Hi all!
Together with another researcher, I have developed a new type of self esteem measure, to use in Ghana and potentially other countries. We wanted to create a measure that is less abstract, and more related to the daily lives of the respondents. Instead of using a typical scale, with statements and a Likert-scale as answer options, we use scenario based questions. We describe certain situations that are reflective of the lived realities of young people, and then let them select how they would most likely respond to that situation. Answer options describe different responses, which reflect different levels of self-esteem.
We now would like to determine the convergent validity of this measure, by letting respondents fill out our self-esteem measure, as well as 2 or 3 established self-esteem measures and determine how they are related to each other. However, I have never conducted such a study before, so I am looking for some guidance. Maybe some tips on how to go about it, literature that gives further guidance or anything else that could be of use.
Thank you in advance!!
Article Verifying survey items for Construct Validity: A two-stage S...
Article Factor Structure, Reliability, and Convergent Validity of th...
https://www.emerald.com/insight/content/doi/10.1108/IJCHM-10-2017-0649/full/html
https://www.researchgate.net/post/How-can-I-carry-out-convergent-or-discriminant-validity-for-a-questionnaire
Dear Sharon,
Convergent validity could be tested by performing a correlation analysis (analyzing the correlation between the scores of your newly developed measure and those of another tool that reflects the same or similar construct).
The following article could be helpful:
Strauss ME, Smith GT. Construct validity: advances in theory and methodology. Annu Rev Clin Psychol. 2009;5:1-25. doi: 10.1146/annurev.clinpsy.032408.153639.
Best,
The responses suggesting that your measure be administered at the same time as established measures would be a reasonable way to explore convergent validity. One thing to keep in mind would be to alternate which measure was administered first to avoid any bias in the order of administration.
One of the old examples of such bias was asking whether reporters should have unfettered access to information in the US or Russia. Depending on whether the question was asked about the US first resulted in very different response patterns than if the question was first asked about Russia.
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts