Dataset and Statistics

<pre>Dataset Description<br>File: data_BM_motorskills.csv<br>Format: CSV <br><br>The file contains 1,590 observations from 53 participants with 18 variables:<br>Participant_id: Unique ID for each participant (53 levels).<br>Trial_onset: Timestamp of trial start (ms).<br>Response_time: Timestamp of participant's response (ms).<br>Reaction_time: Response latency (ms).<br>Accuracy: Binary (1 = correct, 0 = incorrect).<br>Stimulus: Point-light display (PLD) video filename.<br>Experimental_task: Task type ("Action" or "Interaction").<br>Correct_response: Expected response key ("Down", "Left", "Right").<br>Response_key: Actual key pressed by participant.<br>Group: Autistic trait group ("High", "Medium", "Low" based on AQ).<br>AQ_binary: Raw Autism Quotient score.<br>Age: Participant's age (years).<br>Gender: Binary ("f", "m").<br>IQ_RPM: Raven's Matrices score (processed scores).<br>Handedness: Binary ("Left" or "Right").<br>NHPT_Right_Hand, NHPT_Left_Hand: Nine-Hole Peg Test time (seconds) for each hand.<br>Noncompliance: Binary flag for noncompliant trials (0 = valid).<br><br>#####<br><br>Analysis Code Overview<br>The script analysis.R performs data cleaning, demographic checks, and mixed-effects modeling for accuracy and reaction time (RT) analysis. Below are the key steps:<br>RSudio version: 2023.03.0 Build 386<br>Libraries: Uses dplyr, tidyr, lme4, lmerTest, emmeans, performance, interactions.<br><br>1. Data Preparation<br>Removes non-compliant trials (Noncompliance != 1).<br>Excludes left-handed participants and intermediate AQ group (Group != "Medium").<br>Drops NA values and unused factor levels.<br>2. Demographic Checks<br>Compares groups (High vs. Low AQ) on:<br>Gender distribution (χ² test).<br>IQ and age (Wilcoxon tests).<br>3. Accuracy Analysis<br>Model 1: Tests effects of Group and Experimental_task on accuracy (binomial GLMM).<br>Model 2: Adds interaction with motor skills (NHPT_Right_Hand).<br>Follow-up: simple effect analysis via emmeans.<br>4. Reaction Time (RT) Analysis<br>Preprocessing:<br>Excludes RTs for incorrect trials.<br>Removes outliers (1.5×IQR rule) and z-scales RTs within tasks.<br>Models:<br>Model 1: Effects of Group and Task (LMM).<br>Follow-up: simple effect analysis via emmeans.<br>Model 2: Adds interaction with NHPT_Right_Hand.<br>Follow-up: Simple slopes analysis, trends analysis.<br>5. Model Comparison<br>Compares Model 1 and Model 2 for reaction time using AIC, BIC, and R².</pre>