Week 7 - Quant Research Design & Descriptive Stats Part 1 Flashcards
t/f - correlation does not equal causation
true - Causes of health-related phenomena are usually probabilistic rather than deterministic
what is a counterfactual?
what would have happened to the sample people exposed to a causal factor if they simultaneously were not exposed to the causal factor
define an effect in terms of the counterfactual model
the difference between what actually did happen with the exposure and what have happened without it
what is a good model to keep in mind in designing a study to provide cause-and-effect evidence
counterfactual model
define the criteria for causality
Criteria for establishing a cause-and-effect relationship
what are the three criteria for causality
- temporal
- relationship
- no confounders
what are the three criteria for causality
- temporal
- relationship
- no confounders
what is temporal in terms of criteria for causality
A cause must precede an effect in time. Something had to have happened before the cause/effect
what is an example of temporal in terms of criteria for causality
Ex: If we test the hypothesis that smoking causes lung cancer, we need to show that cancer occurred after smoking commenced
Ex: If we were testing the hypothesis that aspartame causes fetal abnormalities, we would have to demonstrate that the abnormalities did not develop before the mothers’ exposure to aspartame
what is relationship in terms of criteria for causality
There must be an empirical relationship between the presumed cause and the presumed effect
what are examples of relationship in terms of criteria for causality
Ex: Must show an association between smoking and cancer - that is, that a higher percentage of smokers than nonsmokers get lung cancer
Ex: Need to find an association between aspartame consumption and fetal abnormalities, that is, that a higher percentage of aspartame users than nonusers had infants with fetal abnormalities
what are no confounders in terms of criteria for causality
The relationship cannot be explained as being caused by a third variable
what are examples of no confounders for criteria of causality
Ex: If smokers tended to also live in urban environments. There would then be a possibility that the relationship between smoking and lung cancer reflects an underlying causal connection between the environment and lung cancer
Ex: Suppose people who used aspartame tended also to drink more coffee than nonusers of aspartame. It would then be possible that any relationship between maternal aspartame use and fetal abnormalities reflects an underlying causal relationship between a substance in coffee and the abnormalities
causality can be enhanced with what types of study design
- experimental
- quasi-experimental
- non-experimental
what are the types of experimental designs?
- true experimental (RCT)
- pretest-posttest (before-after)
- posttest only (after only)
- factorial
- crossover
what is a true experimental design
Researchers are active agents, not passive observers
what is needed for a true experimental design
- manipulation
- control
- randomization
- blinding
define manipulation in an true experimental design
- The researcher does something to at least some participants - aka there is some type of intervention
- Deliberate variation of the IV (the presumed cause) and observe the effect on the outcome (or end point)
define control in an true experimental design
The researcher introduces controls over the research situation, including devising a counterfactual approximation - usually, a control group that doesn’t receive the intervention
define randomization in a true experimental design
The researcher assigns participants to a control or experimental condition on a random basis
define blinding in a true experimental design
Used to prevent biases stemming from awareness. Involves concealing information from participants, data collectors, care providers, intervention agents, or data analysts to enhance objective and minimize expectation bias
define pretest-posttest as an experimental design
With this design researchers examine change. Involves data collection before the intervention and then at multiple post-intervention points (sometimes called repeated measures designs)
define posttest only as an experimental design
Data on the DV are only collected once - after randomization and completion of the intervention
define factorial as an experimental design
Multiple variables are manipulated simultaneously. Permit us to test not only main effects (effects from experimentally manipulated variables) but also interaction effects (effects from combining treatments)
define crossover as an experimental design
Involves exposing the same people to more than one condition (a type of within-subjects design). Has the advantage of ensuring the highest possible equivalence among participants exposed to different conditions
what is a quasi-experimental design
Involves an intervention but lacks randomization or even a control group
what types of designs are quasi-experimental
- pretest-posttest non-equivalent control group
- after-only non-equivalent control group
- one group pretest-posttest
define pretest-posttest non-equivalent control group in terms of quasi-experimental design
Involves two groups of participants, for whom outcomes are measured before and after the intervention; Comparison group is used instead of control group
what is an example of pretest-posttest non-equivalent control group
Ex: Testing effects of new chair yoga intervention for older people; Offered to everyone at a senior center so randomization is not possible. For comparison you’d look at a different senior center that didn’t have the intervention
what is after-only non-equivalent control group in terms of quasi-experimental designs
Unable to have baseline data but still do a comparison posttest
define one group pretest-posttest in terms of quasi-experimental design
Testing a group before an intervention and then testing that same group right after the intervention
what are the two types of non-experimental (observational designs)
- descriptive
- correlational/relationships
what is a descriptive design in terms of a non-experimental design
purpose = to observe, describe, document aspects of a situation as it naturally occurs, sometimes serves as a starting point for hypothesis generation or theory development
what are the two types of descriptive designs
-descriptive correlational
univariate descriptive
what is a descriptive correlational design
- used to describe relationships among variables rather than to support inferences of causality.
- Studies designed to address dx/assessment questions typically uses this design
what is an example of a descriptive correlational design
Ex: Study examining the relationship between respiratory sx and pain experiences in children and adolescents with sickle cell disease
what is a univariate descriptive design
not necessarily focused on only one variable, looks at the frequency of a behavior or description. Primary purpose is to describe the status of each variable and not to relate them to one another
define prevalence in terms of a univariate descriptive design
Prevalence (cross-sectional): done to estimate the prevalence rate of some condition (a disease or behavior like smoking) at a particular point in time. The researcher takes a “snapshot” of the population at risk to determine the extent to which the condition of interest is present
define incidence studies in terms of univariate descriptive
Incidence studies (longitudinal): estimate the frequency of the developing new cases.
what are the two types of correlational/relationship designs
- retrospective design (case control)
- prospective design (cohort)
what is retrospective design in terms of correlational design
studies in which a phenomenon existing in the present is linked to phenomena that occurred in the past
what is an example of retrospective design
antecedent behaviors or health issues that might’ve led to them developing this condition. Looking at both groups and seeing what’s different about them and what’s not
what is prospective design in terms of correlational design
researchers start with a presumed cause and then go forward in time to the presumed effect…looks at what develops over time, usually a longer study over many years
what is time series design
Data collected over an extended period during which an intervention is introduced
What is the key item that makes experimental or quasi-experimental studies different from non-experimental designs?
A. Demographic variables
B. Intervention or treatment
C. Number of groups
D. Sample size
B. Intervention or treatment
Which design includes the administration of more than one treatment to each subject with the treatments being provided sequentially rather than concurrently?
A. Crossover
B. Descriptive
C. Factorial
D. Time series
A. Crossover
What type of design is a systematic investigation of relationships between two or more variables that is meant to explain relationships between variables instead of establishing cause and effect? A. Randomized clinical trial (RCT) B. Correlational design C. Experimental design D. Quasi-experimental design
B. Correlational design
The aim of a study is to evaluate the short and long term effects of smoking cessation strategies tailored to help pregnant adolescents attain and maintain abstinence from smoking. Specifically, the researchers want to compare differences in short and long term smoking behaviors in three groups (Teen Fresh Start, Teen Fresh Start Plus, and Usual Care control). What design would be most appropriate to use?
A. Case control
B. Crossover
C. Descriptive
D. Experimental
D. Experimental
The aim of a study is to examine the relationship of less restrictive restraints with seclusion usage, average years of nursing experience, and mix of staff nursing degrees. What design would be most appropriate to use?
A. Case Control
B. Crossover
C. Descriptive correlational
D. Randomized controlled trial
C. Descriptive correlational