Data from: Exploring Suicidal Behaviour Through Implicit Identity and Control Biases (Summary and Raw Trial-Level Data)

Description

This repository contains two data files from the analyses of the article "Exploring Suicidal Behaviour Through Implicit Identity and Control Biases: Findings from the Death-Implicit Association Test and its Novel Control-Adaptation" examining death-specific implicit associations related to the self-concept and locus of control among psychiatric inpatients with a history of suicidal behavior. The first file summarises key variables used in the analyses, including demographic information, implicit D-scores from two versions of the Death-Implicit Association Test (D-IAT; identity-based (iD-IAT) and control-based (cD-IAT) version), explicit psychological measures, and summarised clinical diagnoses. The second file contains raw trial-level data from both the iD-IAT and cD-IAT versions. Participant IDs were anonymised using randomly generated numbers to ensure privacy. These data were used to investigate differences in implicit cognition among subgroups of suicide attempters, to assess reliability and validity of the iD-IAT and cD-IAT, and their relations to explicit psychological factors. Files Included: rawdata_summarised.sav – Summary dataset with demographics, D-scores, questionnaire scores, and diagnosis categories. rawdata_all_trials.csv – Trial-level raw data for both iD-IAT and cD-IAT.

Steps to reproduce

1. Data Preparation Ensure both files (`rawdata_summarised.sav` and `rawdata_all_trials.csv`) are in your working directory. Load the SPSS summary data file using SPSS or a compatible statistical package (e.g., `haven` in R, `pyreadstat` in Python). Load the raw D-IAT trial-level data in R using 'read.csv'. 2. Statistical Analyses (SPSS) (All performed using SPSS GUI unless otherwise noted) (1) Group Differences Run one-tailed independent samples t-tests using: `Analyze` → `Compare Means` → `Independent-Samples T Test` - Test Variables: `id_iat_dscore`, `cd_iat_dscore` - Grouping Variables: `frequency`, `recency` - Define groups as needed (e.g., 1 = single attempters, 2 = multiple attempters or 1 = lifetime attempters, 2 = recent attempters). (2) Correlations between Implicit and Explicit Measures Use Spearman rank correlations across the full sample with: `Analyze` → `Correlate` → `Bivariate` - Variables: `id_iat_dscore`, `cd_iat_dscore`, `bss_sum`, `bdi_ii_sum`, `mbppas_sum`, `sbq_r_sum`, `ie4_ic_mean`, `ie4_ec_mean`, `pmh_sum` - Select `Spearman` correlation. (Missing data in scales were handled with: `Transform` → `Replace Missing Values`) (3) Hierarchical Logistic Regression Assess the predictive validity of each D-score for group classification: `Analyze` → `Regression` → `Binary Logistic` - Dependent: `frequency` or `recency` - Block 1: `age`, `gender` - Block 2: `mood_disorders`, `anxiety_disorders` - Block 3: `bss_sum` - Block 4: `id_iat_dscore` or `cd_iat_dscore` - Optional model: include both `id_iat_dscore` and `cd_iat_dscore` in final block for joint prediction. (4) ROC Curve Analyses Evaluate discriminant accuracy using: `Analyze` → `Classify` → `ROC Curve` - Test Variables: `id_iat_dscore`, `cd_iat_dscore` - State Variable: `frequency` or `recency` - Define value labels as needed (e.g., 1 = multiple, 0 = single). 3. Statistical Analyses (R) (1) Reliability Analysis Assess permutation-based split-half reliability using the `rapidsplithalf` R package on the raw trial-level data: library(rapidsplithalf) raw_data <- read.csv("insert_path_to_data") rel_iat <- rapidsplit( data = raw_data, subjvar = "ID", diffvars = "congruent", subscorevar = "block_number", aggvar = "latency", splits = 5000, errorhandling = list( type = "fixedpenalty", errorvar = "error", fixedpenalty = 600, blockvar = "block_number" ), standardize = TRUE ) print(rel_iat) plot(rel_iat, type = "all")

Institutions

• Universitat Duisburg-Essen
• Lunds Universitet
• Universitat Leipzig
• Universitat Bern
• Universitatsklinikum Wurzburg

Categories

Licence