How to Create Index Using Principal Component Analysis

CrunchEconometrix videos should be supported by relevant readings from econometrics textbooks, journal articles and other resources to properly harness the simplicity of the video tutorials.
CrunchEconometrix P.E.R.B.A. cruncheconomet...
CrunchEconometrix LinkedIn Page: / cruncheconometrix
For Dataset: www.researchga...
For Stata Do-file: www.researchga...

[@lucifergaming2229]

Hello Dr. Bosede Ngozi Adeleye, I hope you're doing well. Recently, I watched your video of PCA. I constructed 4 indices, in which one index value is 0.3382. So, now my question is, am i supposed to carry on with this exceptional result or can you provide me the link to any other research(es) whose results are like this which will support my this outcome. Or if you can provide me any information for supporting this result. It would be a great help to me. I hope to see your kind response ASAP. thank you.

[@stansilousmunemo3146]

You are a blessing.

[@CrunchEconometrix]

Thanks so much for your kind words. It is deeply appreciated, Sir! 💖

[@RishabhSMukherjee]

Mam your step by step explaination is really awesome.... Just one question can this be also done on a single country since I am working with the data set of a single country trying to create in a similar manner but there are like 16 dimensions a time series data over 18 years time period ?? If yes then instead of KMO for final check what should be done ??

[@CrunchEconometrix]

Yes you can. Same step. From start to finish.

[@RishabhSMukherjee]

Also mam what should be done if it is an unbalanced panel data ?

[@CrunchEconometrix]

It works with an unbalanced panel data.

[@LadyJammyDodgers]

Hi Ngozi, you are a great presenter. Glad to have come across your channel. I am part of your lay audience and would like to know, what is this technique used for?

[@CrunchEconometrix]

Thanks for the kind words, @LadyJammyDodgers. I explained the PCA technique in the introductory video. Kindly watch for in-depth knowledge.

[@fichrieconometrics]

so helpful thanks!

[@CrunchEconometrix]

U're welcome! 🥰

[@divyasharma42]

But, its a panel dataset, might have cross sectional variations. Do we need to declare the dataset as panel before proceeding or is this process carried out on a pooled basis?

[@CrunchEconometrix]

You may declare the dataset as a panel data.

[@Jatindersingh-wo5hf]

Hi, don't we need to normalize variables before applying PCA? How to deal with variables in nominal scale and ratio scale?

[@CrunchEconometrix]

Watch the video again to hear me mention Varimax and Kaiser Normalization. You may want to read more about using PCA on variables with different scales.

[@ParthBasankar]

What can I do to create an index if more than one principal component is showing an eigenvalue above 1?

[@CrunchEconometrix]

Use the component with the highest eigenvalue.