Qualifier_Q68

[Anand Iyer]

Hi,

Nice to be back here after loooong time!

The correlation was found to be r = 0.86 between price (x) and demand of mobile phones (y). Which of the following options could be true?

The given correct answers are:

1. Given two points from the scatter plot of price and demand of mobile phones, one point has a smaller x value and a larger y value than another point.

2. Given two points from the scatter plot of price and demand of mobile phones, one point has a larger x value and a smaller y value than another point.

3. The covariance of price and demand of mobile phones is positive.

I can understand the third, but how is the first and second right options?

If (x1, y1) and (x2, y2) are two points in the line, my answers are 

x2 > x1 and y2 > y1 

x2 < x1 and y2 < y1

Shouldn't this hold true so we get an upward trend?

Can anyone answer this?

[Akshay Malik]

Same doubt 

[CG]

Same doubt here as well

[Amrit Raj]

The first two options are talking about the scatter plot and not the exact line of fit. So the actual points can be above and below the line.

[Anand Iyer]

Well, Amrit, even if that was the reason, it doesn't make them the right answers...they're mere 'possibilities'.  

I'm sure, we're missing something...

--

Cheers,

Anand

[_Music_ Viking]

Also, in the same paper, I don't get how to solve Q.67 and Q.82. I'm not getting those. Can you guys help me with that?

Thank you

[Malabika Guha Mustafi]

If the question is from set 2 then the correlation is negative. I have tried to paste the screen shot but somehow it is not working properly. If it is negative correlation then the options re perfect.

[DHARMA TEJA GODUMALA]

For r= -0.86 ,

Hope this helps you

[Anand Iyer]

Oh. What!

This is from qp1. So there’s some goof up for sure

On Thu, 3 Dec 2020 at 8:56 PM, Malabika Guha Mustafi <malabi...@gmail.com> wrote:

@ Anand

The correlation was found to be r = 0.86 between price (x) and demand of mobile phones (y). Which of the following options could be true?

The given correct answers are:

1. Given two points from the scatter plot of price and demand of mobile phones, one point has a smaller x value and a larger y value than another point.

2. Given two points from the scatter plot of price and demand of mobile phones, one point has a larger x value and a smaller y value than another point.

3. The covariance of price and demand of mobile phones is positive.

I can understand the third, but how is the first and second right options?

If (x1, y1) and (x2, y2) are two points in the line, my answers are 

x2 > x1 and y2 > y1 

x2 < x1 and y2 < y1

Shouldn't this hold true so we get an upward trend?

Correlation  value is positive and strong .

@ Anand, which question set is it ?

for question set 2

On Thursday, 3 December 2020 at 20:04:13 UTC+5:30 _Music_ Viking wrote:

--

Cheers,

Anand

--
You received this message because you are subscribed to a topic in the Google Groups "Discussion forum for Statistics for Data Science I" group.
To unsubscribe from this topic, visit https://groups.google.com/a/nptel.iitm.ac.in/d/topic/ma1002-discuss/TGUeuot4alY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ma1002-discus...@nptel.iitm.ac.in.
To view this discussion on the web visit https://groups.google.com/a/nptel.iitm.ac.in/d/msgid/ma1002-discuss/9675a94a-ae7d-4ab2-956a-b75aa32d63bbn%40nptel.iitm.ac.in.

--

Cheers,

Anand

[Malabika Guha Mustafi]

@ Dharma, I also assumed that. But Anand  has said that the third option is right option.

3. The covariance of price and demand of mobile phones is positive.

If the correlation is negative then how is this option right? 

[Anand Iyer]

Hold on. I’ll send the screenshot 

--

Cheers,

Anand

[CG]

This is from Q1QP1 (Image attached below)

The correlation coefficient is clearly positive. So I wonder how 1 and 2 can be correct. 

I spent quite a while wondering why putting a tick on 1 answer gives 3 marks - literally wasted 6 - 8 minutes reading and re-reading the question to check that only option 3 is the correct answer.

[Anand Iyer]

screenshot attached...

--

Cheers,

Anand

[Akshay Malik]

[Anand Iyer]

oh, everyone sprang to action...news!

i think this needs to be brought to the notice of our course support team.  We've conflicting answers in the answer keys itself then...

--

Cheers,

Anand

[Malabika Guha Mustafi]

Definitely. For strong positive correlation , the two options  can not be right option.

[Anand Iyer]

hmmm, 

finally we all managed to dig something about the QP1 too.  So far, everywhere I'd been reading about QP2 being tough, ambiguous...

Now, thanks to our collective effort, we've produced proof that QP1 also has its share of wrong stuff...haha.

I feel there's at least one more that's wrong with QP1...let me dig that up tomorrow...

--

Cheers,

Anand

[Malabika Guha Mustafi]

Proper digging was initiated by QP2 warrior ;)

[Anand Iyer]

aha, even competition there, eh?

--

Cheers,

Anand

[Malabika Guha Mustafi]

Haaaaa, No competition. I really need you, CG and akshyam's contributions in coming session.

Team please clarify the doubts.

[Anand Iyer]

Just having a little fun, of course...

Thanks, Malabika for following up!

One of the major reasons I'm loving this course are these forums here...and our interactions...Hope it continues the same way, until the end...and beyond...

Coming back to this, I hope the support team looks at it soon enough...and I'll be able to claim my 2 marks out of this:)

--

Cheers,

Anand

[Akshay Malik]

I think the answer lies in the words "could be true" which conveys possibilities.

Remember the question from one of the assignments where box plot was given and we had to find several possibilities. There also the words were "could be".

[Malabika Guha Mustafi]

@ akshyam ,

 I don't think so . It gives totally different direction . We can not consider two points as outlier which can give negative correlations .

[Anand Iyer]

hmm. that's kind of an anti-climax Akshay if that was true:(

--

Cheers,

Anand

[Akshay Malik]

[Akshay Malik]

We do not even need outliers.

[Akshay Malik]

The first options will be true irrespective of whether the covariance is negative or positive. You will certainly find atleast two points which satisfies that first two options.

[Malabika Guha Mustafi]

Why?  How do you get such conclusion? 

[Anand Iyer]

If this is indeed like what Akshay says, this deserves the award for the ultimate twisted question of this course...at least, it's a very strong contender:) 

who could've guessed it, after all?

--

Cheers,

Anand

[Akshay Malik]

@Malabika

Correction : Not always , but it is a possibility that you may find certain points. Because it is asked "which of the options could/may be true?"

[Malabika Guha Mustafi]

No. We can not say at least two points should be like that. Is there a rule ? 

The first options will be true irrespective of whether the covariance is negative or positive. You will certainly find atleast two points which satisfies that first two options.

Then second option is also true .

We can not assume such things as a  right option.

[Malabika Guha Mustafi]

Scary !! This type of logic we were supposed  to think in exam hall! God save us!

On Thursday, 3 December 2020 at 22:18:12 UTC+5:30 akshaym...@gmail.com wrote:

[Malabika Guha Mustafi]

For negative correlation, there is absolute no problem.

[Akshay Malik]

[Akshay Malik]

Yes. And Me and Anand missed an easy tricky question :p (if my reasoning is correct , which may not be)

[Amrit Raj]

I missed the thread. The answer is not wrong.
It says "could be"(so possibilities) and the point need not be on the line of fit. It is on the scatter plot. Strong positive correlation does not mean, all points should lie on the line of fit. Infact it is not necessary for a single point also to lie on the line of fit. Also since we do not know the intercept, all possibilities are there,considering the slope trending upward.( for 0.86 correlation)

[Malabika Guha Mustafi]

If we take such type of possibilities then the student who raised the questions in maths that both are unique( in cup problem) that's also should be considered as true.

Yes , strong correlation never implies that a single point should be lie on the line of fit.. But , it does not imply that two points should be there which gives different directions.

Anyhow, this type of problem is totally unnecessary particular in exam.

[Akshay Malik]

On Friday, December 4, 2020 at 12:18:21 PM UTC+5:30 Akshay Malik wrote:

[Malabika Guha Mustafi]

If we take possibilities then the argument in maths is also true considering possibilities.

If that was wrong considering a special case where m=n=3 then these two options are also not true.

So, what do you expect to apply our logic differently in maths and statistics?