Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
Austin Nichols <[email protected]> |

To |
[email protected] |

Subject |
Re: st: RE: Getting Started with ineqdeco |

Date |
Fri, 13 Jul 2012 10:11:04 -0400 |

One can also fit a distribution via ML using e.g. http://fmwww.bc.edu/repec/bocode/d/dagfit.html using the five quantiles (10th, 30th, 50th, 70th, 90th %iles) available. On Fri, Jul 13, 2012 at 8:33 AM, <[email protected]> wrote: > The Pedant did of course mean to say that there 9 deciles (not 99); > there are 99 percentiles. > > More constructively, on the substantive question of "what to do given > the data to hand?": > > The median of the poorest fifth is the 10th percentile (p10) and the > median of the richest fifth is the 90th percentile (p90). The ratio > p90/p10 is commonly used, at least by many labo(u)r economists, to > summarise inequality. So you could calculate that measure for each > community, and plug that rather than Gini coefficients into whatever > subsequent analysis you are doing. > > > Stephen > ------------------ > Professor Stephen P. Jenkins <[email protected]> > Department of Social Policy and STICERD > London School of Economics and Political Science > Houghton Street, London WC2A 2AE, UK > Tel: +44(0)20 7955 6527 > Changing Fortunes: Income Mobility and Poverty Dynamics in Britain, OUP > 2011, http://ukcatalogue.oup.com/product/9780199226436.do > Survival Analysis Using Stata: > http://www.iser.essex.ac.uk/survival-analysis > Downloadable papers and software: http://ideas.repec.org/e/pje7.html > > > -----Original Message----- > From: Jenkins,S > Sent: 13 July 2012 09:53 > To: '[email protected]' > Subject: st: Getting Started with ineqdeco > > ------------------------------ > Date: Thu, 12 Jul 2012 17:06:36 +0000 (UTC) > From: [email protected] > Subject: st: Getting Started with ineqdeco > > Hi, > > I'm a newbie and statistics is not my long suit to say the least. I have > approximately 250 communities and I want to give each one a gini > coefficient that indicates the degree of income inequality prevailing > there. For each community I have median income for each quintile. It > appears that I have to use ineqdeco, which I have installed on my > computer. After looking at the material on line, I'm not even sure where > to start. I'm guessing that I have to create a single variable for each > community that includes this data, but I'm not sure. Can anybody provide > a "ineqdeco for dummies" description of exactly what I need to do? > > Marty > ======================== > > For each community, you have "grouped" data, not unit-record data > (observations on each and every unit within a community), and appear to > have the median for each quintile group*, not the mean. > > Grouped data should not be used with -ineqdeco- (or other inequality > measure programs on SSC) because they all assume that you have > unit-record data. Put differently, if you do apply them to grouped data, > as is, you will end up under-estimating inequality -- because you are in > effect assuming equality within each quintile group. > > There are methods for estimating inequality indices from grouped data. A > good reference is FA Cowell and F Mehta, (1982). "The estimation and > interpolation of inequality measures", Review of Economic Studies, > 49(2), 273-290, and references therein. See also FA Cowell's book, > Measuring Inequality (several editions). > > Note that these methods typically assume that you have the _mean_ income > within each group, not the _median_. > > In sum, I think your issues are not to do with -ineqdeco- or related > programs. Rather, the fundamental issues concern the (lack of) > information in the data at your disposal. You as researcher have to > decide how to address those. One crude way might be to assume that the > median within each quintile group is the same as the group mean, and > apply one of the grouped data methods. Whatever, with only 5 pieces of > information per community, summarising inequality accurately within each > community is a difficult task > > > Stephen > > * [Pedant's corner] The term "quintile" refers to an income value; > there are 4 quintiles, not 5 (and 99 deciles, not 100, etc.). Talk about > "quintile groups", not "quintiles", or -- more simply -- refer to the > "poorest fifth", ..., "richest fifth". * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

**References**:

- Prev by Date:
**st: RE: Propensity Score Matching with Multiple Categorical Variables with Multiple Categories...Dummy Variables?** - Next by Date:
**st: separate text boxes in each panel of a "by"** - Previous by thread:
**st: RE: Getting Started with ineqdeco** - Next by thread:
**st: Plotting several linear associations in one graph for comparison** - Index(es):