Magic Bullet Poverty Measure?

>how life has gotten worse for folks who have returned to A, and for folks who are displaced in B & C. So you already have your conclusion then, and are looking for ways to quantify it?

Thanks mrbugsentry: there's a simple answer to your question, as well as a complex answer. The simple answer is "yes." The complex answer is, well, a bit more complex. Reputable, albeit by definition, anecdotal, reportage unambiguously presents living conditions for those who fled place A as being "worse for folks who have returned to A, and for folks displaced in B & C." Of course, this reportage is largely focused to those who were already poor to begin with (and as such, self-selecting data, as well as pre-concluded). What my guy wants to do is quantify what has been qualitatively (for the most part) reported and analyzed. (Isn't that the central problem for contemporary social science?) So, while we have a conclusion, and want to quantify it, we're not necessarily married to the particular dress the conclusion is wearing, if you get my drift. We're open (don't we have to be?) to results that might show improvement in the lives of folks who've been displaced from A, but still, we (I mean, "he") still wants "that one indicator which really demonstrates the poverty that" folks who've been displaced from A live in. I hope this makes sense. I realize it's probably more answer than you wanted, but I need all the help I can get with this one. Either there is such a "magic bullet indicator," and it's my job to find it, or there is no such indicator, and I have to explain why.

deejay, can it be a set of indicators or does it have to be a single one? From your first question it sounds like you wanted a set that would measure the changes (qualitatively established as "worse") in living conditions in three places in the first world after a disaster, including the site of the disaster. From your clarification, it sounds like you want one indicator that will demonstrate the quality of "worse" in the same way for each of the three places. This seems problematic to me because the quality of "worse" may be different in the three places. For instance, it would be worth looking at access to utilities in place A, but that might not be relevant in places B and C. Things can get "worse" in different ways.

Carmen: That's a great question, and it may get to the difficulty here. I want a set, precisely for the reasons you state. My guy, however, keeps asking for a single indicator. I'm prepared to tell him that there is no single indicator, but I wanted to make sure, first. More to the point, (and maybe this should have been my question) are some indicators more statistically reliable than others? (uh-oh . . .)

i'm a bit confused: it appears that your boss wants data on a geographical place for a, but individuals for b and c. you can't use geographically based data for b and c, because it will also include data about the residents that didn't come ovver form a. i'm also not sure on how population size, (maybe) education and owners vs renters are good indicators of poverty levels. the real figure seems to be per capita income. i guess the best way would be to draw up a form for all residents of a, b and c, measuring per capitia income, education, and any other measurable factors you wish to use. then make sure that the data is collected only on those actually involved in the original a. and as a side note, this still might be deceiving. if, for example, a is new orleans (?) then a major problem with per capita is that a; the population is now lower, and b: i would guess that there has been a significant increase in wages assiciated with rebuilding, both which would artificially increase per capita income. one way to combat this is to design a subjective survey to have respondents indicate how they think they are doing...but i'm guessing that this isn't a feasable idea for you.

oh, and sorry about my really bad spelling, but i'm also wrestling with a dog at the moment.

LSP: Thanks tons. Are there any unusual or often-overlooked data-measures that strike you as relevant to this problem? And you're right about the "per capita" measure. For a lot of reasons, we're not going to go there.

I am wholly unconcerned with spelling issues on the green, LSP. No worries. Thanks for the great answer.

Hmmm.... well, beyond the statistical reliability of measures (which are absolutely variable) there is the issue of making sure that your indicators actually indicate what you want them to. That is, education levels aren't the same as poverty levels, but in some cases can indicate access to stable, well paid employment, which affect poverty. In these cases education levels taken in the aggregate can indicate something about a population's vulnerability to low wages and poor working conditions. Okay, so you probably know all that, but I just want to lay it out in detail. So, the problem that I see here is the natural disaster. Especially if you are talking about something as large as what happened in New Orleans, the problem that I see you facing is that you don't know what your indicators indicate anymore. What does it mean to be a high school drop out in post-Katrina New Orleans? The way that employment works may have changed. Likewise health indicators may not be spread out among different economic groups in the same way anymore, etc. etc. So this brings me to my next questions: is this an academic or consulting activity? What is the timeline? What is the budget?