1. Subject: Metodología de la investigación y estadística
I.
Dr. Gregorio Gómez Hernández
Group: 3CM8
Castrejón Taboada Ximena
Barroso Pineda Karla Guadalupe
Jiménez García Paloma del Carmen
Roque Garcés Eduardo
Trinidad Silva Erick Francisco
Vega Ovando Fabricio Rafael
Quasi-Experimental Design.
Instituto Politécnico Nacional
Escuela Superior de Medicina
3. Quasi-experimental design
The term "quasi -experimento" refers to
research design or n in which
experimental subjects or groups of study
subjects are not randomized.
Election of groups, in which test a
variable, without any kind of
selection or random or pre-selection process.
4. Intact groups
Set of subjects that in quasi-
experimental designs are not randomly
assigned and are not matched, but were
formed before the experiment
5. Example:
To make an educational experiment, a class can
be arbitrarily divided by alphabetic order or by the
disposition of the seats.
After this selection, the experiment proceeds
much like any other way, with a variable
compared between different groups or while a
period of time.
7. Quasi-experimental Design
This method is very useful to
measure social variables.
The weaknesses inherent
methodology do not weaken data
validity.
8. Advantages:
Provides an approximation to the random
experiment when randomness is not
possible.
It is versatile. As random testing, quasi-
experiments can be used to measure
outcomes at population level or of the
program.
9. For example, a strictly
experimental design implies that mothers
were randomly assigned them to drink
alcohol. This will be illegal because the
possible damage that the study could cause
to the embryos.
So what researchers do is to ask people How
much alcohol did they ingested in their
pregnancies and then assign them to their
respective groups.
10. Disadvantages…
Without a proper random assignment, statistics
tests may be insignificant.
For example, these experimental design do not
consider all pre - existing factors (such as for
mothers: what made them drink alcohol or
not), and do not recognize that the outside
influences to the experiment might have
affected the results.
11. A quasi-experiment
built to analyze the
effects of different
educational programs
in two children
groups, for example,
can generate results
that show that a
program is more
effective than the other.
12. These results do not
stand up to a rigorou
s statistic
analysis because the
researcher also need
to control other
factors that may have
affected the results.
13. One of the children
groups may could have
been a little
more smart or a little
more
motivated to. Without
some form of ramdom
pre-selection, is
dificult to judge the
influence of these
factors.
14. Quasiexperimental
Design
Pre-experimental
Designs
In one group, only
posttest.
With non-equivalent
control group, only
posttest.
One group pretest+
posttest
Quasiexperimental
Designs.
With a control
group
Without a control
group.
Interrupted time
series.
Classification.
15. Pre-experimental designs
They represent the basic modules from which the
rest of the quasi-experimental designs are
configured.
A. Designs single group, only postest.- Lacks
control, so you can not draw causal inferences.
B. Non-equivalent control group, only postest.-
The results are not interpreted in causalities also
without pretest we can not know whether the
differences between groups (posttest) are due to
differential treatment or selection.
C. A single group, pretest + postest.- useful to
suggest hypotheses for future research.
17. With a control group.
A. Control group equivalente.- The allocation
rule groups is not known because you work
with already formed groups, despite this,
who investigates try to select groups as
possible equivalents.
B. Cohortes.- design Cohort: Group of
persons belonging to an institution subject
over a period of time similar circumstances.
It helps to study how a particular event
affects a group (experimental cohort) and
compares it with another that did not live the
event (control group).
18. Without control group.
Sometimes it is not possible to have a control group,
for practical or ethical reasons, as in the medical
treatment.
They have less power to justify causal inferences
designs with control groups.
Design withdrawal treatment.- With the treatment
withdrawal, who is investigating, tries to create
conditions to exercise the function of the control
group.
Repeated treatment design: is available with an
only group in which the research staff introduce,
remove and reinsert the treatment, at different times.
19. Interrupted time series.
Results are observed before and during
treatment, making periodic records. For their
analysis you have to know when treatment is
introduced. If it is effective, subsequent
observations will show a change in the series.
It is often used in:
Social studies.
Educational studies.
Health studies.
Evaluation programs
20. Bibliography
Manterola, Carlos, & Otzen, Tamara. (2015).
Estudios Experimentales 2 Parte: Estudios Cuasi-
Experimentales. International Journal of
Morphology, 33(1), 382-
387. https://dx.doi.org/10.4067/S0717-
95022015000100060