I compiled a questionnaire testing three components of EL, attitude, reported behaviour and knowledge. To test the first two components I used 5-response Likert questions and for knowledge, multiple choice questions. I reported the averages of each component in my research report, but my study leader want a composite "number" for the three components. Can it be done (given the different types of questions) and if it can, how should I weight the components? Any wisdom/literature will be appreciated!
Hennie:
Perhaps these studies can give you an idea of what is possible in answering your question. Hope this link is helpful for you.
Best regards,
Debra
https://www.researchgate.net/publicliterature.PublicLiterature.search.html?type=keyword&search-keyword=compiling+a+composite+number+for+components+of+Environmental+Literacy&search-abstract=&search=Search
Dear Hennie
You can do several statistical analysis to determine factors. The most used is the analysis of main components
Reguards
Dear Hennie,
You can also see:
https://www.researchgate.net/publication/283700257_A_influencia_das_concecoes_sobre_a_literacia_cientifica_no_desenvolvimento_da_cidadania
https://www.researchgate.net/publication/284190138_A_importancia_da_literacia_cientifica_para_os_alunos_do_ensino_secundario_das_regioes_Norte_e_Centro_de_Portugal
Best regards.
Article A influência das conceções sobre a literacia científica no d...
Conference Paper A importância da literacia científica para os alunos do ensi...
Dear Ana
Thanks for the paper
I did not know and I find an interesting contribution.
Best regards.
Dear David, Ana and Debra
Thank you for the answers, I found them really helpful.
Regards
Hennie
Because the knowledge component has a different scoring system, you cannot simply add it to the other two. One option would be to calculate the total score for each of the three, and then transform them all to Z-scores before combining.
With regard to weighting, you should examine the correlations among the three variables. If those correlations are all reasonably similar, then you can add the variables (i.e., give them equal weight). But if one of the variables has higher (or lower) correlations with the others, then you might want to consider factor analysis to get weights for the variables.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire