Can We Trust the Internet to Measure Psychotic Symptoms?

Table 1

Psychometric properties (in square brackets: results for patient groups [Samples 1 and 2] versus “simulators” [Samples 3a and 3b, see text for explanation]).

Variables	Sample 1 (verified Sz diagnosis; 𝑁 = 33)	Sample 2 (probable but unverified Sz diagnosis; Moritz et al., in press, 𝑁 = 113)	Sample 3a (full sample, simulators, 𝑁 = 121)	Subgroup 3b (distinguished experts only, 𝑛 = 86)	Statistics; post-hoc Bonferroni corrected
Means (weighted score; standard deviations)
CAPE positive (range 1–4)	1.86 (0.49)	1.72 (0.46)	2.57 (0.60)	2.59 (0.62)	𝐹(3,263) = 54.66, 𝑃 < .001; [Samples 3a/b] > Samples 1 and 2
CAPE negative (range 1–4)	2.26 (0.59)	2.28 (0.53)	2.24 (0.54)	2.26 (0.54)	𝐹(3,263) = 0.13, 𝑃 > .9
CAPE depression (range 1–4)	2.38 (0.63)	2.28 (0.50)	2.33 (0.49)	2.33 (0.50)	𝐹(3,263) = 0.44, 𝑃 > .7
CAPE lie scale (range 1–4)	n.a.	1.24 (0.27) [5.3% ≥ 8, 2.6% ≥ 9]*	1.59 (0.63) [23.1% ≥ 8, 15.7% ≥ 9]	1.59 (0.66) [20.9% ≥ 8, 17.4% ≥ 9]	𝐹(2,231) = 14.43, 𝑃 < .001; [Samples 3a/b] > Sample 2 (at least 𝑃< .05)
Internal consistency (Cronbach’s 𝛼)					Correlational differences
CAPE positive	.89	.89	.93	.94	𝑃 > .1
CAPE negative	.90	.89	.90	.89	𝑃 > .1
CAPE depression	.88	.80	.79	.79	Sample 1 > [Samples 2 and 3a/b] (at least 𝑃< .05)
Intercorrelations CAPE subscales					Correlational differences
Positive-negative	.61 (𝑃 < .001)	.35 (𝑃 < .001)	−.08 (𝑃 > .3)	−.09 (𝑃 > .4)	[Samples 1 and 2] > [Samples 3a/b] (at least 𝑃< .05)
Positive-depressed	.71 (𝑃 < .001)	.43 (𝑃 < .001)	.04 (𝑃 > .6)	.03 (𝑃 > .7)	[Samples 1 and 2] > [Samples 3a/b] (at least 𝑃< .05)
Negative-depressed	.78 (𝑃 < .001)	.62 (𝑃 < .001)	.69 (𝑃 < .001)	.67 (𝑃 < .001)	𝑃 > .1

Notes. CAPE: Community Assessment of Psychic Experiences.
*Three subjects with scores of 8 were removed from the final sample.