Experimental (Solomon)
Quasi-experimental
Descriptive
Case Study
Correlational