[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

From |
"Nick Winter" <nwinter@policystudies.com> |

To |
<statalist@hsphsun2.harvard.edu> |

Subject |
st: RE: how to generate groups based on some characteristics and obtain the mean/median value for each group |

Date |
Wed, 19 Jun 2002 09:38:47 -0400 |

> -----Original Message----- > From: Yi, Bingsheng [mailto:byi@coba.usf.edu] > Sent: Tuesday, June 18, 2002 7:11 PM > To: statalist@hsphsun2.harvard.edu > Subject: st: how to generate groups based on some > characteristics and obtain the mean/median value for each group > > > Dear Statalisters, > > I wonder whether you will help me figure out the codes to solve the > following problem: > > I have 12 years panel data containing these four variables: > Tobin's q, > size, 4-digit industry code (ind4), and id. For each year, I > want to make > some adjusments in one variable (Tobin's q) based on the > other two variables > (industry and size). First I need to ensure that there are > lat least 10 > firms within each industry. If the number of firms within a > 4-digit industry > code is less than 10, I use 3-digit industry code generated > by gen str4 > ind3=substr(ind4,1,3), see whether the number of firms with > the same 3-digit > industry code is greater or equal to 10, if not, then generate and use > 2-digit industry code. So in the end there are at least 10 > firms within an > industry ( which are classified by 4-digit, 3-digit, 2-digit, > or 1-digit > industry code). The problem is how to get and record the > number of firms in > each industry. For this piece, try something like this. First, generate four variables indicating the 4, 3, 2 and 1-digit industry codes for **ALL** records, named ind1, ind2, ind3, ind4. Then: * generate the number of records in each group forval i=1/4 { sort ind`i' by ind`i': gen num`i'=_N } * group the records gen finalgrp = ind1 forval i=2/4 { replace finalgrp = ind`i' if num`i'>=10 } This should create the variable "finalgrp", which will contain the grouping you desire. THen you can calculate whatever statistics you want in those groups. Nick Winter * * For searches and help try: * http://www.stata.com/support/faqs/res/findit.html * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

- Prev by Date:
**st: RE: graph: cross-plot of means** - Next by Date:
**RE: st: Running do files one line at a time** - Previous by thread:
**st: RE: graph: cross-plot of means** - Next by thread:
**st: Tempname and Tempvar - help desperately needed** - Index(es):

© Copyright 1996–2016 StataCorp LP | Terms of use | Privacy | Contact us | What's new | Site index |