# Directory structure ## Data The data from LAPS I used in chapter 3 is available in the file `laps-i.csv`. The LAPS II data are stored in several different files. `all_data.csv` is the master data set that contains all the information collected in the study. Each rows contains all of the data for a single participant. If you're interested in rerunning the main analyses or running some analyses yourself, it'll probably be easiest to use the file `laps2_full_dataset.csv` below, which is a cleaned-up version of `all_data.csv`. `NAReasons.csv` contains the reasons for missing data; see technical report, Chapter 10. The dataset was split into a training set and a validation (or test) set. Refer to the technical report for the reasons why. The technical report also explains what each variable in these datasets refers to. Training set: - `construct_scores_training.csv`: This file was created using `R Scripts/04_ConstructScores.R` and contains summary scores for the cognitive and questionnaire-based constructs. - `training_set.csv`: This file was created using `R Scripts/03_TrainingTestSplit.R` and contains item-level responses to the cognitive and questionnaire items as well as all other variables collected. Test set: - `test_set.csv`: This file was created using `R Scripts/03_TrainingTestSplit.R` and contains item-level responses to the cognitive and questionnaire items as well as all other variables collected. The full data set was compiled after the predictive models were run on the test set. It contains variables at both the item and construct level. For instance, there is a variable specifying each participant's answer to the third MLAT item at the second data collection as well as a variable specifying the participant's overall MLAT score at the second data collection. See the technical report for the full code book. - `construct_scores.csv`: This file ust contains the summary scores for the cognitive and questionnaire-based constructs. - `laps2_full_dataset.csv`: This contains all variables. You can also install the datasets in R. To that end, go to https://github.com/janhove/LAPS and follow the instructions. ## HTML output of RMarkdown scripts The rendered HTML files for some of the R/RMarkdown scripts, so that the output of these scripts can be read more easily. - `05_PredictionT3_with_T1.html`: Corresponds to `R Scripts/05_PredictionT3_with_T1.Rmd`. ## R scripts The R scripts with which the data cleaning and analysis can be reproduced. - `01_DataCleaning.R`: For reading in the data and wrangling it into an analysable format. The direct links to the datasets have been replaced by `link.to.the.data.set`. Contact jan.vanhove@unifr.ch if you need the raw, unwrangled files. - `02_Checks.R`: Some preliminary checks. - `03_TrainingTestSplit.R`: This script splits the dataset into training and test sets. The split was effected at random, but since the new R version, the same commands give rise to a different random selection. See https://blog.revolutionanalytics.com/2019/05/whats-new-in-r-360.html. The output of scripts 01 through 03 are the CSV files `test_set.csv` and `training_set.csv`, which were put into the working directory for the following scripts (see 'Datasets (training)' and 'Datasets (validation' above). - `04_ConstructScores.R`: This computes construct scores for the cognitive and questionnaire-based constructs and also estimates these scores' reliability (using just the training set). The resultant CSV file (`construct_scores_training.csv`) was put in the working directory (see 'Datasets (training)'). - `05_PredictionT3_with_T1.Rmd`: Predictive modelling of the English data at T3 using T1 information. (Main script for Chapter 4.) ## Figures Figures generated using the scripts. ## Technical report