ManyNumbers: Panamath-Custom AMS Pilot
RAM
2024-08-21
Database
# panamath_custom_data = readxl::read_xlsx("csv_files/processed_data/ams/lf_files/lf_mn_ams_2024-07-24_all.xlsx", sheet = "lf_mn_ams_2024-07-24")
#panamath_custom_data = readxl::read_xlsx("csv_files/processed_data/ams/lf_files/lf_mn_ams_2024-07-24_all.xlsx", sheet = "lf_mn_ams_2024-08-07")
panamath_participants = readxl::read_xlsx("/Volumes/BL-PSY-gunderson_lab/Main/Studies/2023_manynumbers/participant_lists/participantlist_panamath_custom_pilot.xlsx")
panamath_participants = subset(panamath_participants, remove!=1)
panamath_custom_data1 = read.csv("2024_nie_pilot/csv_files/processed_data/ams/lf_files/lf_mn_ams_2024-08-21.csv")
panamath_custom_data2 = read.csv("2024_ucsd_pilot_panamath/csv_files/processed_data/ams/lf_files/lf_mn_ams_2024-08-09.csv")
panamath_custom_data3 = read.csv("2024_iu_pilot_panamath/csv_files/processed_data/ams/lf_files/lf_mn_ams_2024-08-18.csv")
panamath_custom_data=rbind(panamath_custom_data1, panamath_custom_data2, panamath_custom_data3)
panamath_custom_data = panamath_custom_data[which(panamath_custom_data$participant %in% panamath_participants$participant),]Practice trials
Analyses
df_acc_noout_practice = subset(panamath_custom_data, block == "practice")
df_acc_noout_practice = subset(df_acc_noout_practice, participant != "mn024")
df_acc_noout_practice$practice_block = ifelse(df_acc_noout_practice$X <10, "block1","block2")
table(df_acc_noout_practice$participant)##
## 240630a cohen finnley harper lfp1 lfp10 lfp11 lfp2
## 12 6 6 6 12 6 12 12
## lfp4 lfp5 lfp7 lfp8 lfp9 mn046 mn048 mn052
## 6 6 6 12 6 12 12 6
## mn054 mn055 mn056 mn057 mn058 mn059 mn060 mn061
## 12 6 6 6 6 6 12 6
## mn062 mn063 mn064 mn065 MOC2Emily MOC2Liam
## 6 12 12 12 12 6agg_ams_task_practice = aggregate(accuracy ~ participant*numRatio*practice_block*age, df_acc_noout_practice, mean)
descriptive_amstask_practice = summarySE(agg_ams_task_practice, "accuracy", c("numRatio","practice_block"))
kable(descriptive_amstask_practice, table.attr = "style = \"color: white;\"")| numRatio | practice_block | N | accuracy | sd | se | ci |
|---|---|---|---|---|---|---|
| 3.5 | block1 | 30 | 0.6333333 | 0.2949122 | 0.0538433 | 0.1101220 |
| 3.5 | block2 | 13 | 0.6153846 | 0.2995723 | 0.0830864 | 0.1810298 |
| 4.0 | block1 | 30 | 0.6111111 | 0.3042903 | 0.0555556 | 0.1136239 |
| 4.0 | block2 | 13 | 0.5641026 | 0.3438512 | 0.0953672 | 0.2077872 |
##
## block1 block2
## 240630a 2 2
## lfp1 2 2
## lfp11 2 2
## lfp2 2 2
## lfp8 2 2
## mn046 2 2
## mn048 2 2
## mn054 2 2
## mn060 2 2
## mn063 2 2
## mn064 2 2
## mn065 2 2
## MOC2Emily 2 2## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: accuracy ~ +numRatio * practice_block + +(1 | participant)
## Data: agg_ams_task_practice
## Control: glmerControl(optimizer = "bobyqa")
##
## AIC BIC logLik deviance df.resid
## 74.7 84.5 -32.4 64.7 47
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8257 -0.2434 -0.1054 0.5557 1.2649
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0 0
## Number of obs: 52, groups: participant, 13
##
## Fixed effects:
## Estimate Std. Error z value
## (Intercept) -0.470003641049 0.570087663279 -0.824
## numRatio3.5 0.000000005885 0.806225830093 0.000
## practice_blockblock2 1.673976503865 0.870823199943 1.922
## numRatio3.5:practice_blockblock2 -0.393042722239 1.201850399147 -0.327
## Pr(>|z|)
## (Intercept) 0.4097
## numRatio3.5 1.0000
## practice_blockblock2 0.0546 .
## numRatio3.5:practice_blockblock2 0.7436
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) nmR3.5 prct_2
## numRatio3.5 -0.707
## prctc_blck2 -0.655 0.463
## nmRt3.5:p_2 0.474 -0.671 -0.725
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: accuracy
## Chisq Df Pr(>Chisq)
## (Intercept) 0.6797 1 0.40969
## numRatio 0.0000 1 1.00000
## practice_block 3.6952 1 0.05457 .
## numRatio:practice_block 0.1069 1 0.74364
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Experimental trials - All
Figure
| visual_control | numRatio | N | accuracy | sd | se | ci |
|---|---|---|---|---|---|---|
| custom | 2.0 | 7 | 0.5119048 | 0.1311326 | 0.0495635 | 0.1212775 |
| custom | 2.5 | 22 | 0.6060606 | 0.2302831 | 0.0490965 | 0.1021018 |
| custom | 3.0 | 29 | 0.6465517 | 0.1845404 | 0.0342683 | 0.0701954 |
| Panamath | 2.5 | 30 | 0.6388889 | 0.2370530 | 0.0432798 | 0.0885170 |
| Panamath | 3.0 | 30 | 0.6027778 | 0.2632352 | 0.0480600 | 0.0982936 |
Analyses
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: accuracy ~ +numRatio * visual_control * block + +(1 | participant)
## Data: df_acc_noout_experimental
## Control: glmerControl(optimizer = "bobyqa")
##
## AIC BIC logLik deviance df.resid
## 1714.0 1761.1 -848.0 1696.0 1380
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1268 -0.9928 0.3597 0.8267 1.4875
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0.862 0.9285
## Number of obs: 1389, groups: participant, 30
##
## Fixed effects:
## Estimate Std. Error z value
## (Intercept) -0.2365 0.8849 -0.267
## numRatio 0.4540 0.3183 1.426
## visual_controlPanamath 2.0464 1.5465 1.323
## blockblock2 -1.1798 1.3746 -0.858
## numRatio:visual_controlPanamath -0.9846 0.5518 -1.784
## numRatio:blockblock2 0.1580 0.4922 0.321
## visual_controlPanamath:blockblock2 -0.1565 2.3940 -0.065
## numRatio:visual_controlPanamath:blockblock2 0.5655 0.8365 0.676
## Pr(>|z|)
## (Intercept) 0.7893
## numRatio 0.1538
## visual_controlPanamath 0.1858
## blockblock2 0.3907
## numRatio:visual_controlPanamath 0.0744 .
## numRatio:blockblock2 0.7482
## visual_controlPanamath:blockblock2 0.9479
## numRatio:visual_controlPanamath:blockblock2 0.4990
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) numRat vsl_cP blckb2 nmR:_P nmRt:2 vs_P:2
## numRatio -0.951
## vsl_cntrlPn -0.571 0.543
## blockblock2 -0.643 0.612 0.393
## nmRt:vsl_cP 0.547 -0.576 -0.969 -0.351
## nmRt:blckb2 0.615 -0.647 -0.350 -0.960 0.371
## vsl_cntrP:2 0.398 -0.352 -0.677 -0.609 0.626 0.551
## nmRt:vs_P:2 -0.361 0.380 0.638 0.563 -0.659 -0.587 -0.953## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: accuracy
## Chisq Df Pr(>Chisq)
## (Intercept) 0.0714 1 0.78930
## numRatio 2.0346 1 0.15376
## visual_control 1.7509 1 0.18577
## block 0.7367 1 0.39073
## numRatio:visual_control 3.1839 1 0.07437 .
## numRatio:block 0.1030 1 0.74825
## visual_control:block 0.0043 1 0.94789
## numRatio:visual_control:block 0.4571 1 0.49897
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## $emtrends
## block = block1:
## visual_control numRatio.trend SE df asymp.LCL asymp.UCL z.ratio p.value
## custom 0.454 0.318 Inf -0.170 1.078 1.426 0.1538
## Panamath -0.531 0.451 Inf -1.415 0.354 -1.176 0.2395
##
## block = block2:
## visual_control numRatio.trend SE df asymp.LCL asymp.UCL z.ratio p.value
## custom 0.612 0.375 Inf -0.124 1.348 1.630 0.1031
## Panamath 0.193 0.505 Inf -0.797 1.183 0.382 0.7025
##
## Confidence level used: 0.95
##
## $contrasts
## block = block1:
## contrast estimate SE df asymp.LCL asymp.UCL z.ratio p.value
## custom - Panamath 0.985 0.552 Inf -0.0969 2.07 1.784 0.0744
##
## block = block2:
## contrast estimate SE df asymp.LCL asymp.UCL z.ratio p.value
## custom - Panamath 0.419 0.629 Inf -0.8142 1.65 0.666 0.5054
##
## Confidence level used: 0.95## $emmeans
## numRatio = 2.0, block = block1:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.662 0.0765 Inf 0.500 0.793 0.5 1.964 0.0496
## Panamath 0.679 0.0941 Inf 0.476 0.831 0.5 1.735 0.0827
##
## numRatio = 2.5, block = block1:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.711 0.0571 Inf 0.588 0.809 0.5 3.236 0.0012
## Panamath 0.619 0.0683 Inf 0.479 0.741 0.5 1.669 0.0951
##
## numRatio = 3.0, block = block1:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.755 0.0548 Inf 0.633 0.846 0.5 3.796 0.0001
## Panamath 0.554 0.0713 Inf 0.414 0.687 0.5 0.755 0.4500
##
## numRatio = 2.0, block = block2:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.452 0.0938 Inf 0.282 0.634 0.5 -0.508 0.6115
## Panamath 0.703 0.0976 Inf 0.486 0.855 0.5 1.839 0.0659
##
## numRatio = 2.5, block = block2:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.528 0.0698 Inf 0.393 0.660 0.5 0.406 0.6850
## Panamath 0.722 0.0608 Inf 0.589 0.825 0.5 3.154 0.0016
##
## numRatio = 3.0, block = block2:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.603 0.0694 Inf 0.463 0.729 0.5 1.446 0.1481
## Panamath 0.741 0.0585 Inf 0.612 0.839 0.5 3.453 0.0006
##
## Confidence level used: 0.95
## Intervals are back-transformed from the logit scale
## Tests are performed on the logit scale
##
## $contrasts
## numRatio = 2, block = block1:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 0.926 0.509 Inf 0.315 2.718 1 -0.140
## p.value
## 0.8885
##
## numRatio = 2.5, block = block1:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 1.515 0.606 Inf 0.691 3.319 1 1.037
## p.value
## 0.2996
##
## numRatio = 3, block = block1:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 2.478 1.024 Inf 1.103 5.570 1 2.197
## p.value
## 0.0280
##
## numRatio = 2, block = block2:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 0.349 0.210 Inf 0.108 1.135 1 -1.750
## p.value
## 0.0801
##
## numRatio = 2.5, block = block2:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 0.431 0.178 Inf 0.192 0.966 1 -2.044
## p.value
## 0.0410
##
## numRatio = 3, block = block2:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 0.531 0.223 Inf 0.233 1.210 1 -1.506
## p.value
## 0.1321
##
## Confidence level used: 0.95
## Intervals are back-transformed from the log odds ratio scale
## Tests are performed on the log odds ratio scaleExperimental trials - NIE only
Figure
| visual_control | numRatio | N | accuracy | sd | se | ci |
|---|---|---|---|---|---|---|
| custom | 2.5 | 22 | 0.6060606 | 0.2302831 | 0.0490965 | 0.1021018 |
| custom | 3.0 | 22 | 0.6439394 | 0.1910390 | 0.0407297 | 0.0847020 |
| Panamath | 2.5 | 23 | 0.6557971 | 0.2347688 | 0.0489527 | 0.1015216 |
| Panamath | 3.0 | 23 | 0.6014493 | 0.2929695 | 0.0610884 | 0.1266895 |
Analyses
| visual_control | numRatio | block | N | accuracy | sd | se | ci |
|---|---|---|---|---|---|---|---|
| custom | 2.5 | block1 | 10 | 0.6833333 | 0.2383068 | 0.0753592 | 0.1704744 |
| custom | 2.5 | block2 | 12 | 0.5416667 | 0.2117150 | 0.0611169 | 0.1345173 |
| custom | 3.0 | block1 | 10 | 0.7416667 | 0.1687371 | 0.0533594 | 0.1207073 |
| custom | 3.0 | block2 | 12 | 0.5625000 | 0.1745304 | 0.0503826 | 0.1108913 |
| Panamath | 2.5 | block1 | 13 | 0.5833333 | 0.2041241 | 0.0566139 | 0.1233510 |
| Panamath | 2.5 | block2 | 10 | 0.7500000 | 0.2484520 | 0.0785674 | 0.1777319 |
| Panamath | 3.0 | block1 | 13 | 0.5384615 | 0.2670470 | 0.0740655 | 0.1613749 |
| Panamath | 3.0 | block2 | 10 | 0.6833333 | 0.3186585 | 0.1007687 | 0.2279545 |
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: accuracy ~ +numRatio * visual_control * block + +(1 | participant)
## Data: df_acc_noout_experimental2
## Control: glmerControl(optimizer = "bobyqa")
##
## AIC BIC logLik deviance df.resid
## 1257.9 1302.5 -619.9 1239.9 1044
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.5281 -0.9697 0.2735 0.8114 1.4162
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 1.117 1.057
## Number of obs: 1053, groups: participant, 23
##
## Fixed effects:
## Estimate Std. Error z value
## (Intercept) -0.6736 1.7554 -0.384
## numRatio 0.6956 0.6269 1.110
## visual_controlPanamath 2.0501 2.2289 0.920
## blockblock2 0.4528 2.2526 0.201
## numRatio:visual_controlPanamath -1.0762 0.7931 -1.357
## numRatio:blockblock2 -0.5071 0.8019 -0.632
## visual_controlPanamath:blockblock2 0.5834 3.3334 0.175
## numRatio:visual_controlPanamath:blockblock2 0.5293 1.1625 0.455
## Pr(>|z|)
## (Intercept) 0.701
## numRatio 0.267
## visual_controlPanamath 0.358
## blockblock2 0.841
## numRatio:visual_controlPanamath 0.175
## numRatio:blockblock2 0.527
## visual_controlPanamath:blockblock2 0.861
## numRatio:visual_controlPanamath:blockblock2 0.649
##
## Correlation of Fixed Effects:
## (Intr) numRat vsl_cP blckb2 nmR:_P nmRt:2 vs_P:2
## numRatio -0.976
## vsl_cntrlPn -0.787 0.768
## blockblock2 -0.779 0.760 0.630
## nmRt:vsl_cP 0.771 -0.790 -0.975 -0.600
## nmRt:blckb2 0.763 -0.781 -0.600 -0.975 0.617
## vsl_cntrP:2 0.545 -0.512 -0.696 -0.702 0.651 0.658
## nmRt:vs_P:2 -0.526 0.538 0.665 0.673 -0.682 -0.689 -0.958## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: accuracy
## Chisq Df Pr(>Chisq)
## (Intercept) 0.1472 1 0.7012
## numRatio 1.2315 1 0.2671
## visual_control 0.8461 1 0.3577
## block 0.0404 1 0.8407
## numRatio:visual_control 1.8415 1 0.1748
## numRatio:block 0.3999 1 0.5272
## visual_control:block 0.0306 1 0.8611
## numRatio:visual_control:block 0.2073 1 0.6489
## $emtrends
## block = block1:
## visual_control numRatio.trend SE df asymp.LCL asymp.UCL z.ratio p.value
## custom 0.696 0.627 Inf -0.533 1.924 1.110 0.2671
## Panamath -0.381 0.487 Inf -1.334 0.573 -0.782 0.4341
##
## block = block2:
## visual_control numRatio.trend SE df asymp.LCL asymp.UCL z.ratio p.value
## custom 0.189 0.501 Inf -0.793 1.170 0.377 0.7064
## Panamath -0.358 0.687 Inf -1.705 0.988 -0.522 0.6020
##
## Results are averaged over the levels of: numRatio
## Confidence level used: 0.95
##
## $contrasts
## block = block1:
## contrast estimate SE df asymp.LCL asymp.UCL z.ratio p.value
## custom - Panamath 1.076 0.793 Inf -0.478 2.63 1.357 0.1748
##
## block = block2:
## contrast estimate SE df asymp.LCL asymp.UCL z.ratio p.value
## custom - Panamath 0.547 0.850 Inf -1.120 2.21 0.643 0.5201
##
## Results are averaged over the levels of: numRatio
## Confidence level used: 0.95## $emmeans
## numRatio = 2.5, block = block1:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.744 0.0775 Inf 0.567 0.866 0.5 2.619 0.0088
## Panamath 0.605 0.0817 Inf 0.439 0.749 0.5 1.244 0.2136
##
## numRatio = 3.0, block = block1:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.804 0.0652 Inf 0.646 0.903 0.5 3.410 0.0006
## Panamath 0.558 0.0842 Inf 0.393 0.712 0.5 0.688 0.4917
##
## numRatio = 2.5, block = block2:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.562 0.0851 Inf 0.395 0.717 0.5 0.725 0.4685
## Panamath 0.820 0.0629 Inf 0.664 0.913 0.5 3.558 0.0004
##
## numRatio = 3.0, block = block2:
## visual_control prob SE df asymp.LCL asymp.UCL null z.ratio p.value
## custom 0.585 0.0840 Inf 0.417 0.736 0.5 0.997 0.3189
## Panamath 0.792 0.0693 Inf 0.625 0.897 0.5 3.177 0.0015
##
## Confidence level used: 0.95
## Intervals are back-transformed from the logit scale
## Tests are performed on the logit scale
##
## $contrasts
## numRatio = 2.5, block = block1:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 1.897 1.006 Inf 0.6714 5.361 1 1.208
## p.value
## 0.2269
##
## numRatio = 3, block = block1:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 3.250 1.741 Inf 1.1372 9.286 1 2.200
## p.value
## 0.0278
##
## numRatio = 2.5, block = block2:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 0.282 0.154 Inf 0.0963 0.825 1 -2.312
## p.value
## 0.0208
##
## numRatio = 3, block = block2:
## contrast odds.ratio SE df asymp.LCL asymp.UCL null z.ratio
## custom / Panamath 0.371 0.202 Inf 0.1276 1.076 1 -1.825
## p.value
## 0.0680
##
## Confidence level used: 0.95
## Intervals are back-transformed from the log odds ratio scale
## Tests are performed on the log odds ratio scaleAnalyses
##
## Pearson's product-moment correlation
##
## data: .$age and .$Panamath
## t = 2.7574, df = 20, p-value = 0.01215
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## 0.1325611 0.7749646
## sample estimates:
## cor
## 0.524837##
## Pearson's product-moment correlation
##
## data: .$age and .$custom
## t = 4.6606, df = 20, p-value = 0.0001506
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## 0.4310708 0.8765100
## sample estimates:
## cor
## 0.7215437##
## Pearson's product-moment correlation
##
## data: .$custom and .$Panamath
## t = 5.6948, df = 20, p-value = 0.00001423
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## 0.5458695 0.9072530
## sample estimates:
## cor
## 0.7864759