Table 1.

Demographic, psychological, and behavioral details of Dataset 1a

Nonclinical controls t1Nonclinical controls t2Clinical controls t1Clinical controls t2Psychotic t1Psychotic t2
N352022182317
Ageb (yr)27.77 (6.74)27.9 (6.37)40.91 (13.57)40.1 (13)31.22 (7.28)29.9 (7.83)
Gender18 M, 17 F12 M, 8 F11 M, 11 F8 M, 10 F21 M, 2 F17 M, 0 F
Cognitive measures
    IQc107.5 (11.6)108.6 (10.3)97.4 (13.8)99.8 (10.2)88.1 (12.7)87.8 (14.2)
Delusion proneness
    PDI (total)d54.6 (43.1)43.6 (42.5)87.1 (55.2)64.3 (57.3)138.1 (74.2)96.7 (42.6)
    DSSIe2.3 (4.9)2.9 (5.3)4.8 (4.5)4.5 (5.6)15.2 (6.3)8.1 (6.6)
Diagnosis/symptoms
    Diagnoses16 Depression, 3 anxiety and depression, 3 SAD12 Depression, 3 anxiety and depression, 3 SAD18 Scz, 5 bipolar/schizo-affective13 Scz, 4 bipolar/schizo-affective
    MS affective4.6 (1.7)1.0 (1.2)1.8 (1.5)1.5 (1.3)
    MS positive0.3 (0.8)0 (0)6.0 (2.4)1.4 (1.7)
    MS negative0.7 (1.6)1.8 (3.19)1.3 (2.0)0.9 (1.6)
    MS totalf5.5 (2.6)2.8 (3.39)9.1 (3.76)3.7 (3.9)
Beads task
    Initial certainty (1 bead)g0.58 (0.15)0.59 (0.12)0.68 (0.19)0.63 (0.16)0.76 (0.17)0.68 (0.29)
    Initial certainty (3 beads)h0.65 (0.14)0.67 (0.1)0.69 (0.15)0.64 (0.16)0.78 (0.15)0.74 (0.15)
    Disconfirmatory updatingi−0.06 (0.14)−0.03 (0.13)−0.19 (0.3)−0.11 (0.22)−0.29 (0.33)−0.2 (0.3)
    Final certaintyj0.85 (0.2)0.94 (0.11)0.82 (0.16)0.79 (0.23)0.88 (0.11)0.85 (0.23)
  • aDataset 1 includes measures at both baseline (t1) and follow-up (t2). In Dataset 1, verbal IQ was estimated using the Quick Test (Ammons and Ammons, 1962) and delusion proneness using the Peters Delusion Inventory (PDI) (Peters et al., 1999) and Delusions-Symptoms-States Inventory (DSSI) (Foulds and Bedford, 1975). Symptoms were assessed using the Manchester Scale (MS) (Krawiecka et al., 1977). In the tests below, ‘Scz’ refers to the whole Psychotic group. Results are given for ‘Initial certainty’ using both the measure in the original analysis of Dataset 1 (Peters and Garety, 2006), the mean response to the first three beads (3 beads); in Dataset 2, this had to be the mean response to the first three beads in Sequences B and C and two beads in Sequences A and D (2–3 beads), and using the response to the first bead (1 bead).

  • bAt t1: one-way ANOVA F(2,77) = 13.9, p = 10−5. Tukey's HSD: Scz versus Nonclinical controls diff = 3.45, p(adj) = 0.35; Clinical versus Nonclinical controls diff = 13.1, p(adj) = 10−5; Clinical controls versus Scz diff = 9.69, p(adj) = 0.002. At t2: one-way ANOVA F(2,52) = 8.85, p = 0.0005. Tukey's HSD: Scz versus Nonclinical controls diff = 1.98, p(adj) = 0.8; Clinical versus Nonclinical controls diff = 12.2, p(adj) = 0.0006; Clinical controls versus Scz diff = 10.2, p(adj) = 0.007.

  • cAt t1: one-way ANOVA F(2,75) = 16.2, p = 10−6; Tukey's HSD: Scz versus Nonclinical controls diff = −19.5, p(adj) = 10−6; Clinical versus Nonclinical controls diff = −10.1, p(adj) = 0.011; Clinical controls versus Scz diff = 9.36, p(adj) = 0.043. At t2: one-way ANOVA F(2,51) = 14.5, p = 10−5; Tukey's HSD: Scz versus Nonclinical controls diff = −20.8, p(adj) = 10−5; Clinical versus Nonclinical controls diff = −8.8, p(adj) = 0.057; Clinical controls versus Scz diff = 12, p(adj) = 0.01.

  • dAt t1: one-way ANOVA F(2,68) = 12.6, p = 0.00002; Tukey's HSD: Scz versus Nonclinical controls diff = 83.5, p(adj) = 10−5; Clinical versus Nonclinical controls diff = −32.5, p(adj) = 0.094; Clinical controls versus Scz diff = −51, p(adj) = 0.016. At t2: one-way ANOVA F(2,52) = 4, p = 0.024; Tukey's HSD: Scz versus Nonclinical controls diff = 53.1, p(adj) = 0.018; Clinical versus Nonclinical controls diff = −20.7, p(adj) = 0.5; Clinical controls versus Scz diff = −32.4, p(adj) = 0.22.

  • eAt t1: one-way ANOVA F(2,76) = 43, p = 10−13; Tukey's HSD: Scz versus Nonclinical controls diff = 12.9, p(adj) = 10−10; Clinical versus Nonclinical controls diff = 2.52, p(adj) = 0.19; Clinical controls versus Scz diff = −10.4, p(adj) = 10−8. At t2: one-way ANOVA F(2,51) = 3.7, p = 0.032; Tukey's HSD: Scz versus Nonclinical controls diff = 5.2, p(adj) = 0.026; Clinical versus Nonclinical controls diff = 1.65, p(adj) = 0.66; Clinical controls versus Scz diff = −3.56, p(adj) = 0.18.

  • fAt t1: Welch's t(38.4) = −3.62, p = 0.00086, Cohen's d = 1.1. At t2: Welch's t(17.8) = −2.55, p = 0.02, Cohen's d = 1.0.

  • gAt t1: one-way ANOVA F(2,77) = 8.7, p = 0.0004; Tukey's HSD: Scz versus Nonclinical controls diff = 0.18, p(adj) = 0.0003; Clinical versus Nonclinical controls diff = 0.11, p = 0.06; Clinical controls versus Scz diff = −0.08, p(adj) = 0.25. At t2: one-way ANOVA F(2,52) = 0.9, p = 0.4.

  • hAt t1: one-way ANOVA F(2,77) = 6.2, p = 0.003; Tukey's HSD: Scz versus Nonclinical controls diff = −0.14, p(adj) = 0.002; Clinical versus Nonclinical controls diff = 0.04, p = 0.57; Clinical controls versus Scz diff = −0.096, p(adj) = 0.074. At t2: one-way ANOVA F(2,52) = 2.35, p = 0.11; Tukey's HSD: Scz versus Nonclinical controls diff = 0.07, p(adj) = 0.28; Clinical versus Nonclinical controls diff = −0.03, p = 0.8; Clinical controls versus Scz diff = −0.1, p(adj) = 0.1.

  • iAt t1: one-way ANOVA F(2,77) = 6, p = 0.004; Tukey's HSD: Scz versus Nonclinical controls diff = −0.23, p(adj) = 0.003; Clinical versus Nonclinical controls diff = −0.14, p = 0.13; Clinical controls versus Scz diff = 0.097, p(adj) = 0.41. At t2: one-way ANOVA F(2,52) = 2.9, p = 0.062; Tukey's HSD: Scz versus Nonclinical controls diff = −0.18, p(adj) = 0.049; Clinical versus Nonclinical controls diff = −0.08, p = 0.51; Clinical controls versus Scz diff = 0.098, p(adj) = 0.4.

  • jAt t1: one-way ANOVA F(2,77) = 0.71, p = 0.5. At t2: one-way ANOVA F(2,52) = 2.79, p = 0.07; Tukey's HSD: Scz versus Nonclinical controls diff = −0.082, p(adj) = 0.41; Clinical versus Nonclinical controls diff = −0.15, p = 0.057; Clinical controls versus Scz diff = −0.066, p(adj) = 0.57. As reported previously, there were consistent negative correlations between initial certainty (2–3 beads) and disconfirmatory updating in the clinical controls (baseline: ρ = −0.68, p = 0.0005; follow-up: ρ = −0.75, p = 0.0003) and the nonclinical controls (baseline: ρ = −0.52, p = 0.001; follow-up: ρ = −0.43, p = 0.06), but not in the psychotic group (baseline: ρ = −0.30, p = 0.17; follow-up: ρ = 0.17, p = 0.5). There was no consistent correlation between final certainty and either of the other two measures at either time point (p ≥ 0.1 in 11 of 12 comparisons).