While in the process of factor analysis and reliability tests, I find that one of the factors (with only two items loading) have a 'Cronbach's α' being almost 0.4; the other 3 factors of the concept have the typically acceptable '0.7 or more' (with 2 to 5 items each).
Just to note, the loadings for this factor are above 0.7.
Does anyone know how exactly I should treat the factor with the low α? Do I just disregard it? And if so, what about the total variance explained? Do I just subtract this problematic factors?
Thank you very much for your answer. Does this mean that I should not check reliability, or that I should check it differently?
Based on the literature, the factor I am testing is an important one and widely discussed. Thus, I would like to use this despite the low reliability of the scale. I have found Van de Ven and Ferry (1980) arguing that coefficient alpha could be accepted from 0.35 to 0.90 based on different conditions.
thank you very much, yes I am measuring internal consistency.
after an exploratory analysis (Principal component analysis) there was this problematic factor with only two items loaded... the number of items might be one of the reasons for the low alpha.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability