Hi,
I have performed factor analysis in order to categorize 31 variables into major groups, the result yielded 9 factors however, in the rotated component matrix I have found two factors with only one attribute (one variable) but with a high loading value. Is it normal to take a factor with one variable or I have to delete this factor??
Hello Roumeissa Salhi,
Having high loading on the factor is good, especially if the scale is being designed for individual diagnostic (DeVellis, 2003; Nunnally, 1978). However, factors with less than three items are considered “weak and unstable” (Costello & Osborne, 2005, p. 4), and would necessitate ignoring the factor in further analysis, in accordance with Velicer’s and Fava’s (1998) posture that factors with less than three variables with high loadings should not be interpreted.
Blessings,
Hazel Traüffer
You're welcomed. You should write that up in your reporting and include the alpha value.
Blessings,
Hazel Traüffer
Dear Hazel Trauffer ,
Should I include the alpha value for all the items of each factor or only for the overall?
Thank you and regards,
Roumeissa Salhi
Dear Roumeissa,
You should present the Structure Matrix of 9-Factor Solution (i.e. the number of factors with which you finally ended, as well as the Table showing the Total Variance Explained and a Scree plot, showing the plot of the factors. Then, write it up, naming each factor from the item that loads the highest on that factor. Be sure you explain how much of the variance is accounted for by each factor. Once you've done that, discuss the Reliability analysis (Cronbach’s alpha) for each factor; this is where you discuss the overall score, explaining whether or not the score would be different, if items were deleted.
I hope that helps ...
Blessings,
Hazel Trauffer, PhD.
It sounds like your preliminary work was deficient. Did you have an a priori model? It is a bit strange to get 9 factors from just 31 variables...particularly if two of these are singe variable factors. I would expect, perhaps, around six factors each with around four or five variables. It does depend on when the single variable factors emerged in the factor rotations if they are factors 8 and 9 then get rid of them ( particularly if the first seven factors make sense - in terms of your hypothesis)....but, if they appear early in the rotation...which would be very strange... ( I would recommend Equamax with Kaiser normalization) , then they could actually be just what your study is about....but it defeats the purpose of undertaking a factor analysis and (no doubt) your subsequent Cronbach Alpha reliability calculation. I think you need to go back to the drawing board!
Hello Peter Bycroft
I have some models 'Doloi, H., Sawhney, A., Iyer, K. C., & Rentala, S. (2012), Bagaya, O., & Song, J. (2016). Empirical study of factors influencing schedule delays of public construction projects in Burkina Faso. Journal of Management in Engineering, 32(5), 05016014. Analysing factors affecting delays in Indian construction projects. International journal of project management, 30(4), 479-489.'
The single variable factors appeared in the last two factors. I have deleted them, but it appears a new problem that the 7th factor contains only 2 variables in which one variable is important so I can't delete it, and I have see in the article of(doloi et al. 2012)that they keep the last factor which is with 2 variables. so, can I keep this last?or what should I do?
Cordialy
Roumeissa Salhi
Dear Roumeissa Salhi
Under the assumption that Variable and Subject sampling are adequate, factors with one variable and with a high loading value will be easier to interpret. However, it is important to consider the amount of variation explained by that factors under the rotated model and the variable in the context of your model.
Dear Roberto Molteni
In first, the nine factors extracted explained 69.59% of the total variance, however after deleting the 8th and the 9th factors, the amount variation explained by the 7 news factors became 68,014%.
Best regards,
Roumeissa Salhi
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs