Student Mailbag - Week 2

Update: received evening today (28-Oct)

Dear Chandana,
Can we mention our assumptions while making the questionnaire? We feel that the rural market is an important target to be studied. However, since this is a web survey in English, this segment will be completely ignored. Hence we are assuming that the survey is targeted towards the urban market.
Please advise.
Regards,
S

My response:

Yes.
I will make a mention on the blog as well.
Our survey targets the urban SEC A only because of accessability issues. hence, we define our target population as Urban SEC 'A' people skewed towards the younger (20s and early 30s) side.
We will further make the assumption that the snowball sampling method that we will use will produce a probability sample representative of the target population.
In this light, all our results etc are valid and applicable only in the context of this target population we have defined.
Hope that clarifies.
Sudhir

----------------------------------------------------

Hi Prof. Sudhir,
Pardon my ignorance but there is a case to be read for Monday’s class.
Bank of America (Stanford GSB case M213)
I don’t think we have that in the course pack.
Is it available somewhere or I am missing something?
Really sorry if the query was too silly.
Regards,
Kalyan

My Response:

Not at all.
Pls ignore that case. It has been taken out but apparently I missed it while updating the syllabus.
Sorry about the confusion.
Sudhir

Update: received morning today (28-oct)

Sir,

Is there a website or report through which I can get statewise/UT wise SEC classification of the population? To be more specific, I was seeking out the population in Chandigarh & Gurgaon/NCR which would fall under A1 & A2 SEC segment. Any help would be much appreciated

My response:

Hi Harsimran,

No website directly that I know of comes to mind.

However, overall population stats are easily available and I believe the urban SEC-wise population distribution can be overlaid on the total populn of these regions to get some systematic idea of what the sizes are.

In any case, I'd first try contacting LRC staff with this query. Their extensive databases might have something I'm not aware of.
Hope that helps.

_______________________________________________________________________________
I received this email:

Prof. Sudhir,

This is with regards to the question on whether a cross-sectional or longitudinal design is a better fit when response bias is the evaluation criteria. As opposed to what was mentioned in class, I feel longitudinal designs are better for response bias because we can carry out validation or observe any biases by recording a particular respondent’s responses to the same or similar questions at different points in time.
Thanks,
Rangasayee Chandrasekaran

And this was my response:

Thanks, Rangasayee.


Shall put this up on the blog under 'Student contributions'.

I'd say the answer 'depends' on the research questions. Some issues may be quite sensitive to whether the panel was previously quizzed on it and may reflect bias. However, in most cases, I do agree, your point makes sense.

Sudhir

Received this email from 'R' :

Dear Chandana/Pankaj

While taking the survey, I unknowingly skipped over 1 page of questions and the survey did not give me any reminder to complete the questions on that page.

Hope there will be no penalty in marks for skipping(PS:- I am not able to take the survey again)
Pls confirm and provide alternate solution to the issue.

Regards
R

And this was my response:

Don't worry about it, R.


What worries me is that the survey allows folks to skip over entire pages without any problem. perhaps it makes sense to have 1 (easy) question forced (as in, 'answer required for survey to proceed forward') to confront this problem.

Sudhir

More mailbag emails will be putup in the order in which I receive them: