How to Calcualte the Federal Poverty Line in a dataset of youth based on parent measures?

[Nathan E. Fosse]

Presume I have a dataframe of children with parental characteristics (only one child per family is analyzed). Presume we have standard bacground measures such as parental income, education, occupation, family status, and family size.

How can I calculate a variable to determine if that family meets the Federal Poverty Line? I see many ados for Worldbank measures of poverty, but those are incommensurate with American poverty measures.

Thanks much!
- Nate

[ben earnhart]

have you checked out -egen inequal- ? Really, it's overkill/inappropriate for your purposes, but still potentially useful and cool.

You would need to (carefully) add up the incomes of the family members for total family income. When I say carefully, I mean watch out for single or zero-parent households, missing data, different sources of income, categorical variables that need to be coded to a midpoint before adding, etc.

Count the number of adults, elderly, and children in the household.

So now you have three variables: total income, #kids, #adults<age 65, # adults=>age 65.

Then see http://www.census.gov/hhes/www/pover...hld/index.html

Then it's just a bunch of replace-if statements based on the values in the table, the # adults, # elderly, and the # children, No way around it being long, but not *too* bad if you only have a single year.

[Nathan E. Fosse]

Thanks, Ben, I appreciate this. I think I found something that works, and it has to do with distinguishing between poverty thresholds and poverty guidelines. Thresholds are more complex, and worthy of writing an appropriate ado, actually! Thanks, Ben, for the tips and directions.

I've found a great ado ("ssc install povguide") that only requires family size and year. From that scalar you can categorize families in the data. As David Kantor explains below, Poverty Threshold and Poverty Guidelines are two, somewhat important, distinct forms of classification:

The Poverty Guideline is one of two standards in use; the other is the Poverty Threshold, a more complex
calculation, involving the number of children and elderly. The Poverty Threshold is the original standard,
based on the work of Mollie Orshansky; the Poverty Guideline is regarded as a simplified alternative. Also,
the Threshold is used for statistical purposes (e.g., by the U.S. Census Bureau), whereas the Guideline is
used for program eligibility.

[ben earnhart]

That's interesting. When counting up who *should* be eligible (census) they use a micrometer, but when deciding who actually *gets* $ they use a ruler. I suppose based on the burden of documentation and the workload of social workers, makes sense to use a simplified measure. But it makes me wonder how accurate the census figures are. If the burden is too heavy for social workers, imagine the poor census workers out there trying to gather accurate data. Or perhaps it's more an historical artifact than something with a real reason. Huh.

[Nathan E. Fosse]

Hi, Ben -
That's quite an interesting point, presuming the Census is for political reasons (seems reasonable enough). Regarding the Census figures and this distinction, I absolutely agree the guideline/threshold distinction could alter our understanding of "true" prevalence rates of poverty. I would no doubt consider it's far easier to calculate thresholds, and (I'm guessing here) there's been some work to corroborate these associations. I would presume it's something like body mass index; a heuristic that, despite its flaws, is practicable and driven by historical precedence.

Now if I had the time I would definitely consider your great idea of using the egen inequal command and writing the appropriate program that accommodates the 40+ sub-categories for different family configurations. That could be a lot of if else statements!

- Nate

[KerstinMccabe]

Some reports says that there are Americans who already win in the battle of poverty. However, according to United States Census Bureau, there are still lived under poverty line even in the mid 1959’s. Article resource: Americans below poverty line is highest number since records began

[Nathan E. Fosse]

Thanks Kerstin, for the link. I was looking for how the Federal Government defines this, as it is an institutionalized benchmark. While true that we have the highest number of Americans living in poverty, this is also because the country has tripled in size. Nevertheless, I no doubt suspect there are measurement issues at stake. Do you have any consideration of alternative measures based on your work?

Warm regards,
- Nate

[Michael Anbar]

have you checked out -egen inequal- ? Really, it's overkill/inappropriate for your purposes, but still potentially useful and cool.

What package are you referring to? I wasn't able to find anything with -findit inequal- or by searching for "_ginequal.ado" (which would be the name of the file that interfaces with -egen-).

[Nick Cox]

ssc has a feature that helps when exact names are in question. Naturally, at most this helps for stuff on SSC.

. ssc desc e
-------------------------------------------------------------------------------
http://fmwww.bc.edu/repec/bocode/e/
(no title)
-------------------------------------------------------------------------------

PACKAGES you could -net describe-:
    eaalogit          module to estimate endogenous attribute attendance models
    eba               module to perform extreme bound analysis
    ebalance          module to perform Entropy reweighting to create balanced
> samples
    ebrowse           module to to display variables in the Data Editor, ordere
> d by varlist
    eclpci            module to compute Exact and Approximate Confidence Limits
>  for Poisson Counts
    eclplot           module to generate estimates and confidence limits plots
    eealogit          module to estimate endogenous attribute attendance models
    efetch_tools      module to retrieve genetic data from NCBI
    effcon            module to calculate confidence limits for effect size est
> imated from one or two samples from normal distribution
    effects           module to provide a graphical interface for estimation co
> mmands
    egen_inequal      module providing extensions to generate inequality and po
> verty measures
    egenmore          modules to extend the generate function
    egfr              module to calculate estimated glomerular filtration rate
> (eGFR)
    egi               module to calculate Ellison and Glaeser's index of geogra
> phical concentration
    egranger          module to perform Engle-Granger cointegration tests and 2
> -step ECM estimation
    eitc              module to compute U.S. Earned Income Tax Credit receipts
    elapse            module to calculate elapsed time in procedure
    electool          module containing toolkit to analyze electoral data
    elife             module to model annual increments of life expectancy
    elixhauser        module to calculate Elixhauser index of comorbidity
    ellip             module to graph confidence ellipses (version 8)
    ellip6            module to plot confidence ellipses (version 6)
    ellip7            module to plot confidence ellipses (version 7)
    ellipticity       module to compute optimal rank-based multivariate tests o
> f elliptical symmetry and sphericity
    email             module to send emails from Stata via Python
    emailme           module to facilitate sending mail on Windows systems
    emh               module to compute extended Mantel-Haenszel statistics
    enlarge           module to propagate constant values through group
    eofplot           module to plot coefficients or loadings after principal c
> omponent or factor analysis
    eperiod           module to calculate elapsed time between dates
    epiconf           module to assess confounding effects in epidemiological s
> tudies
    epidata           module Data entry system for use with Stata, SPSS, SAS, E
> xcel, text formats
    episens           module for basic sensitivity analysis of epidemiological
> results
    episensrri        module for basic sensitivity analysis for unmeasured conf
> ounders
    epiweek           module to create epidemiological week and equivalent epid
> emiological year
    epsigr            module to save graph in EPSF format in UNIX environment
    eqprhistogram     module for equal probability histogram
    equation          module to Output The Equation of a Regression
    er                module to compute Esteban and Ray measure of polarization
    erate             module to import up-to-date exchange rates for any curren
> cy pairs
    erepost           module to repost the estimation results
    eret2             module to save results in e()
    esetran           module to transform estimates and standard errors in parm
> est resultssets
    esli              module to calculate the equation of a straight line
    est2tex           module to create LaTeX tables from estimation results
    estout            module to make regression tables
    estout1           module to export estimation results from estimates table
    estparm           module to save results from a parmest resultsset and test
>  equality
    estrat            module to perform Endogenous Stratification for Randomize
> d Experiments
    estsave           module to save estimation results with current dataset
    estwrite          module to store estimation results on disk
    etime             module to time how long a program runs
    eurostatuse       module to import data from Eurostat repository into Stata
    eventstudy        module to perform event studies in finance
    eventstudy2       module to perform event studies with complex test statist
> ics
    evhistplot        module to produce event history plot
    ewma              module to calculate exponentially weighted moving average
    ewreg             module to estimate errors-in-variable model with mismeasu
> red regressor
    examples          module to show examples from on-line help files
    excelcol          module to convert a column index into a name of an Excel
> column
    excelsave         module to provide a convenient way to export matched Stat
> a datasets from a specified input directory (also including, if wanted, addit
> ionally recursively matched subsets of files within d
    expandby          module to duplicate observations by variable
    expgen            module to duplicate observations sorted in original order
>  with generated variables
    expl              module to explore folders and files
    explist           module to generate an exponentially-spaced list of number
> s
    extfunnel         module to produce graphical augmentations to the funnel p
> lot to indicate the potential impact of a new study on an existing meta-analy
> sis
    extreme           module to fit models used in univariate extreme value the
> ory
    extremes          module to list extreme values of a variable
-------------------------------------------------------------------------------
(type -ssc describe pkgname- for more information on pkgname)

Ben forgot the underscore.