Research Data for: Video Information Trustworthiness and Quality Assessment Scale (VITAL)

Description

This dataset provides the raw and processed data used for the development and validation of the Video Information Trustworthiness and Quality Assessment Scale (VITAL). The study was conducted on a total of 420 YouTube videos. The dataset is organized into two primary files to support the sequential validation process described in the associated research paper: 1. EFA Dataset (n=210): Contains the evaluation scores used for Exploratory Factor Analysis to determine the initial factor structure of the 10-item VITAL scale. 2. CFA Dataset (n=210): Contains a separate sample of video evaluations used for Confirmatory Factor Analysis to validate the one-dimensional structure. Data Collection and Scoring: The items for the VITAL scale were generated using a novel approach integrating Large Language Models (LLMs) such as ChatGPT, Gemini, and DeepSeek. Videos were evaluated by field experts using a 5-point Likert scale. The final 10-item scale demonstrated high internal consistency with a Cronbach's alpha of 0.903 and a McDonald's Omega of 0.911. This data supports the findings presented in the manuscript submitted to Computers & Education.

Steps to reproduce

Step 1: Item Generation: The initial item pool was created through a literature review and the use of five Large Language Models (ChatGPT, GEMINI, Co-Pilot, Perplexity, and DeepSeek) with prompted and unprompted guidance. Step 2: Content Validity: 10 field experts evaluated the 35-item draft using the Lawshe technique. Items with a Content Validity Ratio (CVR) below 0.62 were removed, resulting in an 11-item form. Step 3: Data Collection: Videos related to "Industry 4.0" were collected from YouTube using the YouTube API and Google Colab. Videos were filtered based on exclusion criteria (non-English, <60 seconds, >60 minutes, etc.). Step 4: Scoring: Two experts independently scored the videos using a 5-point Likert scale. Inter-rater agreement was confirmed using the Kappa method. Step 5: Statistical Analysis: Exploratory Factor Analysis (EFA) was performed on the first sample (n=210). Confirmatory Factor Analysis (CFA) was performed on a second, independent sample (n=210) to validate the single-factor structure. Analyses were conducted using IBM SPSS 25, IBM SPSS AMOS V24, and JASP 0.95.4.0.

Institutions

• Ondokuz Mayıs University

  Samsun, Samsun

Categories

Licence