Data for: It's about Time: Delay-Dependent Forgetting of Item- and Contextual-Information

Published: 27-08-2020| Version 1 | DOI: 10.17632/4n7j9ry468.1
Talya Sadeh


Data included: Exp_[1-3]_data.csv: Table with all experiments’ raw data (see description below) in long format. W_dictionary.xlsx: Dictionary with 800 words (in Hebrew), the wordpool for all three experiments and each word’s numeric code - the corresponding index in semantic distance matrix. W_Semantic_Distances.csv: A 800*800 matrix which holds the semantic distances between each pair of words in the dictionary. Semantic distances between each word pair were derived from the semantic network constructed by Kennet et al., (2011). In this model, semantic distance is measured by pathway length (=lowest number of paths needed to cross between two nodes), where a distance of 1 signifies high semantic similarity and increasing pathway length signifies lower semantic similarity.