dear researchers,
we need currently to measure the validity of an instrument (distinctive, predictive etc.) but with the usage of SPSS PROGRAM. please if you have some distintive article or ppt slides or courses which could help us to finish my colleague's thesis.
if you have practical steps to measure validity with the use of factor analysis we will also grateful to you dear researchers.
Hello there,
There are various types of validity foe questionnaires, and although some may be considered as more important than others, well... I would argue that similarly to fit indices in SEM - it is good practice to implement at least some of them instead of relying on one 'superior' type of validity.
CONSTRUCT VALIDITY - ability of a psychometric tool to measure what it is supposed to measure. This one is the most important !!!
e.g. Anxiety scale should measure anxiety, not PTSD
CONTENT VALIDITY - does a measure contain fairly full scope of issues pertaining to the concept it measures (e.g. depression)?
FACE VALIDITY - does a measure seems to pertain to a main topic? Perhaps a discussion with experts and non-experts would be advised here.
CRITERION VALIDITY - based on two components
A/ Predictive validity - Does a measure predict future (performance, behaviours, etc.)
B/Concurrent validity - Does a measure predict current performance, behaviour etc.
Also internal consistency (reliability expressed in Cronbach's alpha is extremely important and should be added to validity tests while assessing questionnaires.
Regards, Witold
No, SPSS does not have specific programs for the measurement of validity. As Witold explains, many forms of validity involves predictions about the relationship of a variable to other variables, and this is typically asses with correlation ands regression.
Factor analysis typically involves assessing the structure of the relationships among the variables that make up a scale or scales. You can test predictions about that structure using Confirmatory Factor Analysis (which requires the AMOS module in SPSS). If you don't need to do formal tests, then you can use "face validity" to assess statements such as: There is only one factor in the data.
As I know, to measure the validity of the measurement if we use spss, the common method is use factor analysis. SPSS just give us information the loading factors everh item we have. I suggest you use confirmatory factor analysis with LISREL or AMOS, so we can really have informations about our item. Is it good or not and fit or not with our construct.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS