Design
|
I
|
Randomized trial: Comparison of two or more groups or conditions in an experiment with random assignment to group or to sequence of conditions in a repeated-measures design
|
II
|
Non-RCT: Comparison of two or more groups or treatments in a quasi-experiment without randomization to group, condition, or sequence
|
III
|
Non-RCT: Comparison of one group pre- and post-treatment
|
IV
|
Single-subject design: One subject measured at intervals throughout an intervention
|
N
|
A Narratives and case studies
|
Sample size
|
A
|
n ≥ 50 persons per condition or group or observations in a single subject design
|
B
|
n ≥ 20 persons per condition or group or observations in a single subject design
|
C
|
n < 20 persons per condition or group or observations in a single subject design
|
Internal validity
|
1
|
High internal validity: No strong alternate explanation for outcome or other threats to validity, such as attrition, unblinded evaluation, unequal treatment, or spontaneous recovery
|
2
|
Moderate internal validity: No strong alternate explanation for outcome but one or two threats to validity exist
|
3
|
Low internal validity: Does not meet criteria for 1 or 2
|
External validity
|
a
|
High external validity: Participants represent population and treatment represents current practice or has strong theoretical support and the research was done in a natural (home or clinic) setting
|
b
|
b Moderate external validity: Has two of the criteria listed for a
|
c
|
Low external validity: Has one or fewer criteria listed for a
|