Skip to main content

Table 1 Values of simulation parameters

From: Evaluating the impact of calibration of patient-reported outcomes measures on results from randomized clinical trials: a simulation study based on Rasch measurement theory

Characteristics

Parameter

Values

PRO instrument

Number of items J

4, 7, 10

Number of response categories M (response options from 0 to M-1)

3, 5

Item locations (δj) and category thresholds (δjl)  distribution

First archetype:

\({\delta}_j=-0.25+\frac{0.5}{J-1}\left(j-1\right),\kern0.5em j=1\dots J\) (Regularly spaced between − 0.25 + 0.25)

δjl defined as the percentiles from a normal distribution with SD of 2.5

Second archetype:

\({\delta}_j=-1+\frac{2}{J-1}\left(j-1\right),\kern0.5em j=1\dots J\) (Regularly spaced between −1 + 1),

δjl defined as the percentiles from a normal distribution with SD of 1.5

Calibration sample

Sample size Nvalidation

100, 250, 500

Variance

1, 2

Mean of latent trait

0

Trial sample

Sample size Ntrial

50, 100, 200, 500 (per group)

Effect size (Standardized mean difference between groups) γ

0, 0.2, 0.5

Mean of latent trait μ

0, 0.5, 2

Variance within each group

1