How 'Factor Analysis' works for Marketing Research Data?

Let’s see conversation between the Statistics Professor Aditya Kaul and student Monil Taploo.

Monil a passionate marketing student, currently working on his research paper, he has done with data collection, his research design is causal in nature, his analysis plan is ready, but he has to deal with 100 variables exist in his datafile. He has taken appointment of his professor Aditya Kaul. Basically, he wanted to get guidance on his approach.

Professor Aditya is sitting in his cabin, cabin’s the window facing the garden filled with lush green apple trees. Sunlight is coming through French styled window and it is constantly pouring the energy inside the room.

Prof. Aditya Kaul is reviewing the analysis plan given by Monil, the moment he has read a name of statistical technique he has asked.

“Why do we need Factor Analysis?”

Monil: "For Dimension Reduction! sir!"

Aditya: "Then, why can’t we use this every time we are carrying substantial number of Variables."

Monil: "Yes, we can use that, can’t we? "(replied with cracking voice showed up lack of confidence)

Aditya: "See Try to understand the usefulness of Factor Analysis.

Factor Analysis can help you to group the variables with minimum amount of loss of information."

Monil Thinking in his mind (ohh Dear Kaul sir …That is true and I know this already)

Monil couldn’t show some excitement on his face by that stamen of Prof Kaul.

Aditya further said “Look Monil, when you have 100 variables in your data file”

“And you do have your Analysis Plan in your hand but carrying those variable and execution of the plan becomes tedious or time taking"

“In many cases variables are having the similar pattern, so at sample data level its repetition of many patterns”

"In this situation, inherently you will start looking for innovative idea, about how you can ease the rigorous execution of the analysis plan. How difficult it is to carry whole colony of variables for analysis that is divided into several steps !!”

“Correct me If I am Wrong”

Monil nodded “Yes sir”

but his facial expressions were revealing confused state of his mind. His twitching eyebrows, starry eyes, and shallow breathing indicate not getting through the exact narrative of Prof Kaul.

Monil is not having technical background, but he wants to understand.

Aditya Kaul recognises the fact that Monil Taploo is confused with the narration.

But What to do is the question.

Aditya Kaul paused briefly, took a deep breath, and looked out his cabin window. He noticed beautiful apple trees and several women carrying bags walking through the garden.

Aditya got one idea and he started explaining with enthusiasm, with that hope, this time he would be able to explain this concept to Monil Taploo.

Facing Monil, Prof. Aditya Kaul said with an enthusiastic smile, "Monil, you mentioned you're planning a one-day trip to Manali."

Monil has shown some level of surprising expressions on his face.

(Monil was thinking What that has to do with Factor Analysis???)

He said “Yes Sir” with little bit surprise in his voice

Aditya: "Tell me how many blue Jeans are you going to carry along with you?"

He found this question weird but said “ONLY ONE” “yes”. beside this he wore blue jeans today.

Aditya: "How Many Blue-Jeans you do have at your Home?"

He said “well many... if I count them, total 6”

Aditya: “Then why you are keeping only on this trip?”

He said, “Yes because, they are blue jeans, though the shades are different”

“I will pick only one, what is the point of keeping all of them on the trip, even if I am on 2- or 3-days trip, I prefer to keep only one blue-jeans, rather I will experiment with different color like I can carry one baclk jeans too”

Aditya Kaul: “Ok that sounds logical”

“Then, why do you carry all those 100 variables in your analysis, while many of them are correlated, their pattern of increase/decrease is same or similar like the shades of your blue jeans are different more or less”

“Dear, you may perceive like this way, Applying Factor Analysis, we are grouping the variables which are showing the high correlations among themselves. Or we are picking up the most relevant, or say, picking up the most important variable from the bundle of highly correlated variables. Off course there is minimum amount of loss of information. But what is a point of carrying all of them in your analysis plan for Causal research?”

The moment of knowing the logic brings smile on Monil’s face.

Finally, he understood concept of Factor analysis. What he knew before was the procedure to carry out the Factor Analysis using tools like R, AMOS, SAS etc.

“Ohh ... Sir I have done this Factor Analysis So many times, Now I Know Why do we need that”

With enthusiasm he asked “sir Correct me I am wrong”

“This is the reason, we do Exploratory Factor Analysis If we don’t have any idea about the constructs and manifest variables loaded to it, right”

“secondly, I will check which set of few variables are having high correlations with other variables, if they are meaning they do represent their pattern on larger extent. So these are my key variables. Rather than using the whole colony of variables I will focus on those key who are having potential to explain the correlation of all others”

Aditya: “Yes dear, if you don’t have theory in your mind, or, you do not have it after doing literature study even, then study the correlations among the variables and derive the factors using Exploratory Factor Analysis (EFA)”

Aditya asked, Curiously and with the intention of Cross-checking Monil “If you do have theory in your hand then ???”

Monil Confidently said “Sir then we may choose Confirmatory Factor Analysis, we may draw the factors, we may load the variables to each factors as per our theoretical understanding and carry out analysis to confirm it”

“Well in this we need to ensure Reliability and Validity”

“Then After defining factors, we can set up relationship structure as per our theoretical understanding and check what we are thinking is correct or not ... I mean theoretical model is fitting well on data or not”

Aditya “great”

They both had smile on their faces while on the same moment chai-Wala knocked the Door, he stepped into the room, reached to the table, and served tea.

They both were enjoying the tea near the window facing the garden with some apple trees. Sun is radiating the golden energy and professor is navigating the energy of good thoughts.

Jay Hind!

Rooshabhkumar

References

Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis (6th ed.). Pearson Prentice Hall.