Query regarding global single item in formative scale development.

[hitesh kumar]

I have used a 1-7 point scale to measure the items in my scale, however, to address the redundency procedure, i added a global single item that I measure for each dimension on a 1-10 point.

At the time of questionnaire development, I did this with thinking it would reduce CMB. 

Apart from this explanation, I don't have any other sufficient reason, also I am doubtful if this CMB reason would be valid or not. How can I justify the use of different anchors as I have already collected and analysed the data? 

Please guide me and suggest any reference material or study supporting my approach.