My name is Filippos Filippou and I teach Greek Traditional Dance at the Democritus University of Thrace/Greece. I am currently engaged in the construction of a diagnostic test of conceptual knowledge and motor skills.
The tool will be aimed at adult youth with Down syndrome and will include 50 pictures since we took into account that these individuals have better visual than auditory memory. The 50 images were derived from a question posed to 45 dance teachers, graduates of the Departments of Physical Education Science, specializing in Greek Traditional Dance and the curricula of the Ministry of Education. The aim of the study is to examine the effect of two different programs - Greek traditional dance one and movement games the other - on vocabulary learning.
The initial study involved 66 boys and girls aged 20-31 years who were diagnosed with Down's syndrome and mild mental retardation. Frequency analysis revealed that of the total participants, 22 subjects recognized the picture depicting the boy, 23 the picture depicting the girl, 24 the picture depicting the fingers, 1 person the shoulder, and scarf 5 people. Not one person identified any of the remaining 45 images.
To ensure the validity of a test (whether for images or any other type of assessment), you need to analyze different statistical measures. The key statistics and methods depend on what aspect of validity you are evaluating:
1️⃣ Content Validity (Does the test measure what it should?)
✔ Expert Review & Item Analysis – Ensure that the test items (images) accurately represent the intended concept. ✔ Inter-Rater Reliability – If multiple evaluators are scoring, check agreement levels using Cohen’s Kappa or Intraclass Correlation (ICC).
2️⃣ Construct Validity (Is the test measuring the right construct?)
✔ Factor Analysis – Use Exploratory Factor Analysis (EFA) or Confirmatory Factor Analysis (CFA) to see if the test items (images) align with the intended constructs. ✔ Correlation with Established Tests – Compare results with a validated test using Pearson’s or Spearman’s correlation.
3️⃣ Criterion-Related Validity (Does the test predict real-world outcomes?)
✔ Concurrent Validity – Check how well the test correlates with an already validated measure. ✔ Predictive Validity – Use Regression Analysis to see if the test predicts future performance or behavior.
4️⃣ Reliability Measures (Indirectly Supports Validity)
✔ Cronbach’s Alpha (α) – Measures internal consistency (should be ≥ 0.7). ✔ Test-Retest Reliability – Use Intraclass Correlation (ICC) to check stability over time.
The above answer is an AI answer and very general. We can start a discussion. Three questions:
1. Is there a clear purpose and outcome? Extant literature?
2. Any existing measures to check concurrently?
3. This survey seems to be an assessment to choose the correct label. Are there choices? Timing? Approximations? Considering different levels of cognitive ability?
Recomendo a segunda resposta e acrescento sugerindo dois instrumentos: 1. Índice de Validade do Conteúdo (CVI) com especialistas.
2. Kuder-Richardson (KR-20) para itens binários (reconheceu ou não reconheceu) isso ajudar-te-á saber quem reconheceu ou não, uma vez que a pesquisa sugere fazer escolha da imagem correta. E como sugere a 3 discussão de David, deve se ter em conta os timing e aproximações e outros factores, que devem ser definidos para facilitar a validação do teste e evitar viés.
- Badges
- Science topic
- Similar topics
- Mathematics