08102023

.Rproj.user/093035E2/pcs/files-pane.pper

@@ -5,5 +5,5 @@
             "ascending": true
         }
     ],
-    "path": "~/OSF/osf.io-ueryq/semanticdecision/frequentist_analysis/model_diagnostics"
+    "path": "~/OSF/osf.io-ueryq/semanticdecision/power_analysis/results"
 }

.Rproj.user/093035E2/pcs/source-pane.pper

@@ -1,3 +1,3 @@
 {
-    "activeTab": -1
+    "activeTab": 0
 }

.Rproj.user/093035E2/pcs/windowlayoutstate.pper

@@ -1,14 +1,14 @@
 {
     "left": {
-        "splitterpos": 482,
-        "topwindowstate": "HIDE",
-        "panelheight": 1041,
-        "windowheight": 1079
+        "splitterpos": 348,
+        "topwindowstate": "NORMAL",
+        "panelheight": 753,
+        "windowheight": 791
     },
     "right": {
-        "splitterpos": 684,
+        "splitterpos": 494,
         "topwindowstate": "MINIMIZE",
-        "panelheight": 1041,
-        "windowheight": 1079
+        "panelheight": 753,
+        "windowheight": 791
     }
 }

.Rproj.user/093035E2/sources/per/t/98902CB9

@@ -0,0 +1,26 @@
+{
+    "id": "98902CB9",
+    "path": "~/OSF/osf.io-ueryq/semanticdecision/bayesian_analysis/semanticdecision_brms_weaklyinformativepriors_exgaussian.R",
+    "project_path": "semanticdecision/bayesian_analysis/semanticdecision_brms_weaklyinformativepriors_exgaussian.R",
+    "type": "r_source",
+    "hash": "602524222",
+    "contents": "",
+    "dirty": false,
+    "created": 1696285940543.0,
+    "source_on_save": false,
+    "relative_order": 1,
+    "properties": {
+        "source_window_id": "",
+        "Source": "Source",
+        "cursorPosition": "7,0",
+        "scrollLine": "0"
+    },
+    "folds": "",
+    "lastKnownWriteTime": 1696285960,
+    "encoding": "UTF-8",
+    "collab_server": "",
+    "source_window": "",
+    "last_content_update": 1696285960873,
+    "read_only": false,
+    "read_only_alternatives": []
+}

.Rproj.user/093035E2/sources/per/t/98902CB9-contents

@@ -0,0 +1,156 @@
+
+
+# Part of Study 2: Semantic decision
+
+# Part of Bayesian analysis
+
+# Model with weakly-informative priors (SD = 0.2)
+
+
+library(dplyr)
+library(brms)  # Version 2.17.0
+
+# Read in data
+semanticdecision = read.csv('semanticdecision/data/final_dataset/semanticdecision.csv')
+
+# Define priors
+
+# In the function `set_prior` used below, the argument `class` often represents
+# groups of parameters. For instance, class 'b' refers to fixed effects.
+
+# Since most priors do not specify the negative/positive direction of effects,
+# distributions are first specified for entire classes, and those are
+# overridden wherever a direction is specified.
+
+weaklyinformative_priors =
+  c(set_prior('normal(0, 0.2)', class = 'Intercept'),
+    set_prior('normal(0, 0.2)', class = 'b'),
+    set_prior('normal(0.1, 0.2)', class = 'b',
+              coef = 'z_orthographic_Levenshtein_distance'),
+    set_prior('normal(-0.1, 0.2)', class = 'b',
+              coef = 'z_word_frequency'),
+    set_prior('normal(0, 0.2)', class = 'sd'),  # automatically truncated to keep positive values only
+    set_prior('normal(0, 0.2)', class = 'sigma'),  # automatically truncated to keep positive values only
+    set_prior('lkj(2)', class = 'cor')  # standard, regularising prior on random-effects covariance to aid convergence
+  )
+
+# Model
+
+semanticdecision_brms_weaklyinformativepriors_exgaussian =
+
+  brm(
+
+    # Dependent variable
+    z_RTclean ~
+
+      # FIXED EFFECTS
+
+      # Controlled lexical variables
+      z_word_frequency + z_orthographic_Levenshtein_distance +
+
+      # Controlled word concreteness, important because the task is a word concreteness assessment.
+      z_word_concreteness +
+
+      # Control interaction between word concreteness and vocabulary size to allow a rigorous analysis of the
+      # interaction between word co-occurrence and vocabulary size, and the interaction between word visual
+      # rating and vocabulary size.
+      z_word_concreteness : z_vocabulary_size +
+
+      # Control interaction between word concreteness and participant gender to allow a rigorous analysis
+      # of the interaction between visual rating and participant gender.
+      z_word_concreteness : z_recoded_participant_gender +
+
+      # Important control due to the relationship between this variable and vocabulary size (Pexman & Yap, 2018;
+      # http://germel.dyndns.org/psyling/pdf/2018_PexmanYap_SemCat_IndividualDifferences.pdf)
+      z_information_uptake +
+
+      # Control interactions with information uptake to allow a rigorous analysis of the interactions with
+      # vocabulary size specified below.
+      z_word_cooccurrence : z_information_uptake +
+      z_visual_rating : z_information_uptake +
+
+      # Main effects of interest
+      z_vocabulary_size + z_recoded_participant_gender +
+      z_word_cooccurrence + z_visual_rating +
+
+      # Interactions of interest
+      z_word_cooccurrence : z_vocabulary_size +
+      z_visual_rating : z_vocabulary_size +
+      z_word_cooccurrence : z_recoded_participant_gender +
+      z_visual_rating : z_recoded_participant_gender +
+
+      # RANDOM EFFECTS: maximal structure constructed following the guidelines of Brauer and Curtin (2018;
+      # https://psych.wisc.edu/Brauer/BrauerLab/wp-content/uploads/2014/04/Brauer-Curtin-2018-on-LMEMs.pdf).
+      # Interactions only require random slopes if all variables involved vary within the same units.
+
+      # By-participant random slopes
+
+      # Below, the random slopes for control covariates (i.e., 'z_word_frequency', 
+      # 'z_orthographic_Levenshtein_distance' and 'z_word_concreteness') were removed 
+      # due to non-convergence, inspired by Remedy 11 from Table 17 in Brauer and 
+      # Curtin (2018). However, whereas Brauer and Curtin constrain such a removal 
+      # to cases in which the covariate does not interact with any effects of 
+      # interest, the random slopes for 'z_word_concreteness' were removed below 
+      # because the interactions between this covariate and the effects of interest 
+      # were control covariates, not interactions of interest. That is, they 
+      # were not critical to the research question.
+
+    (z_word_cooccurrence + z_visual_rating | Participant) +
+
+      # By-word random slopes
+
+      # Below, the random slopes for the control covariate (i.e., 'z_information_uptake') 
+      # were removed due to non-convergence, inspired by Remedy 11 from Table 17 in Brauer 
+      # and Curtin (2018). However, whereas Brauer and Curtin constrained such a removal to 
+      # cases in which the covariate does not interact with any effects of interest, the 
+      # random slopes for 'z_information_uptake' were removed below because the interactions 
+      # between this covariate and the effects of interest were control covariates, not 
+      # interactions of interest. That is, they were not critical to the research question.
+
+    (z_vocabulary_size + z_recoded_participant_gender | Word),
+
+    data = semanticdecision, prior = weaklyinformative_priors,
+
+    # Following the prior predictive checks, an ex-gaussian distribution is set below.
+    # This distribution is well suited for the analysis of response times (Bürkner et al., 
+    # 2022, https://cran.r-project.org/web/packages/brms/brms.pdf; Rodríguez-Ferreiro 
+    # et al., 2020, see supplementary materials via https://doi.org/10.7717/peerj.9511). 
+    # Furthermore, the use of link functions, instead of transformations of the 
+    # dependent variable, largely preserves the interpretability of the coefficients 
+    # (Knief & Forstmeier, 2021; Lo & Andrews, 2015, https://doi.org/10.3389/fpsyg.2015.01171).
+
+    family = exgaussian(),
+
+    seed = 123,  # allow exact replication of results
+    warmup = 2000,  # warmup iterations
+    iter = 8000, chains = 16,  # (iter - warmup) x chains = total post-warmup draws (brms v2.17.0)
+    cores = 5,  # parallel computation
+    control = list(adapt_delta = 0.99, max_treedepth = 15)  # facilitate convergence
+  )
+
+
+# NOTE: Output takes up more than 15G, so it was removed from the project due to the 50G
+# limit on OSF.io. If you wish to save the output, please uncomment the lines below.
+# saveRDS(semanticdecision_brms_weaklyinformativepriors_exgaussian,
+#         'semanticdecision/bayesian_analysis/results/semanticdecision_brms_weaklyinformativepriors_exgaussian.rds')
+
+# Model summary
+summary(semanticdecision_brms_weaklyinformativepriors_exgaussian) %>%
+  saveRDS('semanticdecision/bayesian_analysis/results/semanticdecision_summary_weaklyinformativepriors_exgaussian.rds')
+
+# Posterior predictive check
+# Next, a posterior predictive check is performed to examine the sensitivity of 
+# the model to priors varying in informativeness (for similar procedures, see 
+# Rodríguez-Ferreiro et al., 2020, https://doi.org/10.7717/peerj.9511; 
+# Stone et al., 2020, https://doi.org/10.7717/peerj.10438). The posterior 
+# predictive checks from all models are compared in the folder 
+# 'posterior_predictive_checks'.
+pp_check(semanticdecision_brms_weaklyinformativepriors_exgaussian, ndraws = 100) %>%
+  saveRDS('semanticdecision/bayesian_analysis/posterior_predictive_checks/semanticdecision_posteriorpredictivecheck_weaklyinformativepriors_exgaussian.rds') 
+
+# Plot posterior distribution of each fixed effect (i.e., effects commencing with 'b_')
+mcmc_plot(semanticdecision_brms_weaklyinformativepriors_exgaussian, 
+          type = 'areas', variable = '^b_', regex = TRUE) %>%
+  saveRDS('semanticdecision/bayesian_analysis/results/posteriordistributions_semanticdecision_weaklyinformativepriors_exgaussian.rds')
+
+

.Rproj.user/093035E2/sources/prop/3406D2DF

@@ -0,0 +1,6 @@
+{
+    "source_window_id": "",
+    "Source": "Source",
+    "cursorPosition": "7,0",
+    "scrollLine": "0"
+}

.Rproj.user/093035E2/sources/prop/INDEX

@@ -5,6 +5,7 @@
 ~%2FOSF%2Fosf.io-ueryq%2Fsemanticdecision%2Fbayesian_analysis%2Fposterior_predictive_checks%2Fsemanticdecision_posteriorpredictivechecks.R="2EA34ACD"
 ~%2FOSF%2Fosf.io-ueryq%2Fsemanticdecision%2Fbayesian_analysis%2Fprior_predictive_checks%2Fsemanticdecision_priorpredictivecheck_diffusepriors_exgaussian.R="434AA05B"
 ~%2FOSF%2Fosf.io-ueryq%2Fsemanticdecision%2Fbayesian_analysis%2Fsemanticdecision_brms_diffusepriors_exgaussian.R="98D87F8C"
+~%2FOSF%2Fosf.io-ueryq%2Fsemanticdecision%2Fbayesian_analysis%2Fsemanticdecision_brms_weaklyinformativepriors_exgaussian.R="3406D2DF"
 ~%2FOSF%2Fosf.io-ueryq%2Fsemanticdecision%2Fcorrelations%2Fsemanticdecision_correlations.R="3EDC62B8"
 ~%2FOSF%2Fosf.io-ueryq%2Fsemanticdecision%2Fdata%2Fsemanticdecision_data_preparation.R="5F25266B"
 ~%2FOSF%2Fosf.io-ueryq%2Fsemanticdecision%2Ffrequentist_analysis%2Flexical_covariates_selection%2Flexical_covariates_selection_semanticdecision_lmerTest.R="A4DD5B32"

semanticdecision/bayesian_analysis/semanticdecision_brms_weaklyinformativepriors_exgaussian.R

@@ -109,6 +109,8 @@ semanticdecision_brms_weaklyinformativepriors_exgaussian =
 
     (z_vocabulary_size + z_recoded_participant_gender | Word),
 
+    data = semanticdecision, prior = weaklyinformative_priors,
+
     # Following the prior predictive checks, an ex-gaussian distribution is set below.
     # This distribution is well suited for the analysis of response times (Bürkner et al., 
     # 2022, https://cran.r-project.org/web/packages/brms/brms.pdf; Rodríguez-Ferreiro