Perceptions, uses and training needs regarding artificial intelligence among pre-service teachers at the University of Cantabria: survey dataset (2024/2025)

Description:This dataset contains the responses of 226 students enrolled in the Early Childhood Education (n = 104) and Primary Education (n = 122) teacher-training degrees at the University of Cantabria (Spain) to a questionnaire on perceptions, uses and training needs regarding artificial intelligence (AI). The data come from a quantitative study with a descriptive-comparative, cross-sectional design. A non-probability sample was obtained during the 2024/2025 academic year through an online questionnaire. Instrument An ad hoc questionnaire of closed-ended items across 14 sections was administered electronically via Microsoft Forms. The questionnaire also included 12 optional open-ended items, which are neither analysed in the associated study nor included in this deposit. The closed-ended items are grouped into eight dimensions: 1. Sociodemographic information: gender, age and degree programme.2. Interest in and familiarity with AI: one Likert item on familiarity and one dichotomous item on the wish to receive training, plus 5 Likert items on desired training topics, shown only to those who wish to receive training.3. Use of AI in the university context: one dichotomous item and 4 frequency items, shown only to those who report using AI.4. Information about AI: 2 Likert items and 1 ordinal frequency item on actively seeking information.5. Information sources: 2 Likert items on preferred source profiles and 8 frequency items on information media, shown only to those who seek information about AI.6. Perceived relevance of AI for personal and professional life (4 Likert items).7. Perceived impact of AI on society and science (5 Likert items).8. Views on introducing AI training at three educational stages: Early Childhood/Primary, Secondary/Upper Secondary/Intermediate Vocational Training, and University/Higher Vocational Training. This dimension comprises 3 dichotomous items, with 12 Likert items conditional on an affirmative answer. Attitude and appraisal items use a 4-point agreement scale (Strongly disagree, Disagree, Agree, Strongly agree). Use and information-behaviour items use a 4-point frequency scale (Never, Rarely, Often, Usually/Always). An even-numbered scale with no midpoint was chosen deliberately to avoid central tendency and elicit a clear position. The questionnaire was piloted with 15 students from both degrees who were not included in the final sample. The pilot aimed to identify comprehension problems and ambiguous wording, and to estimate completion time. Contents of the deposit - `data.csv`: anonymised data matrix (226 cases × 52 variables), one row per participant. UTF-8 CSV, comma-separated, integer values; missing values are empty cells.- `variables.xlsx`: codebook (in Spanish) describing each variable, its dimension, value codes, skip logic and a breakdown of missing values (not applicable vs. no response), with methodological notes. Usage notes Numeric codes are reversed relative to the order in which the scales were presented: for agreement items, 0 = Strongly agree and 3 = Strongly disagree; for frequency items, 0 = Usually/Always and 3 = Never. The items `problema_ciencia` and `ventaja_ciencia` have been recoded to this same order, as they followed a different one in the source file; the correspondence is documented in the codebook. Empty cells correspond both to items not shown because of the questionnaire's skip logic and to items left unanswered. Gender is included in its binary form, as used in the analyses; the three-category version has been excluded from the deposit for data minimisation. Ethical considerations Participation was anonymous, voluntary and based on informed consent. The data contain no personal identifiers, IP addresses, timestamps or any other information that could identify participants.