How to prepare index using factor score (arrived from factor analysis)

[Anita Makkar]

Respected sir

i wish to know abt preparation of index using factor score

i have applied factor analysis and found there are 4 major factors that can be further used
now i want to prepare an index using these scores

how it can be prepared

further how to find out a Taylor value in the index so that a range can be defined to categorize the data

with regards
Anita Makkar
Research Scholar
Haryana School of Business
GJU S&T, Hisar

[Neeraj Kaushik]

Dear Anita
I've no idea about What is Index? Where is it used?
Plz explain what it is?
Regards
Neeraj

--
You received this message because you are subscribed to the Google Groups "DataAnalysis" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataanalysistrai...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[VARUN ARORA]

Are you talking about discriminant function analysis????

[Anita Makkar]

Sir
here index is a statistical indicator representing the value of number in a given series

in factor score some have positive values and others has negative value
how to convert them in a uniform way so that they can be used for comparison with each other

[VARUN ARORA]

You cannot generate an index from Factor Analysis itself. Factor analysis is not a tool to develop an index, theoretically, it provides amount of covariances/correlations and hence it should be able to generate an index but remember these covariances and correlations are relative in nature - seeking the mutual association among different variables in question and necessarily not show same relationship with the outcome. This is just like the situation while standing at a bus stop to reach to Nehru Place, you are told that you cannot go to Nehru Place directly but you will have to first reach to Connaught Place - you are told that bus no. out of the 20 buses that carry you from that bus stop bus nos. 258, 312, 415 and 366 go to Connaught Place and then out of 112 buses carrying you  from that bus stop, bus No. 56, 72, 44 and 31 take you to Nehru Place i.e. bus nos. 258, 312, 415, 56, 72, 44 and 31 form the part of your solution - and this is where the role of factor analysis ends in a normal situation i.e. reduction of factors involved in your goal achievement from 132 to just 8.  That is why factor analysis is called data reduction tool i.e. sieving out those items which are available in the universe but which are not part of solution to your problem. Once factor analysis is over - you can generate the index using either logistic regression, linear regression or discriminant function analysis as the case may be depending on the nature and type of problem and data collected by you.


VARUN

[Anita Makkar]

Respected sir

i m also using factor analysis to combine the variables that are closely related (correlated) with each other.
further i wish to use factor scores that are available in the input sheet to prepare an index just to get benchmark value on the basis of which i can categorize the data in different categories

take the example of regional disparity in haryana
in haryana there are 21 districts there is disparity in these districts on the basis of agriculture, industries, etc
suppose 30 variables that are related to agriculture, industries and others
we apply factor analysis to combine highly correlated variable and get factors relating to agriculture and industries
now we get factor score in the input sheet of these factors against each district
some of score having positive value, others have negative value
how to create a statistical series i.e. index on the basis of which we can further categorize the districts into developed, moderately developed and undeveloped areas????

with regards
Anita Makkar

[VARUN ARORA]

For sure. We can on the basis of scale developed following factor analysis classify the districts into categories as indicated - but for that we need to validate the scale first. I understand you must have checked the sphericity and BI.

Varun

[Anita Makkar]

good morning sir

how to develop the scale that u said on the basis of scale developed through factor analysis
there is some formula through which an index can be prepared using factor score

what is Talyor value? for what purpose it is used?

Regards
Anita Makkar

[VARUN ARORA]

Dear Anita,

After factor analysis, you rule out certain items as they are rendered redundant, you reorgnize them on the basis of their congruency with each other into some dimensions. Thus the reduced scale is the accepted scale.

Scale development is as much a scientific exercise as it is artistic. The skillful understanding of the dimensions/factors make it relevant to understand and for each scale the methodology can be defined in terms of steps - there are objective tests to check and validate whether you are going in right direction as stated earlier. However, the actual inception of a scale for a working problem depends on much of skillful use of the factors obtained and subjective knowledge.

If your data set can be shared then we would be able to comment more on that.

2. Taylor value - With respect to Taylor value. It seems to be referring to Taylor's value added model. Is it so?

Varun