0% found this document useful (0 votes)
26 views26 pages

Experimental Methods

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
26 views26 pages

Experimental Methods

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 26

Experimental Methods

from Evidence Based Design


by Kopec, Sinclair and Mathes
Methods
• Research Question is clear
– Satisfies the empirical criterion
– We know what data is needed
• We can focus on how to collect the data,
and
• We know what to do with the data once
collected
Three main questions
1. Design
2. Data Collection, and
3. Data analysis
Quantitative, Qualitative or Both

Quantitative Qualitative
• Experimental comparison • an ethnographic case
based on intervention or study, focusing on
treatment interpretations,
meanings and the
cultural significance of
some behaviour
Mixed
• perhaps a quantitative survey is to be followed by
qualitative interviewing
What really?
Ask yourself these:
1. What data does the research question need? Re-write if
necessary.
2. Which one is your research?
1. making standardized comparisons, dimensions, quantifying
relationships between variables or measuring variance =
Quantitative, or
2. is it more interested in studying a phenomenon or situation in detail,
holistically and in context, focusing on interpretations and/or
processes = Qualitative
3. What do literature say about your topic?
4. What are the practical consequences?
5. Which way would we learn more?
6. Which sort of research is more 'my style'?
e a s u r e
To m
or
e a s u r e
o t t o m
n Internal validit
matches the q
y means the m
uestion.
ethod
s e a r c h
R e R e
e s t io n s e a r c h
Qu Design Da t a
Five Main Ideas of Design
The design is the basic plan for a research,
five main questions:
1. What strategy will be followed?
2. Within what framework?
3. From whom will the data be collected?
4. How will the data be collected?
5. How will the data be analysed?
Strategy
• the internal logic or rationale
• the reasoning, or the set of ideas, by
which the study intends to proceed in
order to answer its research questions
Quantitative Strategy
• experiment
• quasi-experiment
• correlation
Qualitative Strategy
• case studies
• ethnography
• grounded theory
Framework
• means conceptual framework
• the things being studied, and their relationship
to each other

• Quantitative designs typically have well


developed prespecified frameworks

• Qualitative designs show much more variability,


(clear prespecified, to 'first approximations', to
no prespecified framework at all)
Framework - possibilities
1. show in a diagram
if predetermined Cause Effect
2. Initial,
approximation or
for modification
3. None yet but will
be unfolded in the
study
Sample
• From whom will the data be collected
• Who or what will be studied
• You cannot study everyone everywhere
doing everything (Miles & Huberman,
1994)
• Plan for sampling rather than
afterthought
• Probabilistic or not
Sample in Quantitative
• Probabilistic or non-probabilistic
• Type of sampling:
• Sample size
Instruments and procedures
Quantitative Qualitative
• measuring instruments • the researcher is seen as
(sound level meters, the primary instrument
watthour meter, for data collection and
thermometers, lux analysis
meters, etc.) • documents, diaries and
• rating scales or journals,
observation schedules
• or develop an instrument
(pretesting)
Experimental Methods

Experimental methods are


highly systematic
approaches to identifying
"cause and effect"
Like many data gathering
relationships
techniques and processes,
experimentation requires
tightly controlled
procedures.
• the experiment can be
replicated repeatedly by
researchers who will
achieve the same results,
thereby increasing the
validity.
However, life itself (human In other words, an
or animal) serves as a experiment involving
confounding or perplexing people is extremely
variable because behaviors difficult to replicate
have only a certain degree because we are all so
of reliability. diverse
Types

• True experiment
• Quasi experiment
• Single-subject
experiment
• Double-blind, placebo-
controlled experiment
True Experiment

-when one type of subject is A simple example: to test the effectiveness of


using the color red on the walls of an exercise
exposed to one type of condition room

1. randomly selects only 10 people from a very


It is used to show a direct cause- specific group of 100 people who have been
and-effect relationship. It is evaluated as the type of person to use an
considered the simplest of all exercise room.
2. Then the researcher would divide the two
experiments groups equally, and have one group go into
the red room and the other go into a white
room.
as one type of subject(s), 3. Once the subjects have spent the
randomly selected, would have predetermined amount of time in the space,
only a slim chance of providing doing a predetermined activity, they leave
the room and have the inside of ·their
extraneous variables in such mouths swabbed to measure the stress
minimal conditions. hormones present in the saliva
Quasi Experiment

A quasi experiment is also referred • An example - a design


to as a natural experiment. In this researcher might conduct - to
scenario, the researcher quite compare subject behavior at a
literally controls nothing. The quasi revolving door in a epartment
experiment is most like an store versus subject behavior at
observational study. The biggest a sliding door in a department
difference is that the subject(s) are store.
not selected-randomly or otherwise.
• By comparing performance and
The subjects are already in the
"box" and their responses to the
operational success, the
stimulus is observed and recorded, researcher can perhaps
then compared with the results of extrapolate the best door
other similar experiments option for a department store
design.
Single Subject Experiment

a single case or subject is • A good example of this would


be investigating what conditions
studied over a long period
various metals require to
acquire a patina (texture and
most like a longitudinal color changes due to
weathering).
study • To scientifically document the
changes, researchers would be
required to maintain a control
group of the same types of
metals to validate their results
as being directly "caused" by
the experimented conditions
Double-blind experiment

two groups are studied and • The control group in this case
no one knows who is being is given the experimental
agent. The placebo-cont
studied rolled group thinks they are
being exposed to the
experimental agent, but
• they are not. The differences
between the two groups are
measured. The double-blind,
placebo-controlled experiment
• is always held to a higher
standard of scientific detail
and accuracy

You might also like