HW#2 Answers
0 views
Skip to first unread message
Kartik Ramachandran
Apr 7, 2011, 12:22:32 PM4/7/11
to collins-omis3...@googlegroups.com
Hi Jonathan,
Regarding HW2, how did you get the tcrit to be 1.73.
From the annova, we have k-1 = 5 ==>k = 6 ==> N-k = 14.
So we need to use tinv(0.05,14) which is 2.14
Please let me know what you think..
Kartik Ramachandran
Apr 7, 2011, 12:22:57 PM4/7/11
to collins-omis3...@googlegroups.com
I am referring to Qn2 in HW2.
Thanks,
Kartik
Jonathan Carandang
Apr 8, 2011, 1:53:22 PM4/8/11
to Collins OMIS357 Spring 2011
Hi Kartik,
For Problem 1
I have 20 data sets, which gives you N=20
I have 1 independent variable (Parents' Ave Height), which gives you 2
degrees of freedom including the intercept --> K=2
N-K = 18
K-1 = 1
with 95% confidence level (alpha = 0.05), I got tcrit of 1.73
For Problem 2, (I revised mine posted one due to typo error. Refer
below for the update values.)
I have 20 data sets, which gives you N=20
I have 3 independent variables (Price, # of SalesPersons, and News
Advert), which gives you 4 degrees of freedom including the intercept
--> K=4
N-K = 16
K-1 = 3
with 95% confidence level (alpha = 0.05), I got
Tcrit of 1.75 and
Fcrit of 3.24
Note that I dropped the competitors price and Radio Advert, as they
did not pass the t-crit test.
Thanks for questioning.
Regard
Jonathan
On Apr 7, 9:22 am, Kartik Ramachandran <kramach1...@gmail.com> wrote:
> Hi Jonathan,
> Regarding HW2, how did you get the tcrit to be 1.73.
>
> From the annova, we have k-1 = 5 ==>k = 6 ==> N-k = 14.
> So we need to use tinv(0.05,14) which is 2.14
>
> Please let me know what you think..
>
> On Wed, Apr 6, 2011 at 4:09 PM, Jonathan Carandang
> <joncarand...@yahoo.com>wrote:
>
> - Show quoted text -
For Problem 1
I have 20 data sets, which gives you N=20
I have 1 independent variable (Parents' Ave Height), which gives you 2
degrees of freedom including the intercept --> K=2
N-K = 18
K-1 = 1
with 95% confidence level (alpha = 0.05), I got tcrit of 1.73
For Problem 2, (I revised mine posted one due to typo error. Refer
below for the update values.)
I have 20 data sets, which gives you N=20
I have 3 independent variables (Price, # of SalesPersons, and News
Advert), which gives you 4 degrees of freedom including the intercept
--> K=4
N-K = 16
K-1 = 3
with 95% confidence level (alpha = 0.05), I got
Tcrit of 1.75 and
Fcrit of 3.24
Note that I dropped the competitors price and Radio Advert, as they
did not pass the t-crit test.
Thanks for questioning.
Regard
Jonathan
On Apr 7, 9:22 am, Kartik Ramachandran <kramach1...@gmail.com> wrote:
> Hi Jonathan,
> Regarding HW2, how did you get the tcrit to be 1.73.
>
> From the annova, we have k-1 = 5 ==>k = 6 ==> N-k = 14.
> So we need to use tinv(0.05,14) which is 2.14
>
> Please let me know what you think..
>
> On Wed, Apr 6, 2011 at 4:09 PM, Jonathan Carandang
>
>
>
> > Here are my answers.
>
> > Let me know if any of yours comes out different.
>
> > Regards
> > Jonathan- Hide quoted text -
>
>
> > Here are my answers.
>
> > Let me know if any of yours comes out different.
>
> > Regards
>
> - Show quoted text -
Savitha
Apr 10, 2011, 10:36:05 AM4/10/11
to Collins OMIS357 Spring 2011
Hi Jonathan,
For problem2 when we picked the model with price, news ads,
#salespersons the beta's were
Intercept 1715.049479
price -49.04641917
#salespersons 156.6809521 <<----
news ad($) 0.215196687
In our discussions yesterday we thought that the beta for
#salespersons 156 is high i.e we need 156 sales people to increase the
sale by 1 unit.
So the question is if we should eliminate this variable also and have
a 2 variable (price & news ads) model.
In my calculation the R^2 for the 2 variable mode is 0.72 (our
original value was 0.77).
Savitha
For problem2 when we picked the model with price, news ads,
#salespersons the beta's were
Intercept 1715.049479
price -49.04641917
#salespersons 156.6809521 <<----
news ad($) 0.215196687
In our discussions yesterday we thought that the beta for
#salespersons 156 is high i.e we need 156 sales people to increase the
sale by 1 unit.
So the question is if we should eliminate this variable also and have
a 2 variable (price & news ads) model.
In my calculation the R^2 for the 2 variable mode is 0.72 (our
original value was 0.77).
Savitha
Kartik Ramachandran
Apr 10, 2011, 11:09:46 AM4/10/11
to collins-omis3...@googlegroups.com
Hi,
156 should be interpreted as follows in my opinion;
156 should be interpreted as follows in my opinion;
An increase in 1 sales person increases the sales by 156 units.
Thanks,
Kartik
Sent from my iPhone
alex smirnov
Apr 10, 2011, 1:34:31 PM4/10/11
to collins-omis3...@googlegroups.com, Kartik Ramachandran
I agree with 1 salesperson = 156 units
i don't think it makes sense the other way around
On Sun, Apr 10, 2011 at 8:09 AM, Kartik Ramachandran <krama...@gmail.com> wrote:
Hi,
156 should be interpreted as follows in my opinion;
An increase in 1 sales person increases the sales by 156 units.
Thanks,
Kartik
Sent from my iPhone
Savitha
Apr 10, 2011, 9:49:25 PM4/10/11
to Collins OMIS357 Spring 2011
Thanks Karthik and Alex.
You are right don't know why I thought the other way.
If we just plug these numbers back into the eqn it is easy to see it.
I am glad I asked the question.
Savitha
On Apr 10, 10:34 am, alex smirnov <alex.smir...@gmail.com> wrote:
> I agree with 1 salesperson = 156 units
>
> i don't think it makes sense the other way around
>
> On Sun, Apr 10, 2011 at 8:09 AM, Kartik Ramachandran
> <kramach1...@gmail.com>wrote:
You are right don't know why I thought the other way.
If we just plug these numbers back into the eqn it is easy to see it.
I am glad I asked the question.
Savitha
On Apr 10, 10:34 am, alex smirnov <alex.smir...@gmail.com> wrote:
> I agree with 1 salesperson = 156 units
>
> i don't think it makes sense the other way around
>
> On Sun, Apr 10, 2011 at 8:09 AM, Kartik Ramachandran
>
> > Hi,
> > 156 should be interpreted as follows in my opinion;
>
> > An increase in 1 sales person increases the sales by 156 units.
>
> > Thanks,
> > Kartik
>
> > Sent from my iPhone
>
> > Hi,
> > 156 should be interpreted as follows in my opinion;
>
> > An increase in 1 sales person increases the sales by 156 units.
>
> > Thanks,
> > Kartik
>
> > Sent from my iPhone
>
Savitha
Apr 10, 2011, 10:06:15 PM4/10/11
to Collins OMIS357 Spring 2011
Although the sales due to additional news-ads seem to be just 0.2151
(kind of low).
I tried to run regression on just price and # of sales. The sales
dropped even further (-92) for increase in price, while the sales due
to additional sales person increase to 235 (75 more sales).
But the R^2 dropped to 0.61 (with 3 variables it was 0.77)
I guess it is safe to conclude to use the 3 variable (price, #of sales
persons, news) as it has higer confidence level (0.77).
Savitha
(kind of low).
I tried to run regression on just price and # of sales. The sales
dropped even further (-92) for increase in price, while the sales due
to additional sales person increase to 235 (75 more sales).
But the R^2 dropped to 0.61 (with 3 variables it was 0.77)
I guess it is safe to conclude to use the 3 variable (price, #of sales
persons, news) as it has higer confidence level (0.77).
Savitha
Crystal Diaz
Apr 10, 2011, 10:34:20 PM4/10/11
to collins-omis3...@googlegroups.com
Savitha,
I am working on this problem right now, and I have found the same to be true. I have choosen to work with the same 3 variables because they are the strongest as indicated by the t-stat and when we drop the other two variables the r2 remains at 0.77.
-Crystal
PS. If anyone is at school I am in LUCAS 200B and will be here for another hour or so.
I am working on this problem right now, and I have found the same to be true. I have choosen to work with the same 3 variables because they are the strongest as indicated by the t-stat and when we drop the other two variables the r2 remains at 0.77.
-Crystal
PS. If anyone is at school I am in LUCAS 200B and will be here for another hour or so.
Reply all
Reply to author
Forward
0 new messages