Lecture 11 Flashcards
When we are unable to randomly assign participants to groups but find an existing group that appears similar to the experimental group we can use ___ ___ ___ design.
Nonequivalent comparison groups
A research design having some but not all of the characteristics of a true experiment is called a ___-___
Quasi-experiment
In most quasi-experimental designs, what is the missing element?
The random assignment of subjects to the control and experimental groups
What design does this represent?
O1 x O2
O1 O2
Quasi-experimental design
True or false: because of not randomly assigning research groups, quasi-experimental design is not as rigorous as experimental designs
True
True or false: Because treatment and control groups may not be comparable at the beginning of a design (baseline), quasi-experimental designs are not as rigorous as experimental designs
True
True or false: quasi-experimental designs are not as feasible or flexible as experimental for social studies
False
What is one way to strengthen the internal validity of the nonequivalent comparison groups design?
The use of multiple pretests
True or false: By using the same pretest at a different time point before intervention begins, we can detect whether one group is already engaged in a change process while the other is not.
True
What is this a representation of?
O1 O2 x O3
O1 O2 O3
Quasi-experiment
Comparison group pre-test/post-test
A ___-___ design is when a single population group is studied over a period during which an intervention takes place. It may be a simple pre-test/post-test design, or several measurements that are made both before and after an intervention.
Time-series
A simple time-series design is also called a ___ ___ design, with no control (comparison) group. It can be a group or a ___ (one person).
single case; case
This is an example of what design?
O1 O2 O3 O4 O5 x O6 O7 O8 O9 O10
A time-series (single case) design
Interrupted time series design
What is the purpose of using more tests in the time series design?
Mitigates threats to internal validity
By adding time-series analysis to the nonequivalent comparison groups, multiple time-series designs are ____ than single time-series designs.
stronger
What design is this an example of?
O1 O2 O3 O4 O5 x O6 O7 O8 O9 O10
O1 O2 O3 O4 O5 O6 O7 O8 O9 O10
Multiple time-series design
A ___-___ study examines the phenomenon by taking a ___-___ of it as one point in time.
cross-sectional; cross-section
True or false: because simple correlations at one point in time do not produce causal inference, using cross-sectional designs may help to rule out the alternative explanations through multivariate statistical procedures
True
True or false: case-control designs are not a type of quasi-experimental design
False
Case-control designs compare groups of cases that have had contrasting outcomes and then collect ____ data about past differences that might explain the differences in outcomes.
retrospective
This is an example of what design:
A study trying to find that people who smoke (the FACTOR) are more likely to be diagnosed with lung cancer (the OUTCOME). The experimental group (CASES) were people with lung cancer, the control group (COMPARISON) were people without lung cancer, and some of each group were smokers. If a larger proportion of the cases with cancer smoke then the control group, than it implies that the hypothesis that smoking cases cancer is valid.
Case-control
What was a great triumph of the case-control study developed through the efforts of many scientists, including Doll, in the 1950s?
Demonstration of the link between smoking and lung-cancer.
What did Doll’s case control study concerning smoking and lung cancer show?
A significant correlation between the two in large case control studies. (Tobacco companies had argued for years that this type of study cannot prove causality. Eventually the results of double-blind prospective studies confirmed the causal link which the case-control studies suggested)
It is now generally believed that smoking accounts for __% of lung cancer deaths in the U.S.
a) 67%
b) 97%
c) 87%
c) 87%
What are the practical pitfalls of carrying out experiments and quasi-experiments in social work agencies? (4)
- FIDELITY of the intervention
- CONTAMINATION of the control condition (the control group and the experiment group interact)
- RESISTANCE to the case assignment protocol
- PROBLEMS in client recruitment and retention
Qualitative methods offer a number of techniques that researchers can use to observe research pitfalls. They are:
- ___ shadowing (following along and observing the intervention practice)
- ___ observation during training or group supervision (to identify differences between the stated intervention and what is actually done)
- ___ conversations with agency staff (to identify compliance problems with research protocols)
- ___ or ___ practitioner-client sessions (assess the fidelity of the intervention)
- ___ logs and ___ logs (to assess the fidelity of the intervention and identify major organizational changes that may impede the compliance with the research protocols
- Ethnographic
- Participatory
- Informal
- Videotaping; audiotaping
- Practitioner; event
True or false: An experimental design or quasi-experimental design is not more likely to derive a more robust estimate of the effect of an intervention that a non-experimental design
False (it is more likely to do so)
True or false: in experimental and quasi-experimental designs we can test the intervention variables one by one. This is not a holistic approach.
True
This is a description of the limitations of what?
X1 ----------- Y X2 ... ... Xn
Testing intervention variables one by one