ch. 7
analytical (epidemiology)
Absolute risk Relative risk Odds ratio Attributable risk
chi-square
Association between nominally scaled variables
judgmental approach
Based on the beliefs or experience of experts
divergent criterion cutoff score
Construct-Related Validity •Cutoff scores can be found using the ____ group method with two groups that are clearly different •The distribution for each group is plotted •The point in the curves where the scores overlap is used as a _____
combination approach
Involves use of multiple available sources -Experts -Prior experience -Empirical data -Norms
case series (observational, epidemiology)
Notes cases at a particular time or place
normative approach
Uses norm-referenced data to set standards with a theoretically accepted criterion
analytical epidemiology
causes and prevention of mortality and morbidity
case-control (observational, epidemiology)
compares known cases of mortality or morbidity with matched non-cases
proportionate mortality or morbidity study (observational, epidemiology)
compares results of a study group to the population
analytical (epidemiology)
convert measures of incidence and prevalence into estimates of risk
incidence (epidemiology)
The number, proportion, rate, or percentage of new cases of mortality and morbidity
prevalence (epidemiology)
The number, proportion, rate, or percentage of total cases of mortality and morbidity
cross-sectional (observational, epidemiology)
takes a snapshot of identifiable groups at one time
epidemiology
the study of the distribution and determinants of health-related states and events in populations and the applications of this study to the control of health problems
criterion-referenced test (CRT)
used to make categorical decisions
kappa (K) interobserver
•Widely used technique that allows for correction for chance agreements --to assess ____ agreement --to examine the agreement between a predictor and a criterion that is nominally scaled •Theoretically ranges from −1 to +1 --Negative values of K imply that the proportions of agreement resulting from chance are greater than those from observed agreements
empirical approach
-Uses an external criterion measure -Cut off scores directly established based on the data available on this external attribute
randomized clinical trial (experimental, epidemiology)
Randomly assigns participants to treatments or exposures
community trial (experimental, epidemiology)
Randomly assigns whole communities to treatments or exposures
phi coefficient
correlation between two dichotomous variables
descriptive epidemiology
describes the frequency and distribution of mortality and morbidity according to time, place, and person
judgemental approach normative approach empirical approach combination approach
developing criterion-referenced standards
cohort (observational, epidemiology)
longitudinal; generally tracks populations long term
limitations of CRT
•Cutoff scores always involve some subjective judgment •Misclassifications can be severe •Participants who attain the cutoff level may not be motivated to continue to improve
proportion of agreement
•Established by adding the proportions in cells that are consistently classified •Ranges from 0 to 1; the higher the value the more consistent (correct) the classifications •Does not account for chance •2 x 2 Contingency Table (Stability)
150 75
•For substantial health benefits, adults should do at least: •___ minutes/wk moderate activity OR •___ minutes/wk of vigorous PA or a equivalent combination
epidemiology
•Fundamental science of public health •May help identify risk factors of mortality and morbidity
advantages of CRT
•Represent specific, desired performance levels that are explicitly linked to a criterion •Absolute standards, independent of the proportion of the population that meets the standard •Achievement is based on reaching the standard, not on competing •Performance is linked to specific outcomes •Specific diagnostic evaluations can be made to improve performance to the criterion level •Participants know exactly what is expected of them
criterion-referenced test (CRT)
•Typically involve nominal data (although may include continuous data as well) •ones involving continuous data use cutoff scores •Constructed to yield measurements that are directly interpretable in terms of specific performance standards •Performance standards are generally specified by defining a class or domain of tasks that should be performed by the participant