[Limdep Nlogit List] grouped data regression

Sat May 8 02:11:34 EST 2010

The 1 to 3 coding was just an example. I didn't want to overwhelm the list
with data (which apparently I did with a ton of emails, sorry). So yes, I
have 11 categories and I created 10 limit variables. In each limit variable
there are two possible values, one for each group. Is it normal that the
program reports the average of these values?

Andreas 

-----Original Message-----
From: limdep-bounces at limdep.itls.usyd.edu.au
[mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of William Greene
Sent: Friday, May 07, 2010 7:04 PM
To: Limdep and Nlogit Mailing List
Subject: Re: [Limdep Nlogit List] grouped data regression

It is giving you results based on the data you supplied. Apparently

your LHS variable is coded 1 to 11, not 1 to 3. I don't know what is

going on with the censoring limits; those are your data.

/Bill

----- Original Message -----

From: "Andreas Drichoutis" <adrihout at aua.gr>

To: "Limdep and Nlogit Mailing List" <limdep at limdep.itls.usyd.edu.au>

Sent: Friday, May 7, 2010 10:55:11 AM GMT -05:00 Colombia

Subject: Re: [Limdep Nlogit List] grouped data regression

Ok got it. It seems like it worked. Why do I get the below output? These are

not the limits I used. It seems to me it is taking an average of the limit

variables for the 2 groups. Is this ok?

| Censoring Thresholds for the 11 cells:      |

|  y   Lower   Upper y   Lower    Upper       |

|  1 *******    1.00  2    1.00   13.00       |

|  3   13.00   24.50  4   24.50   37.00       |

|  5   37.00   50.50  6   50.50   66.50       |

|  7   66.50   86.00  8   86.00  111.00       |

|  9  111.00  144.50 10  144.50  191.00       |

| 11  191.00 *******                          |

Andreas 

-----Original Message-----

From: limdep-bounces at limdep.itls.usyd.edu.au

[mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of William Greene

Sent: Friday, May 07, 2010 3:30 PM

To: Limdep and Nlogit Mailing List

Subject: Re: [Limdep Nlogit List] grouped data regression

The dependent variable is coded 1,2,3. The limit values give the interior

values. For group 1, you give limit values 1 and 15. For group 2, your

limit values are 4,18.  You have two variables low,high.  Low takes value

1 or 4. High takes values 15 or 18.

/B. Greene

----- Original Message -----

From: "Andreas Drichoutis" <adrihout at aua.gr>

To: "Limdep and Nlogit Mailing List" <limdep at limdep.itls.usyd.edu.au>

Sent: Friday, May 7, 2010 7:22:19 AM GMT -05:00 Colombia

Subject: Re: [Limdep Nlogit List] grouped data regression

I still don't understand how to structure this in limdep. Let's take an

example. Assume y* is reported for group 1 like this: y*<1, 1<=y*<15,

y*>=15. For group 2 y* is reported as: y*<4, 4<=y*<18, y*>=18. How do you

set the limit for the extreme J category? Is it the same as for the J-1

category or should I leave it blank? Is the below coding correct?

y

Group

Limit1

Limit2

1

1

1

2

1

15

3

1

15

1

2

4

2

2

18

3

2

18

2

2

18

3

1

15

2

1

15

I did the above coding and then run the syntax:

 Grouped ; lhs=y ; rhs=one,x1,x2 ; limits=limit1,limit2 $

But I get this error: 

Error    49: GROUPED Y > # limits + 1. Bad data or too few limit values.

What am I doing wrong?

Andreas 

-----Original Message-----

From: limdep-bounces at limdep.itls.usyd.edu.au

[mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of William Greene

Sent: Friday, May 07, 2010 1:23 PM

To: Limdep and Nlogit Mailing List

Subject: Re: [Limdep Nlogit List] grouped data regression

Andreas. This appears not to be reported in the manual. 

But, the boundary values in the GROUPED_DATA command may be

variables rather than fixed values (or a mixture of the two).

This will allow you to do the setup you are looking for. 

Note that there must be the same number of bins for both

groups.

/Bill Greene

----- Original Message -----

From: "Andreas Drichoutis" <adrihout at aua.gr>

To: "Limdep and Nlogit Mailing List" <limdep at limdep.itls.usyd.edu.au>

Sent: Thursday, May 6, 2010 2:15:26 PM GMT -05:00 Colombia

Subject: [Limdep Nlogit List] grouped data regression

Dear all,

I have an interval category dependent variable that differs for two

treatments e.g. the limits are 25, 35, 45, 55 for one treatment and 20, 30,

40, 50 for another treatment. How do I combine these in a single grouped

data regression model? The way I read it in E27-61, this is not possible. Is

there a workaround?

Yours,

Andreas Drichoutis

_______________________________________________

Limdep site list

Limdep at limdep.itls.usyd.edu.au

http://limdep.itls.usyd.edu.au

_______________________________________________

Limdep site list

Limdep at limdep.itls.usyd.edu.au

http://limdep.itls.usyd.edu.au

_______________________________________________

Limdep site list

Limdep at limdep.itls.usyd.edu.au

http://limdep.itls.usyd.edu.au

_______________________________________________

Limdep site list

Limdep at limdep.itls.usyd.edu.au

http://limdep.itls.usyd.edu.au

_______________________________________________

Limdep site list

Limdep at limdep.itls.usyd.edu.au

http://limdep.itls.usyd.edu.au

_______________________________________________

Limdep site list

Limdep at limdep.itls.usyd.edu.au

http://limdep.itls.usyd.edu.au