OSA analysis
1 participants info
LEARNING EXPS:
OSA 2: N = 102 (no exclusion due to no attention checking)
OSA 3: N = 221 (35 exclusion due to failing both attention checkers)
combined included subs N = 323
SURPRISE & EXPLANATION EXPS:
OSA1: N = 242 (qualtrics, half expected and half unexpected, has attention-checking questions)
OSA4: N = 80 (4 key trials and 9 fillers, has checker questions) (2 excluded)
combined included subs N = 322
2 LEARNING
2.1 all trials
OSA 2 & OSA 3
data_curve_analysis(OSA_learning_core, "score")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 1888.8
Scaled residuals:
Min 1Q Median 3Q Max
-1.1921 -0.9877 0.7649 0.9691 1.1736
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.01461 0.1209
Residual 0.23581 0.4856
Number of obs: 1292, groups: subject, 323
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 4.892e-01 2.784e-02 1.275e+03 17.568 <2e-16 ***
violation_number1 1.548e-02 3.821e-02 9.660e+02 0.405 0.685
violation_number2 4.025e-02 3.821e-02 9.660e+02 1.053 0.292
violation_number3 -9.288e-03 3.821e-02 9.660e+02 -0.243 0.808
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.686
viltn_nmbr2 -0.686 0.500
viltn_nmbr3 -0.686 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.489 0.0278 1275 0.420 0.559
1 0.505 0.0278 1275 0.435 0.574
2 0.529 0.0278 1275 0.460 0.599
3 0.480 0.0278 1275 0.410 0.550
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.01548 0.0382 966 -0.405 1.0000
violation_number0 - violation_number2 -0.04025 0.0382 966 -1.053 1.0000
violation_number0 - violation_number3 0.00929 0.0382 966 0.243 1.0000
violation_number1 - violation_number2 -0.02477 0.0382 966 -0.648 1.0000
violation_number1 - violation_number3 0.02477 0.0382 966 0.648 1.0000
violation_number2 - violation_number3 0.04954 0.0382 966 1.296 1.0000
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 312826, p-value = 0.9828
alternative hypothesis: two.sidedOSA2 only
data_curve_analysis(filter(OSA_learning_core, exp == "OSA2"), "score")boundary (singular) fit: see help('isSingular')Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 602.6
Scaled residuals:
Min 1Q Median 3Q Max
-1.1996 -1.1013 0.8063 0.9046 1.0226
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.0000 0.0000
Residual 0.2485 0.4985
Number of obs: 408, groups: subject, 102
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 0.49020 0.04936 404.00000 9.930 <2e-16 ***
violation_number1 0.10784 0.06981 404.00000 1.545 0.123
violation_number2 0.05882 0.06981 404.00000 0.843 0.400
violation_number3 0.06863 0.06981 404.00000 0.983 0.326
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.707
viltn_nmbr2 -0.707 0.500
viltn_nmbr3 -0.707 0.500 0.500
optimizer (nloptwrap) convergence code: 0 (OK)
boundary (singular) fit: see help('isSingular')
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.490 0.0494 404 0.366 0.614
1 0.598 0.0494 404 0.474 0.722
2 0.549 0.0494 404 0.425 0.673
3 0.559 0.0494 404 0.435 0.683
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.1078 0.0698 303 -1.545 0.7406
violation_number0 - violation_number2 -0.0588 0.0698 303 -0.843 1.0000
violation_number0 - violation_number3 -0.0686 0.0698 303 -0.983 1.0000
violation_number1 - violation_number2 0.0490 0.0698 303 0.702 1.0000
violation_number1 - violation_number3 0.0392 0.0698 303 0.562 1.0000
violation_number2 - violation_number3 -0.0098 0.0698 303 -0.140 1.0000
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 32028, p-value = 0.5401
alternative hypothesis: two.sidedOSA3 only
data_curve_analysis(filter(OSA_learning_core, exp == "OSA3"), "score")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 1290.4
Scaled residuals:
Min 1Q Median 3Q Max
-1.2340 -0.9373 -0.6648 0.9894 1.2861
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.02027 0.1424
Residual 0.22955 0.4791
Number of obs: 884, groups: subject, 221
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 0.48869 0.03362 862.95230 14.535 <2e-16 ***
violation_number1 -0.02715 0.04558 659.99999 -0.596 0.552
violation_number2 0.03167 0.04558 659.99999 0.695 0.487
violation_number3 -0.04525 0.04558 659.99999 -0.993 0.321
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.678
viltn_nmbr2 -0.678 0.500
viltn_nmbr3 -0.678 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.489 0.0336 863 0.405 0.573
1 0.462 0.0336 863 0.377 0.546
2 0.520 0.0336 863 0.436 0.605
3 0.443 0.0336 863 0.359 0.528
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 0.0271 0.0456 660 0.596 1.0000
violation_number0 - violation_number2 -0.0317 0.0456 660 -0.695 1.0000
violation_number0 - violation_number3 0.0452 0.0456 660 0.993 1.0000
violation_number1 - violation_number2 -0.0588 0.0456 660 -1.291 0.9865
violation_number1 - violation_number3 0.0181 0.0456 660 0.397 1.0000
violation_number2 - violation_number3 0.0769 0.0456 660 1.688 0.5516
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 144644, p-value = 0.6581
alternative hypothesis: two.sided2.2 first block (first three trials)
OSA 2 & OSA 3
data_curve_analysis(filter(OSA_learning_core, trial_id <= 3), "score")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 950.1
Scaled residuals:
Min 1Q Median 3Q Max
-1.1199 -0.9254 -0.8136 1.0053 1.1581
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.008207 0.09059
Residual 0.241095 0.49101
Number of obs: 646, groups: subject, 323
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 0.43073 0.04131 641.94019 10.427 <2e-16 ***
violation_number1 0.07430 0.05674 603.29061 1.309 0.1909
violation_number2 0.11853 0.05600 501.82663 2.117 0.0348 *
violation_number3 0.02234 0.05612 609.66111 0.398 0.6907
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.723
viltn_nmbr2 -0.726 0.529
viltn_nmbr3 -0.731 0.537 0.536
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.431 0.0414 642 0.327 0.534
1 0.505 0.0393 642 0.407 0.603
2 0.549 0.0386 642 0.453 0.646
3 0.453 0.0384 642 0.357 0.549
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.0743 0.0569 603 -1.306 0.7682
violation_number0 - violation_number2 -0.1185 0.0561 502 -2.112 0.2112
violation_number0 - violation_number3 -0.0223 0.0563 610 -0.397 1.0000
violation_number1 - violation_number2 -0.0442 0.0549 600 -0.806 1.0000
violation_number1 - violation_number3 0.0520 0.0544 491 0.954 1.0000
violation_number2 - violation_number3 0.0962 0.0542 582 1.776 0.3815
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 79210, p-value = 0.6915
alternative hypothesis: two.sidedOSA2 only
data_curve_analysis(filter(OSA_learning_core, trial_id <= 3, exp == "OSA2"), "score")boundary (singular) fit: see help('isSingular')Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 306
Scaled residuals:
Min 1Q Median 3Q Max
-1.2084 -1.0001 0.7917 0.8519 1.1251
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.00 0.0
Residual 0.25 0.5
Number of obs: 204, groups: subject, 102
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 0.43750 0.07217 200.00000 6.062 6.57e-09 ***
violation_number1 0.13657 0.09918 200.00000 1.377 0.170
violation_number2 0.16667 0.10206 200.00000 1.633 0.104
violation_number3 0.06250 0.09918 200.00000 0.630 0.529
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.728
viltn_nmbr2 -0.707 0.514
viltn_nmbr3 -0.728 0.529 0.514
optimizer (nloptwrap) convergence code: 0 (OK)
boundary (singular) fit: see help('isSingular')
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.438 0.0722 200 0.256 0.619
1 0.574 0.0680 200 0.403 0.746
2 0.604 0.0722 200 0.422 0.786
3 0.500 0.0680 200 0.329 0.671
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.1366 0.0992 200 -1.377 0.8502
violation_number0 - violation_number2 -0.1667 0.1020 100 -1.633 0.6336
violation_number0 - violation_number3 -0.0625 0.0992 200 -0.630 1.0000
violation_number1 - violation_number2 -0.0301 0.0992 200 -0.303 1.0000
violation_number1 - violation_number3 0.0741 0.0962 100 0.770 1.0000
violation_number2 - violation_number3 0.1042 0.0992 200 1.050 1.0000
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 8040, p-value = 0.6018
alternative hypothesis: two.sidedOSA3 only
data_curve_analysis(filter(OSA_learning_core, trial_id <= 3, exp == "OSA3"), "score")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 653.3
Scaled residuals:
Min 1Q Median 3Q Max
-1.0854 -0.8890 -0.7937 1.0609 1.1706
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.01159 0.1077
Residual 0.23793 0.4878
Number of obs: 442, groups: subject, 221
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 4.270e-01 5.042e-02 4.378e+02 8.468 3.81e-16 ***
violation_number1 4.318e-02 6.907e-02 3.823e+02 0.625 0.532
violation_number2 1.004e-01 6.749e-02 3.956e+02 1.488 0.138
violation_number3 4.872e-03 6.799e-02 3.923e+02 0.072 0.943
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.719
viltn_nmbr2 -0.738 0.538
viltn_nmbr3 -0.732 0.533 0.549
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.427 0.0506 438 0.300 0.554
1 0.470 0.0482 438 0.349 0.591
2 0.527 0.0457 438 0.413 0.642
3 0.432 0.0465 438 0.315 0.548
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.04318 0.0693 382 -0.623 1.0000
violation_number0 - violation_number2 -0.10042 0.0678 396 -1.482 0.8352
violation_number0 - violation_number3 -0.00487 0.0683 392 -0.071 1.0000
violation_number1 - violation_number2 -0.05723 0.0659 382 -0.868 1.0000
violation_number1 - violation_number3 0.03831 0.0665 391 0.576 1.0000
violation_number2 - violation_number3 0.09554 0.0646 361 1.480 0.8352
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 36802, p-value = 0.8719
alternative hypothesis: two.sided2.3 first trial
OSA 2 & OSA 3
data_curve_analysis(filter(OSA_learning_core, trial_id == 1), "score", use_lmer = FALSE)Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Call:
lm(formula = formula_lm, data = data)
Residuals:
Min 1Q Median 3Q Max
-0.6000 -0.5273 0.4000 0.4348 0.6000
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.52727 0.06710 7.858 1.23e-13 ***
violation_number1 0.07273 0.08966 0.811 0.418
violation_number2 0.03794 0.08995 0.422 0.674
violation_number3 -0.12727 0.09489 -1.341 0.181
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 0.4976 on 245 degrees of freedom
Multiple R-squared: 0.0219, Adjusted R-squared: 0.009926
F-statistic: 1.829 on 3 and 245 DF, p-value: 0.1425
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.527 0.0671 245 0.358 0.696
1 0.600 0.0595 245 0.450 0.750
2 0.565 0.0599 245 0.414 0.716
3 0.400 0.0671 245 0.231 0.569
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.0727 0.0897 245 -0.811 1.0000
violation_number0 - violation_number2 -0.0379 0.0899 245 -0.422 1.0000
violation_number0 - violation_number3 0.1273 0.0949 245 1.341 0.7243
violation_number1 - violation_number2 0.0348 0.0844 245 0.412 1.0000
violation_number1 - violation_number3 0.2000 0.0897 245 2.231 0.1597
violation_number2 - violation_number3 0.1652 0.0899 245 1.837 0.3373
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 10810, p-value = 0.2293
alternative hypothesis: two.sidedOSA2 only
data_curve_analysis(filter(OSA_learning_core, trial_id == 1, exp == "OSA2"), "score", use_lmer = FALSE)Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Call:
lm(formula = formula_lm, data = data)
Residuals:
Min 1Q Median 3Q Max
-0.5625 -0.5357 0.4375 0.4474 0.5500
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.45000 0.11349 3.965 0.00014 ***
violation_number1 0.10263 0.14022 0.732 0.46594
violation_number2 0.08571 0.14860 0.577 0.56538
violation_number3 0.11250 0.17024 0.661 0.51027
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 0.5076 on 98 degrees of freedom
Multiple R-squared: 0.006503, Adjusted R-squared: -0.02391
F-statistic: 0.2138 on 3 and 98 DF, p-value: 0.8866
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.450 0.1130 98 0.161 0.739
1 0.553 0.0823 98 0.343 0.762
2 0.536 0.0959 98 0.292 0.780
3 0.562 0.1270 98 0.240 0.885
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.10263 0.140 98 -0.732 1.0000
violation_number0 - violation_number2 -0.08571 0.149 98 -0.577 1.0000
violation_number0 - violation_number3 -0.11250 0.170 98 -0.661 1.0000
violation_number1 - violation_number2 0.01692 0.126 98 0.134 1.0000
violation_number1 - violation_number3 -0.00987 0.151 98 -0.065 1.0000
violation_number2 - violation_number3 -0.02679 0.159 98 -0.168 1.0000
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 1961, p-value = 0.6241
alternative hypothesis: two.sidedOSA3 only
data_curve_analysis(filter(OSA_learning_core, trial_id == 1, exp == "OSA3"), "score", use_lmer = FALSE)Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Call:
lm(formula = formula_lm, data = data)
Residuals:
Min 1Q Median 3Q Max
-0.6562 -0.5714 0.3438 0.4146 0.6667
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.57143 0.08291 6.892 1.63e-10 ***
violation_number1 0.08482 0.11997 0.707 0.4807
violation_number2 0.01394 0.11289 0.123 0.9019
violation_number3 -0.23810 0.11421 -2.085 0.0389 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 0.4905 on 143 degrees of freedom
Multiple R-squared: 0.0602, Adjusted R-squared: 0.04049
F-statistic: 3.054 on 3 and 143 DF, p-value: 0.03051
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.571 0.0829 143 0.362 0.781
1 0.656 0.0867 143 0.437 0.876
2 0.585 0.0766 143 0.392 0.779
3 0.333 0.0785 143 0.135 0.532
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.0848 0.120 143 -0.707 1.0000
violation_number0 - violation_number2 -0.0139 0.113 143 -0.123 1.0000
violation_number0 - violation_number3 0.2381 0.114 143 2.085 0.1555
violation_number1 - violation_number2 0.0709 0.116 143 0.613 1.0000
violation_number1 - violation_number3 0.3229 0.117 143 2.760 0.0392
violation_number2 - violation_number3 0.2520 0.110 143 2.297 0.1153
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 3485, p-value = 0.05463
alternative hypothesis: two.sided3 SURPRISE
3.1 preprocessing
3.2 plots and models
Coordinate system already present. Adding new coordinate system, which will
replace the existing one.
Coordinate system already present. Adding new coordinate system, which will
replace the existing one.
Coordinate system already present. Adding new coordinate system, which will
replace the existing one.
OSA 1 & OSA 4
data_curve_analysis(OSA_surprise_core, "surprise")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 4692.9
Scaled residuals:
Min 1Q Median 3Q Max
-2.75591 -0.67017 -0.03989 0.71665 3.10394
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.8476 0.9207
Residual 1.6431 1.2818
Number of obs: 1296, groups: subject, 324
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.23148 0.08768 958.85926 14.05 <2e-16 ***
violation_number1 2.47531 0.10071 969.00000 24.58 <2e-16 ***
violation_number2 3.18827 0.10071 969.00000 31.66 <2e-16 ***
violation_number3 3.48148 0.10071 969.00000 34.57 <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.574
viltn_nmbr2 -0.574 0.500
viltn_nmbr3 -0.574 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.23 0.0877 959 1.01 1.45
1 3.71 0.0877 959 3.49 3.93
2 4.42 0.0877 959 4.20 4.64
3 4.71 0.0877 959 4.49 4.93
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -2.475 0.101 969 -24.578 <.0001
violation_number0 - violation_number2 -3.188 0.101 969 -31.658 <.0001
violation_number0 - violation_number3 -3.481 0.101 969 -34.569 <.0001
violation_number1 - violation_number2 -0.713 0.101 969 -7.079 <.0001
violation_number1 - violation_number3 -1.006 0.101 969 -9.991 <.0001
violation_number2 - violation_number3 -0.293 0.101 969 -2.911 0.0037
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 487938, p-value < 2.2e-16
alternative hypothesis: two.sidedOSA 1 only
data_curve_analysis(filter(OSA_surprise_core, exp == "OSA1"), "surprise")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 3478.1
Scaled residuals:
Min 1Q Median 3Q Max
-2.63159 -0.60793 -0.00032 0.66600 3.09224
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.7776 0.8818
Residual 1.6141 1.2705
Number of obs: 968, groups: subject, 242
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.19835 0.09941 731.90328 12.05 <2e-16 ***
violation_number1 2.61570 0.11550 723.00000 22.65 <2e-16 ***
violation_number2 3.38017 0.11550 723.00000 29.27 <2e-16 ***
violation_number3 3.72314 0.11550 723.00000 32.23 <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.581
viltn_nmbr2 -0.581 0.500
viltn_nmbr3 -0.581 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.20 0.0994 732 0.949 1.45
1 3.81 0.0994 732 3.565 4.06
2 4.58 0.0994 732 4.330 4.83
3 4.92 0.0994 732 4.673 5.17
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -2.616 0.115 723 -22.647 <.0001
violation_number0 - violation_number2 -3.380 0.115 723 -29.266 <.0001
violation_number0 - violation_number3 -3.723 0.115 723 -32.235 <.0001
violation_number1 - violation_number2 -0.764 0.115 723 -6.619 <.0001
violation_number1 - violation_number3 -1.107 0.115 723 -9.588 <.0001
violation_number2 - violation_number3 -0.343 0.115 723 -2.970 0.0031
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 276668, p-value < 2.2e-16
alternative hypothesis: two.sidedOSA 4 only
data_curve_analysis(filter(OSA_surprise_core, exp == "OSA4"), "surprise")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 1191.7
Scaled residuals:
Min 1Q Median 3Q Max
-2.46457 -0.66872 -0.02504 0.53703 2.80926
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.9547 0.9771
Residual 1.6214 1.2733
Number of obs: 328, groups: subject, 82
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.3293 0.1772 229.4576 7.50 1.39e-12 ***
violation_number1 2.0610 0.1989 243.0000 10.36 < 2e-16 ***
violation_number2 2.6220 0.1989 243.0000 13.19 < 2e-16 ***
violation_number3 2.7683 0.1989 243.0000 13.92 < 2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.561
viltn_nmbr2 -0.561 0.500
viltn_nmbr3 -0.561 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.33 0.177 229 0.883 1.78
1 3.39 0.177 229 2.944 3.84
2 3.95 0.177 229 3.505 4.40
3 4.10 0.177 229 3.651 4.54
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -2.061 0.199 243 -10.364 <.0001
violation_number0 - violation_number2 -2.622 0.199 243 -13.185 <.0001
violation_number0 - violation_number3 -2.768 0.199 243 -13.921 <.0001
violation_number1 - violation_number2 -0.561 0.199 243 -2.821 0.0104
violation_number1 - violation_number3 -0.707 0.199 243 -3.557 0.0014
violation_number2 - violation_number3 -0.146 0.199 243 -0.736 0.4625
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 29824, p-value < 2.2e-16
alternative hypothesis: two.sided4 EXPLANATION
4.1 NUMBER OF EXPLANATION
4.1.1 all explanations included
Warning: Removed 2 rows containing non-finite outside the scale range
(`stat_summary()`).
Removed 2 rows containing non-finite outside the scale range
(`stat_summary()`).Warning: Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
OSA 1 & OSA 4
data_curve_analysis(OSA_explan_core, "explanation_num")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 1319
Scaled residuals:
Min 1Q Median 3Q Max
-2.5850 -0.0819 -0.0292 0.4344 4.5803
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.05712 0.2390
Residual 0.12294 0.3506
Number of obs: 1294, groups: subject, 324
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 0.65338 0.02360 993.40235 27.680 <2e-16 ***
violation_number1 0.28489 0.02757 967.71095 10.332 <2e-16 ***
violation_number2 0.27872 0.02757 967.71095 10.108 <2e-16 ***
violation_number3 0.26641 0.02760 968.08540 9.653 <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.585
viltn_nmbr2 -0.585 0.501
viltn_nmbr3 -0.585 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.653 0.0236 993 0.594 0.712
1 0.938 0.0236 991 0.879 0.997
2 0.932 0.0236 991 0.873 0.991
3 0.920 0.0236 993 0.861 0.979
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.28489 0.0276 967 -10.332 <.0001
violation_number0 - violation_number2 -0.27872 0.0276 967 -10.108 <.0001
violation_number0 - violation_number3 -0.26641 0.0276 968 -9.653 <.0001
violation_number1 - violation_number2 0.00617 0.0275 967 0.224 1.0000
violation_number1 - violation_number3 0.01848 0.0276 967 0.670 1.0000
violation_number2 - violation_number3 0.01231 0.0276 967 0.446 1.0000
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 353566, p-value = 1.359e-07
alternative hypothesis: two.sidedOSA 1 only
data_curve_analysis(filter(OSA_explan_core, exp == "OSA1"), "explanation_num")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 1127.4
Scaled residuals:
Min 1Q Median 3Q Max
-2.2858 -0.0718 -0.0214 0.3934 4.5101
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.06114 0.2473
Residual 0.14396 0.3794
Number of obs: 967, groups: subject, 242
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 0.53552 0.02916 762.43756 18.36 <2e-16 ***
violation_number1 0.35704 0.03453 722.57683 10.34 <2e-16 ***
violation_number2 0.36531 0.03453 722.57683 10.58 <2e-16 ***
violation_number3 0.34878 0.03453 722.57683 10.10 <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.594
viltn_nmbr2 -0.594 0.501
viltn_nmbr3 -0.594 0.501 0.501
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 0.536 0.0292 762 0.463 0.609
1 0.893 0.0291 760 0.820 0.965
2 0.901 0.0291 760 0.828 0.974
3 0.884 0.0291 760 0.811 0.957
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.35704 0.0345 722 -10.339 <.0001
violation_number0 - violation_number2 -0.36531 0.0345 722 -10.578 <.0001
violation_number0 - violation_number3 -0.34878 0.0345 722 -10.099 <.0001
violation_number1 - violation_number2 -0.00826 0.0345 722 -0.240 1.0000
violation_number1 - violation_number3 0.00826 0.0345 722 0.240 1.0000
violation_number2 - violation_number3 0.01653 0.0345 722 0.479 1.0000
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 205154, p-value = 8.124e-10
alternative hypothesis: two.sidedOSA 4 only
data_curve_analysis(filter(OSA_explan_core, exp == "OSA4"), "explanation_num")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: -26.7
Scaled residuals:
Min 1Q Median 3Q Max
-4.4984 -0.2778 -0.0415 0.0776 7.1291
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.01164 0.1079
Residual 0.04241 0.2059
Number of obs: 327, groups: subject, 82
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.00000 0.02567 283.78402 38.951 <2e-16 ***
violation_number1 0.07317 0.03216 242.24543 2.275 0.0238 *
violation_number2 0.02439 0.03216 242.24543 0.758 0.4490
violation_number3 0.02451 0.03228 242.71322 0.759 0.4483
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.626
viltn_nmbr2 -0.626 0.500
viltn_nmbr3 -0.624 0.498 0.498
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.00 0.0257 284 0.935 1.06
1 1.07 0.0257 284 1.009 1.14
2 1.02 0.0257 284 0.960 1.09
3 1.02 0.0258 285 0.960 1.09
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.07317 0.0322 242 -2.275 0.1426
violation_number0 - violation_number2 -0.02439 0.0322 242 -0.758 1.0000
violation_number0 - violation_number3 -0.02451 0.0323 242 -0.759 1.0000
violation_number1 - violation_number2 0.04878 0.0322 242 1.517 0.6532
violation_number1 - violation_number3 0.04866 0.0323 242 1.508 0.6532
violation_number2 - violation_number3 -0.00012 0.0323 242 -0.004 1.0000
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 20017, p-value = 0.9733
alternative hypothesis: two.sided4.1.2 no give up answers
Warning: Removed 2 rows containing non-finite outside the scale range
(`stat_summary()`).
Removed 2 rows containing non-finite outside the scale range
(`stat_summary()`).Warning: Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
OSA 1 & OSA 4
data_curve_analysis(filter(OSA_explan_core, giveup == FALSE), "explanation_num")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 90.1
Scaled residuals:
Min 1Q Median 3Q Max
-4.0959 -0.2651 -0.0803 0.0257 6.2645
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.02145 0.1465
Residual 0.04853 0.2203
Number of obs: 751, groups: subject, 282
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.00888 0.01830 694.01560 55.140 < 2e-16 ***
violation_number1 0.08071 0.02314 557.70006 3.488 0.000526 ***
violation_number2 0.04011 0.02341 560.47616 1.714 0.087146 .
violation_number3 0.01663 0.02392 558.25970 0.695 0.487136
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.616
viltn_nmbr2 -0.609 0.501
viltn_nmbr3 -0.592 0.489 0.493
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.01 0.0183 701 0.963 1.05
1 1.09 0.0187 703 1.043 1.14
2 1.05 0.0190 708 1.001 1.10
3 1.03 0.0197 721 0.976 1.07
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.0807 0.0232 578 -3.485 0.0032
violation_number0 - violation_number2 -0.0401 0.0234 581 -1.712 0.3266
violation_number0 - violation_number3 -0.0166 0.0239 579 -0.695 0.6503
violation_number1 - violation_number2 0.0406 0.0233 542 1.744 0.3266
violation_number1 - violation_number3 0.0641 0.0238 544 2.690 0.0368
violation_number2 - violation_number3 0.0235 0.0238 526 0.985 0.6503
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 105853, p-value = 0.9461
alternative hypothesis: two.sidedOSA 1 only
data_curve_analysis(filter(OSA_explan_core, giveup == FALSE, exp == "OSA1"), "explanation_num")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 92
Scaled residuals:
Min 1Q Median 3Q Max
-3.5260 -0.2519 -0.0855 0.0397 6.0071
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.02828 0.1682
Residual 0.04824 0.2196
Number of obs: 460, groups: subject, 200
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.01309 0.02438 436.17550 41.556 <2e-16 ***
violation_number1 0.08010 0.03049 343.46990 2.627 0.009 **
violation_number2 0.04708 0.03095 344.72466 1.521 0.129
violation_number3 0.01170 0.03201 343.29845 0.366 0.715
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.630
viltn_nmbr2 -0.621 0.529
viltn_nmbr3 -0.597 0.514 0.522
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.01 0.0244 440 0.952 1.07
1 1.09 0.0243 438 1.032 1.15
2 1.06 0.0248 441 0.998 1.12
3 1.02 0.0263 450 0.959 1.09
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.0801 0.0305 360 -2.623 0.0546
violation_number0 - violation_number2 -0.0471 0.0310 361 -1.519 0.5190
violation_number0 - violation_number3 -0.0117 0.0321 360 -0.365 0.7559
violation_number1 - violation_number2 0.0330 0.0299 318 1.106 0.7559
violation_number1 - violation_number3 0.0684 0.0309 317 2.215 0.1375
violation_number2 - violation_number3 0.0354 0.0308 302 1.148 0.7559
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 39716, p-value = 0.949
alternative hypothesis: two.sidedOSA 4 only
data_curve_analysis(filter(OSA_explan_core, giveup == FALSE, exp == "OSA4"), "explanation_num")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 10.9
Scaled residuals:
Min 1Q Median 3Q Max
-4.2859 -0.2930 -0.0443 0.0741 6.7585
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.01273 0.1128
Residual 0.04755 0.2181
Number of obs: 291, groups: subject, 82
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 0.99988 0.02727 258.80613 36.668 <2e-16 ***
violation_number1 0.08144 0.03566 218.12596 2.284 0.0233 *
violation_number2 0.02720 0.03586 220.00399 0.759 0.4489
violation_number3 0.02665 0.03601 220.17750 0.740 0.4601
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.604
viltn_nmbr2 -0.602 0.466
viltn_nmbr3 -0.599 0.461 0.462
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.00 0.0273 257 0.931 1.07
1 1.08 0.0290 266 1.008 1.15
2 1.03 0.0292 267 0.954 1.10
3 1.03 0.0294 268 0.953 1.10
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.081440 0.0357 214 -2.283 0.1404
violation_number0 - violation_number2 -0.027198 0.0359 216 -0.758 1.0000
violation_number0 - violation_number3 -0.026645 0.0360 217 -0.740 1.0000
violation_number1 - violation_number2 0.054242 0.0370 214 1.467 0.7126
violation_number1 - violation_number3 0.054795 0.0372 217 1.472 0.7126
violation_number2 - violation_number3 0.000553 0.0373 215 0.015 1.0000
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 15880, p-value = 0.9751
alternative hypothesis: two.sided4.2 EVALUATION OF EXPLANATIONS
4.2.1 all explanations included
Warning: Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
OSA 1 & OSA 4
data_curve_analysis(OSA_explan_core, "explanation_eval")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 4939.6
Scaled residuals:
Min 1Q Median 3Q Max
-2.1030 -0.6957 -0.1337 0.7501 2.5690
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.8723 0.934
Residual 2.0598 1.435
Number of obs: 1295, groups: subject, 324
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.73148 0.09513 1019.96941 18.201 < 2e-16 ***
violation_number1 0.46605 0.11276 967.61932 4.133 3.89e-05 ***
violation_number2 -0.13889 0.11276 967.61932 -1.232 0.218
violation_number3 -0.47387 0.11286 968.01058 -4.199 2.93e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.593
viltn_nmbr2 -0.593 0.500
viltn_nmbr3 -0.592 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.73 0.0951 1020 1.49 1.97
1 2.20 0.0951 1020 1.96 2.44
2 1.59 0.0951 1020 1.35 1.83
3 1.26 0.0953 1022 1.02 1.50
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.466 0.113 968 -4.133 0.0001
violation_number0 - violation_number2 0.139 0.113 968 1.232 0.2184
violation_number0 - violation_number3 0.474 0.113 968 4.199 0.0001
violation_number1 - violation_number2 0.605 0.113 968 5.365 <.0001
violation_number1 - violation_number3 0.940 0.113 968 8.328 <.0001
violation_number2 - violation_number3 0.335 0.113 968 2.968 0.0061
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 284360, p-value = 6.384e-05
alternative hypothesis: two.sidedOSA 1 only
data_curve_analysis(filter(OSA_explan_core, exp == "OSA1"), "explanation_eval")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 3704.3
Scaled residuals:
Min 1Q Median 3Q Max
-1.9280 -0.6998 -0.3243 0.7641 2.5307
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.7456 0.8635
Residual 2.1475 1.4654
Number of obs: 968, groups: subject, 242
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 1.42975 0.10934 803.81919 13.076 < 2e-16 ***
violation_number1 0.49174 0.13322 723.00000 3.691 0.00024 ***
violation_number2 -0.02479 0.13322 723.00000 -0.186 0.85241
violation_number3 -0.40496 0.13322 723.00000 -3.040 0.00245 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.609
viltn_nmbr2 -0.609 0.500
viltn_nmbr3 -0.609 0.500 0.500
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 1.43 0.109 804 1.156 1.70
1 1.92 0.109 804 1.648 2.20
2 1.40 0.109 804 1.131 1.68
3 1.02 0.109 804 0.751 1.30
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.4917 0.133 723 -3.691 0.0010
violation_number0 - violation_number2 0.0248 0.133 723 0.186 0.8524
violation_number0 - violation_number3 0.4050 0.133 723 3.040 0.0074
violation_number1 - violation_number2 0.5165 0.133 723 3.877 0.0006
violation_number1 - violation_number3 0.8967 0.133 723 6.731 <.0001
violation_number2 - violation_number3 0.3802 0.133 723 2.854 0.0089
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 162100, p-value = 0.005192
alternative hypothesis: two.sidedOSA 4 only
data_curve_analysis(filter(OSA_explan_core, exp == "OSA4"), "explanation_eval")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 1186.2
Scaled residuals:
Min 1Q Median 3Q Max
-2.3788 -0.7247 0.0771 0.7473 2.2556
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.5334 0.7304
Residual 1.7929 1.3390
Number of obs: 327, groups: subject, 82
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 2.6220 0.1684 279.0081 15.567 < 2e-16 ***
violation_number1 0.3902 0.2091 241.8397 1.866 0.06323 .
violation_number2 -0.4756 0.2091 241.8397 -2.274 0.02382 *
violation_number3 -0.6755 0.2099 242.2945 -3.219 0.00146 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.621
viltn_nmbr2 -0.621 0.500
viltn_nmbr3 -0.619 0.498 0.498
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 2.62 0.168 279 2.20 3.05
1 3.01 0.168 279 2.59 3.44
2 2.15 0.168 279 1.72 2.57
3 1.95 0.169 281 1.52 2.37
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.390 0.209 242 -1.866 0.1265
violation_number0 - violation_number2 0.476 0.209 242 2.274 0.0715
violation_number0 - violation_number3 0.675 0.210 242 3.219 0.0059
violation_number1 - violation_number2 0.866 0.209 242 4.141 0.0002
violation_number1 - violation_number3 1.066 0.210 242 5.078 <.0001
violation_number2 - violation_number3 0.200 0.210 242 0.952 0.3418
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 16802, p-value = 0.0006838
alternative hypothesis: two.sided4.2.2 no give up answers
Warning: Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
Removed 1 row containing non-finite outside the scale range (`stat_summary()`).
OSA 1 & OSA 4
data_curve_analysis(filter(OSA_explan_core, giveup == FALSE), "explanation_eval")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 2402.9
Scaled residuals:
Min 1Q Median 3Q Max
-3.16152 -0.59350 0.06258 0.67453 2.28994
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.3426 0.5853
Residual 1.1435 1.0693
Number of obs: 752, groups: subject, 282
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 2.78057 0.08479 722.50253 32.795 < 2e-16 ***
violation_number1 0.88420 0.11111 588.31958 7.958 9.05e-15 ***
violation_number2 -0.02338 0.11235 591.84901 -0.208 0.835200
violation_number3 -0.42594 0.11486 590.12699 -3.709 0.000228 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.637
viltn_nmbr2 -0.631 0.495
viltn_nmbr3 -0.613 0.484 0.486
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 2.78 0.0848 722 2.57 2.99
1 3.66 0.0868 724 3.45 3.88
2 2.76 0.0884 726 2.54 2.98
3 2.35 0.0919 734 2.12 2.58
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.8842 0.111 588 -7.952 <.0001
violation_number0 - violation_number2 0.0234 0.112 592 0.208 0.8353
violation_number0 - violation_number3 0.4259 0.115 590 3.706 0.0007
violation_number1 - violation_number2 0.9076 0.112 553 8.081 <.0001
violation_number1 - violation_number3 1.3101 0.115 555 11.401 <.0001
violation_number2 - violation_number3 0.4026 0.115 536 3.494 0.0010
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 91771, p-value = 2.187e-05
alternative hypothesis: two.sidedOSA 1 only
data_curve_analysis(filter(OSA_explan_core, giveup == FALSE, exp == "OSA1"), "explanation_eval")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 1414.3
Scaled residuals:
Min 1Q Median 3Q Max
-3.12056 -0.65101 0.08915 0.64610 2.42294
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.247 0.497
Residual 1.032 1.016
Number of obs: 461, groups: subject, 200
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 2.84400 0.10172 453.26114 27.960 < 2e-16 ***
violation_number1 0.90770 0.13582 372.18002 6.683 8.56e-11 ***
violation_number2 0.06756 0.13777 375.21803 0.490 0.62417
violation_number3 -0.43793 0.14253 374.85700 -3.072 0.00228 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.670
viltn_nmbr2 -0.661 0.512
viltn_nmbr3 -0.636 0.496 0.498
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 2.84 0.102 454 2.59 3.10
1 3.75 0.102 453 3.50 4.01
2 2.91 0.104 453 2.65 3.17
3 2.41 0.111 456 2.13 2.68
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.9077 0.136 378 -6.672 <.0001
violation_number0 - violation_number2 -0.0676 0.138 381 -0.490 0.6248
violation_number0 - violation_number3 0.4379 0.143 381 3.067 0.0046
violation_number1 - violation_number2 0.8401 0.135 338 6.207 <.0001
violation_number1 - violation_number3 1.3456 0.140 338 9.609 <.0001
violation_number2 - violation_number3 0.5055 0.141 321 3.595 0.0011
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 34272, p-value = 0.0005648
alternative hypothesis: two.sidedOSA 4 only
data_curve_analysis(filter(OSA_explan_core, giveup == FALSE, exp == "OSA4"), "explanation_eval")Warning in jonckheere.test(data[[dv]], as.numeric(data$violation_number)): Sample size > 100 or data with ties
p-value based on normal approximation. Specify nperm for permutation p-value$model_summary
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: formula_lmer
Data: data
REML criterion at convergence: 979.2
Scaled residuals:
Min 1Q Median 3Q Max
-2.82239 -0.60185 0.03984 0.66667 2.05175
Random effects:
Groups Name Variance Std.Dev.
subject (Intercept) 0.4634 0.6807
Residual 1.3375 1.1565
Number of obs: 291, groups: subject, 82
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 2.6575 0.1490 242.7931 17.834 < 2e-16 ***
violation_number1 0.8475 0.1893 211.0831 4.477 1.24e-05 ***
violation_number2 -0.1667 0.1904 212.9626 -0.875 0.3823
violation_number3 -0.4021 0.1912 213.0728 -2.103 0.0367 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) vltn_1 vltn_2
viltn_nmbr1 -0.585
viltn_nmbr2 -0.584 0.466
viltn_nmbr3 -0.581 0.461 0.462
$emmeans
violation_number emmean SE df lower.CL upper.CL
0 2.66 0.149 245 2.28 3.03
1 3.51 0.158 256 3.11 3.90
2 2.49 0.159 257 2.09 2.89
3 2.26 0.160 259 1.85 2.66
Degrees-of-freedom method: kenward-roger
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 4 estimates
$pairwise_comparisons
contrast estimate SE df t.ratio p.value
violation_number0 - violation_number1 -0.848 0.189 214 -4.476 <.0001
violation_number0 - violation_number2 0.167 0.190 216 0.875 0.4714
violation_number0 - violation_number3 0.402 0.191 216 2.102 0.1102
violation_number1 - violation_number2 1.014 0.196 214 5.167 <.0001
violation_number1 - violation_number3 1.250 0.198 216 6.320 <.0001
violation_number2 - violation_number3 0.235 0.198 214 1.189 0.4714
Degrees-of-freedom method: kenward-roger
P value adjustment: holm method for 6 tests
$jonckheere_test
Jonckheere-Terpstra test
data:
JT = 13901, p-value = 0.01491
alternative hypothesis: two.sided