CH 9 - Exam 2 Flashcards
Experimental, randomized control trials, and quasi-experimental design studies are all examples of research designs that use an ___.
intervention
A __ study compares 2 groups but does not use an intervention.
observation
___ variables are variables that can influence the outcome and weaken the link b/t the independent and dependent variable.
cofounding
randomization, crossover, homogeneity, matching, and statistical control are methods to ___ the influence of ___ ___
limit
cofounding variables
withholding critical information to advert bias is known as ___
blinding
Who can be blinded during a study? (4 groups)
- participants
- the people preforming the intervention
- other staff
- data collectors
More blinded groups = ___ bias in the study
less
What are the 2 types of time frames that can be used for collecting data?
- cross sectional
- longitudinal
relative timing includes retrospective studies that look ___, and prospective cohort study that looks ___.
- backwards
- forwards
A retrospective study is a ___ (1° or 2°) study. Why?
2° - b/c it is looking at research that as already been completed (backwards) by someone other than the researcher
What are the key design features (8ish)
- intervention
- type of comparison
- control of cofounding variable
- bias
- blinding
- time frame (data collection)
- relative timing
- location
Cause and Effect : ___ is the independent variable while ___ is the dependent variable
cause = independent effect = dependent
What are the 3 traits / criteria for establishing a casual relationship
- temporal
- relationship
- cofounders
temporal requires a cause to ___ an effect
precede
___ requires as association to be shown b/t cause and effect
relationship
___ requires it to be proven that the effect cannot be caused by a third variable
cofounders
A randomized controlled trial has what 3 things?
- intervention
- a control
- participants are chosen randomly
describe the pretest / post test design
- measure the outcome indicator before = baseline data, and after the intervention (posttest data)
How is randomization done today?
What is another mundane ex of randomization ?
- computers
- pulling names out of a hat
A ___ design is used when you give more than one tx but separate the tx and use randomization to determine the ordering.
crossover
What is a limitation for RTCs?
What is one way to address this
- blinding : have to edu the participants which would mean the researcher and participant would not be blinded to the study
- if edu is given in an asynchronous then the research could still be blinded
What is usual care in regards to the comparison in PICO?
- when you are comparing the therapy in place to a new intervention
An ___ ___ is tested out 2 txs for one outcome to see which one gets better results.
Is this ethical? Why?
What if the results are similar?
- alternative treatment
- yes : bc both txs should benefit the participants
- if similar the study loses power to say one is better than the other
A ___ is a tx expected to have no effect to compare with the intervention
placebo
An attention control condition is when the control group gets ___ but not the ____.
- attention
- intervention
The independent and control groups are receiving edu about weight loss, but only the independent group is given a weight loss supplement. This is an example of ___ ___ ___
attention control condition
___ ___ gives both groups the intervention, but the control groups waits until late to get it.
delayed treatment
What is the most ethical type of comparison to make for PICO?
delayed treatment
The inability to test all interventions in this format, to randomized certain variables, and difficulties getting administrative approval to carry out a study are all examples of what?
limitations of a RCT
A ___ ___ study examines an intervention but does not include randomization.
quasi-experimental
Does a quasi-experimental study include a control group?
What is another name for the control group in this type of study?
- not always
- comparison group AKA: non-equivalent control group
Describe a quasi experimental study?
- the intervention is offered to everyone in a population
- have another similar population that is not offered the intervention
- compare the outcomes after a preset time
- -> will often use pre&post test data for this
If a quasi-experimental study only collects post test data it is ___ (more/less) reliable
less
A ___ ___ design collects data over an extended period, introducing an intervention, then collecting data for a length of time.
What type of study would this increase reliability?
- time series
- quasi-experimental that did not have a comparison group
Which type of study is easier to conduct RTC or quasi?
What is one example of why this may be? (think about participants)
- quasi = easier
- some ppl do not want to be randomized and put into a control group b/c they want the intervention –> quasi avoids this
What are 2 disadvantages of a quasi-experimental study?
- dec generalizability d/t inc bias
- inc chance for cofounding variables to influence the results
A non experimental study is one in which an ___ is not introduced and is AKA an ___ ___
- intervention
- observational study
Correlational research shows ____ but does not prove ___ ___ ___
- association
- cause and effect
A ___ design begins with a presumed cause and looks for a presumed effect.
This is AKA ___ design
- cohort
- prospective
What type of design looks retrospectively, starting with the outcome, and looks for the result?
A case control design
How might a case control design further identify causative variables?
compare individuals with the outcome/effect to those who do not
What type of study observes for prevalence but does not determine a relationship?
descriptive
A ___ ___ study determines if traits have a relationship that is not casual.
descriptive correlational
Researchers measure the incidence hunger AND thirst in patients with new onset diabetes without relating the factors that cause diabetes. What type of study is this?
descriptive correlation
Why are non-experimental studies important?
some interventions cannot be tested experimentally, but the relationships can be tested
Can non-experimental studies be used to support practice change?
no
Non-experimental studies can lead to ___ ___ and ___ ___ but not specific ___.
- other studies
- educational recommendations
- interventions
When is data collected in a cross sectional design?
at 1 point in time
When is data collected during a longitudinal study?
data is collected multiple times over and extended period
happens over a LONG time
Script, homogeneity, matching, and statistical control are all used to ___ the study and help reduce ___
- control
- bias
___ is a method or restricting a study to individual who have cofounding variables and subsequently cancel out the effect of the cofounding variable.
homogeneity
Age is a cofounding variable therefore the researchers restricted participants to only those in a certain age range. This is an example of ___
homogeneity
___ is developing a comparison group with the same cofounding variables as the control.
When might this become difficult?
- matching
- more than 2 variables
___ ___ can be used to account for cofounding variables
statistical programs
What 2 factors relate to statistical validity?
- statistical power
- interventions that are clearly defined and different
Large sample size = ___ (more/less) power
more
For an intervention to be clearly defined it needs to be ___ from the ___
- different
- control
Describe internal validity
the extent it can be inferred that the independent variable causes the outcome
What 5 things can threaten internal validity?
- temporal ambiguity : cause does not occur before the effect
- selection : not randomized
- Hx : outside events that can influence participants cooperation
- maturation : time’s effect on outcome
- attrition : ppl dropping out of study (impair diversity = cofounding variable)
What types of studies are more likely to face threats of internal validity? (2)
- quasi experimental
- correlation
___ ___ concerns the ability to apply the study findings to the general population.
The sample needs to be ___ of the pop
- external validity
- representative
What are 2 things to increase external validity?
- multi-site studies
- systematic reviews
___ ___ involves inference from the particulars of the study to the higher order constructs they are intended to represent
construct validity
What is the biggest weakness to construct validity?
blinding : participants may be aware of the intervention and it may effect the outcome
