Experimental, Quasiexperimental, Single-Case Research and Internet based experiments And Article Critique discusses various research designs including experimental, quasi-experimental, single-case, and internet-based experiments. Experimental research uses random assignment and manipulation of independent variables to test causation. Quasi-experimental research lacks random assignment. Single-case research examines the effect of an intervention on an individual subject using repeated measures. Internet-based experiments can reach large, diverse samples but have validity issues like self-selection and dropout. The article provides details on the characteristics, strengths, limitations, and standards for each research design.
2. Experimental Research
Definition: Testing an idea to determine whether it influences an outcome or dependent
variable.
Key Characteristics:
Random Assignment: Process of assigning individuals at random to groups or to different
groups
Control over Extraneous Variables: Controlling influences of selection of participants, the
procedures, the statistics, the design likely to affect the outcome.
Pretest-posttest, covariates, Matching Participants, Homogenous samples, Blocking variables
Manipulation of Treatment Conditions: Steps-Identify a treatment variable and its levels or
conditions, manipulate treatment conditions
Outcome measures: Dependent variable that is the presumed effect of the treatment variable.
Group comparisons: Obtaining scores for individuals or groups on the dependent variable
and comparing the means and variance between the groups.
Threats to validity: History(time passes), Maturation, Selection, Mortality, Interaction,
Testing, et.
Cresswel, (2014); Frankel, Wallen & Hyun, (2012)
3. Experimental Research
A ‘true’ experiment includes several key features:
one or more control groups
one or more experimental groups
random allocation to control and experimental groups
pretest of the groups to ensure parity
post-test of the groups to see the effects on the dependent
variable
one or more interventions to the experimental group(s)
isolation, control and manipulation of independent
variables
Cohen, Mannion, & Morrison (2007)
5. Experimental Research
How to Design an Experimental Research
Define your research objectives
Formulate hypotheses: H0 and H1
Set up your research design
Select instruments
Select appropriate levels at which to test your hypotheses
Assign persons to groups randomly
Carry out the experiment meticulously
Analyze the data
Muijs(2004); page:334
6. Experimental Research
True Experimental Designs: 7
Pretest-Posttest Controlled Experimental Group Design
Experiment
group
R(Random O1(Observation) X (treatment)
Assignment)
Control group R
O3
O2
O4
Two control Groups and One Experimental Group
Pretest-Posttest Design
Experiment
R
O1
Control
R
O3
Control
R
X
O2
O4
X
O5
The Posttest Control-Experiment Group Design
Experiment
R
Control
R
X
O1
O2
Cohen, Mannion, & Morrison (2007)
7. Experimental Research
The Posttest Two Experimental Group Designs
Experiment
R
X1
O1
Experiment
R
X2
O2
The Pretest-Posttest Two Experiment Groups Design
Experiment R
O1
X1
O2
Experiment R
O3
X2
O4
Matched Pairs Design
Factorial Design
Low
Receive Health Lecture
Smoking Number
Medium
Receive Health Lecture
Smoking Number
High
Receive Health Lecture
Smoking Number
Low
Receive Standard Lecture
Smoking Number
Medium
Receive Standard Lecture
Smoking Number
High
Receive Standard Lecture
Smoking Number
Cohen, Mannion, & Morrison (2007)
8. Experimental Research
Parametric Design
Poor Readers
Token
Number of Correct Word
Average Readers
Token
Number of Correct Word
Good Readers
Token
Number of Correct Word
Outstanding Readers Token
Number of Correct Word
Control
Number of Correct Word
Repeated Measures Design
G1
O
X1
O
X2
O
X3
O
G2
O
X2
O
X3
O
X1
O
G3
O
X3
O
X1
O
X2
O
G4
O
X2
O
X1
O
X3
O
G5
O
X3
O
X2
O
X1
O
Cohen, Mannion, & Morrison (2007)
9. Experimental Research
Poor Experimental Designs:
One-shot Case Study
X
O
One-Group Pretest-Posttest Design
O
The Static-Group(Non-Equivalent)
X
X
O
O
Comparison Design:
The Static-Group(Non-Equivalent)
Pretest-Posttest Design:
O
O
O
X
O
O
Frankel, Wallen, & Hyun, 2012
10. Experimental Research
True Experimental Designs:
The Randomized Posttest Only Control Group Design
Treatment
R
Control
X
O
R
O
The Randomized Pretest-Posttest Only Control Group Design
Treatment
R
O
Control
R
X
O
O
O
The Randomized Solomon Four Group Design
Treatment
R
O
X
O
Control
R
O
C
O
Treatment
R
X
O
Control
R
C
O
Random Assignment with Matching
Frankel, Wallen, & Hyun, 2012
11. Experimental Research
Single Group Designs
The One-shot Case Study
One group pretest-posttest design
Time series designs
Control Group Design with Random Assignment
Pretest-posttest control group design
Posttest only control group design
One-variable multiple condition design
Gall, Gall, &Borg, 2003
12. Experimental Research
Between Group Designs
True experimental design: (Randomized)Pretest-Posttest design or
Posttest only design
Quasi experimental design: (Un-randomized)Pretest-Posttest design
or Posttest only design
Factorial design
Within Group/Individual Designs
Repeated measures design: Interrupted(One experiment) or
Equivalent (More than one experiment)
Single subject designs: Multiple baseline design or Alternating
treatments
Creswell (2014)
13. Experimental Research
Strengths:
Causality: The best type for testing hypotheses about
cause-and-effect relationships
Manipulation of independent variable
Help to see whether the treatment made difference.
Go beyond description and prediction, beyond the
identification of relationship-what causes them.
Frankel, Wallen & Hyun, (2012)
14. Experimental Research
Limitations: Difficult to
Control some variables
Address all threats
Ethical issues: Control group may be disadvantaged by
not receiving treatment or vice versa.
15. Quasi-experimental
“quasi” means, in essence, “sort of.” = quasiexperiment is a “sort of” experiment.
Definition: A quasi-experiment is a study that includes
a manipulated independent variable but lacks
important controls (e.g., random assignment), or a
study that lacks a manipulated independent variable
but includes important controls. Includes nonrandom
assignment-matching.
More threat to internal validity: maturation selection,
mortality, interaction of selection, history, testing,
instrumentation, regression- Cresswell (2014)
16. Quasi-Experimental Research
How to Design an Experimental Research
Define your research objectives
Formulate hypotheses: H0 and H1
Set up your research design
Select instruments
Select appropriate levels at which to test your hypotheses
Assign persons to groups randomly (only experimental
design)
Carry out the experiment meticulously
Analyze the data
Muijs(2004); page:334
17. Quasi-experimental
Types:
A Pre-experimental Design: The one group pretest-posttest
O1 X O2
A Pre-experimental Design: The one group posttest only design
X O1
A Pre-experimental Design: The posttests only non-equivalent
groups design
A Quasi-experimental design: The pretest-posttest nonequivalent groups design
Experimental
O1 X O2
Comparison
O3
O4
The One Group Time Series
19. Single-Case Research- Definition
Key Features:
Single - one subject
Standard conditions
Repeated
measurement
Effectiveness or
productivity
Three components:
(a) repeated measurement,
(b) baseline phase, and (c)
treatment phase.
alternative to group
designs.
Group designs compare the
performance of one sample
of individuals (e.g., people
who don’t smoke, or rabbits
who don’t have smoke
blown into their cages) with
another (e.g., people who
do smoke, or rabbits who do
have smoke blown into their
cages).
Single-subject designs
compare the performance of
an individual before and
after a specified
intervention.
Alberto& Troutman, 1995;Best& Khan, 1998,Tekin (2002),
20. A-B Design
Regardless of the research design, the line graphs used to illustrate the data
contain a set of common elements.
Dependent measure
Condition identifications
Baseline
8
Praise
7
Frequency of disruptions
Independent variable
Condition change line
6
5
4
3
Ordinate
Data points
Data path
2
Abscissa
1
0
0
1
2 3 4
Unit of time
5
6
7 8
Day
9 10 11 12 13 14 15 16
Measure of time
21. Single-Case Research- Types
A-B-A-B Designs: Reversibility-last experimental control or
no functional relationships
Number of
fulfilled
assignments
and without
token(A) and
treatment
with
tokens(B).
(Choen, Mannion, & Morrison, 2007; Kennedy, 2005)
22. Single-Case Research- Types
B-A-B Designs: an intervention already placed
Sometimes an individual’s behavior is so severe that
the researcher cannot wait to establish a baseline
Or an intervention already placed so researcher must
begin with an intervention. In this case, a B-A-B design
is used. The intervention is followed by a baseline
followed by the intervention.
26. Single-Case Research
Strengths
Researcher can establish a cause-and-effect
relationship between treatment and behavior using
only a single participant
See the effect of a treatment on a single participant
Flexibility – development of the design depends on
participant’s responses
By using comparative designs, compare and
contrast the results of the studies easily
27. Single-Case Research
Limitations
Problem with generalizations since designs use only
one participant
Multiple observations can affect participant’s
responses
Absence of statistical controls and reliance on
visual inspection of the data
28. Internet based experiments
Three data collection method through Internet;
Nonreactive data collection
Online Surveys
Web based experiments (Reips,2002)
29. Internet based experiments
Why?
Speed,
Low cost,
Experimenting around a
clock,
A high degree of automation
of the experiment, a wider
sample.
Large number of
participants
High statistical power
Protection of anonymity
Huge representativeness
There is little evidence in the literature that Internetbased surveys achieve higher response rates, as a general
rule, than conventional surveys
Reips (2002)
30. Internet based experiments
Form of emails to emails-plus-attachments of the questionnaire itself, to
emails directing potential respondents to a web site, or simply to web sites.
Although email surveys tend to attract greater response than web-based
surveys, web-based surveys have the potential to reach greater numbers of
participants
Page layout options should be simple not advanced
Avoid open-ended questions not to distrupt participants attention
Confirming of each item can be difficult for those who have less developed
computer skills.
Keep the introduction to the questionnaire short (no more than one
screen), informative (e.g. of how to move on) and avoiding giving a long list
of instructions.
Keep the response categories close to the question
Cohen,
31. Internet based experiments
Advantages:
Ease access to a large number of
demographically and culturally diverse
participants
Specific participant population
Better generalizability of findings to
population, more settings or situations
Avoidance of time constrains,
organizational problems: scheduling
difficulties, as thousands of participants
may participate simultaneously
Reduction of experimenter effects,
demand characteristics
Cost saving of personnel hours,
equipment, administration
Greater openness of the research
process
Access to the number of
nonparticipants
Ease access for participants
Public control of ethical issues
Highly voluntary participation
High participation: High statistical
power
Detectability of motivational
Reips (2002)
32. Internet based experiments
Disadvantages:
Possible multiple submission:
warning about multiple
submission, blocking using
same IP address, or handing
out passwords-one time
password, participant pool or
online panel, control by
collecting personal
identification, controlling
internal consistency
Self-selection: can be
controlled by using the
multiple site entry technique.
Dropout: Promising
immediate feedback, giving
financial incentives, by
personalization
Misunderstood instructions:
Pretest of materials and
providing the participants with
the opportunity for giving
feedback
The comparative basis is
relatively low.
External validity is limited by
their dependence on computer
Reips, (2002)
33. Internet based experiments
Dillman et al. (1999) three ways to overcome problem of
differential expertise in computer usage:
having the instructions for how to complete the item
next to the item itself at the start of the questionnaire
asking the respondents at the beginning about their
level of computer expertise, and, if they are more
expert, offer omitting instruction part and, if they are
less experienced, directing them to instructions
having a ‘floating window’ that accompanies each
screen and which can be maximized for further
instructions.
Cohen
35. Internet based experiments
16 Standards:
5. Consider multiple site entries
1. Consider to use web-based
6. Run survey both online and
software tool to create survey
2. Pretest the instrument for
clarity of instructions
availability on different
platforms
3. Make a decision about
offline for comparision
7. If dropout is to be avoided use
the warm-up technique
8. Use dropout to determine
whether there is motivational
confounding
advantages out-weigh the
disadvantages
4. Check your web survey for
configuration errors
Reips, (2002)
36. Internet based experiments
16 Standards:
13. Perform consistency checks
9. Use high-hurdle technique,
14. Keep logs
incentive information
10. Ask filter questions at the
15. Report and analyze drop out
rates
beginning of the experiment to
encourage serious and complete 16. The experimental materials
should be kept available on the
responses.
Internet, as they will often give
11. Check for obvious naming of
a much better impression of
files, conditions, passwords
what was done than any verbal
description could convey.
12. Use , if needed to avoid
multiplication, participant tools
or password techniques
Reips, (2002)
37. References
Cohen, L., Manion, L., & Morrison, K. (2013). Research methods in education. Routledge.
Creswell, J. W. (2014). Educational research: Planning, conducting and evaluating, quantitative
and qualitative. Pearson International Edition.
Fraenkel, J. R., Wallen, N. E., & Hyun, H. H. (2012). How to design and evaluate research in
education. McGraw-Hill International Edition.
Gall, M. D., Gall J. P. & Borg, W. R. (2003). Educational research: An introduction. Pearson.
Kennedy, C. H. (2005). Single-case designs for educational research. Financial Times/Prentice
Hall.
Reips U. D. (2002). Theory and techniques of web based experimenting. In B. Batinic,
U.D. Reips, & M. Bosnjak (Eds.) Online Social Sciences. Seattle Hogrefe & Huber.
Reips, U. D. (2002). Standards for Internet-based experimenting. Experimental Psychology
(formerly Zeitschrift für Experimentelle Psychologie), 49(4), 243-256.
Tekin, E. (2000). Karşılaştırmalı tek denekli araştırma modelleri. Özel Eğitim Dergisi,
2(4), 1-12.