Clinical Research Exam 2 |71
Questions with Solutions
Internal validity - -level of control in designing and conducting a study
limit potential threats to the outcome
- External validity - -extend or generalize conclusions to other
participants/settings
- Threats to internal validity - -history, maturation, testing, instrumentation,
mortality, selection bias
- History threat - -outside influence occurred during the course of the study
outcomes observed at the end of a study might be due to this outside
influence
- Instrumentation threat - -changes in either physical equipment or human
observers between the pretest and posttest
- Maturation threat - -increases in performance due to the participants'
growth and development
increases in performance due to recovery over time
- Selection threat - -participants/groups differ in a systemic way, rather than
in a random way, prior to a study
- Mortality threat - -participants drop out before the end of a study
loss of participants may not be random
- Possible ways an investigator may control threats as part of a research
design - -
- Nonexperimental research - -conditions are pre-existing or inherent
- Survey - -learn more about prevalence of conditions
practice patterns about groups of people, opinions, and attitude
investigator collects a representative sample of population
- Types of survey questions - -yes/no
categorical
rating scale
cumulative
open-ended
, - Survey format - -questionnaire
interview
combination
- Survey face validity - -more informal
having "normal" people review it
- Survey content validity - -more formal
sending it to experts to be reviewed
- Simple random sampling - -identify population and randomly select
members of a population to "sample"
- Systematic sampling - -identify population and then select at a systematic
interval (e.g., interview every 4th person on the list)
- Stratified random sampling - -divides population into divisions/subgroups
and random sampling within a subgroup
- Convenience sampling - -choosing individuals who are easiest to reach
(this is not good)
- Sources of possible bias - -failure to identify all members of population
convenience sampling
doesn't truly represent what we are trying to learn
constituting a sample based on volunteers
- Survey combating threats - -sample size; more = better
maximize the response rate
- Correlation/regression - -examines the relationship between 2+ variables
correlation determines strength of relationship
- Correlation research methods - -physical measurements
naturalistic observations
existing data (archival)
- Regression helps determine _ - -causation
helps predict outcome
- Case study/case report - -single participant (or series of individual reports)
intensive presentation of information--retrospective in nature
no control aspect
- Group comparisons (case-controlled) - -comparing individuals with a
condition/characteristic to a group without same condition/characteristic
Questions with Solutions
Internal validity - -level of control in designing and conducting a study
limit potential threats to the outcome
- External validity - -extend or generalize conclusions to other
participants/settings
- Threats to internal validity - -history, maturation, testing, instrumentation,
mortality, selection bias
- History threat - -outside influence occurred during the course of the study
outcomes observed at the end of a study might be due to this outside
influence
- Instrumentation threat - -changes in either physical equipment or human
observers between the pretest and posttest
- Maturation threat - -increases in performance due to the participants'
growth and development
increases in performance due to recovery over time
- Selection threat - -participants/groups differ in a systemic way, rather than
in a random way, prior to a study
- Mortality threat - -participants drop out before the end of a study
loss of participants may not be random
- Possible ways an investigator may control threats as part of a research
design - -
- Nonexperimental research - -conditions are pre-existing or inherent
- Survey - -learn more about prevalence of conditions
practice patterns about groups of people, opinions, and attitude
investigator collects a representative sample of population
- Types of survey questions - -yes/no
categorical
rating scale
cumulative
open-ended
, - Survey format - -questionnaire
interview
combination
- Survey face validity - -more informal
having "normal" people review it
- Survey content validity - -more formal
sending it to experts to be reviewed
- Simple random sampling - -identify population and randomly select
members of a population to "sample"
- Systematic sampling - -identify population and then select at a systematic
interval (e.g., interview every 4th person on the list)
- Stratified random sampling - -divides population into divisions/subgroups
and random sampling within a subgroup
- Convenience sampling - -choosing individuals who are easiest to reach
(this is not good)
- Sources of possible bias - -failure to identify all members of population
convenience sampling
doesn't truly represent what we are trying to learn
constituting a sample based on volunteers
- Survey combating threats - -sample size; more = better
maximize the response rate
- Correlation/regression - -examines the relationship between 2+ variables
correlation determines strength of relationship
- Correlation research methods - -physical measurements
naturalistic observations
existing data (archival)
- Regression helps determine _ - -causation
helps predict outcome
- Case study/case report - -single participant (or series of individual reports)
intensive presentation of information--retrospective in nature
no control aspect
- Group comparisons (case-controlled) - -comparing individuals with a
condition/characteristic to a group without same condition/characteristic
