Migration from Dupage Pumas

I’m new here. How can I get data for those who moved out of dupage? I want to know where these people are going.

In IPUMS USA, there are several migration variables available that indicate an individual’s residence either 1 or 5 years ago. Geographic variables, such as counties and cities are not always identified due to confidentiality purposes, however, it looks like Du Page County, Illinois is identifiable in recent IPUMS USA samples. As such, you may want to look at using MIGCOUNTY1 (county of residence 1 year ago) to identify individuals who lived in Du Page County, Illinois. Alternatively, you could consider using MIGPUMA1, for which you would just need to identify the Migration Public Use Microdata Areas (MIGPUMA) that are of interest to you.

Historically, there is also information on residence 5 years ago, such as MIGCOUNTY5 and MIGCITY5, however, these variables are most recently available in the 2000 5% sample.

In addition to using a migration variable to help capture those who moved from Du Page, you will then want to have a geographic variable that captures each individual’s current residence. For this, again, you may want to look at COUNTYFIP or CITY, however, not all counties and cities will be identified. Similar to use of migration PUMAs, you could also use PUMA to find where each individual that previously lived in Du Page (either defined by its county or PUMAs) now lives.

Hi Michelle,

Thank you for your reply. How can I read the MIGPUMA1 codes? It looks like I should access a GIS program in order to be able to read them. Do you have the information in excel or so?

Thank you.


A little update on this. I decided to stop working with the PUMAs and work with STATEFIP, MIGCOUNTY, and PERWT. I-m creating a pivot table to get the results, but the total population that I get 97,079 is far from the current population for the county: 930,000. I cannot understand why. Shouldn’t the total population be close to the county’s current population? Would you please help me understand what am I missing?

I’ll address both of your questions just in case you decide to work with PUMAs in the future.

For more complex geographic variables, we have a page dedicated to geographic tools. You can find several pages related to PUMAs, however, I think this page would be of particular interest to you as it addresses 2010 migration PUMAs and includes a map that could help you identify the MIGPUMAs of interest to your analysis.

As for your question relating to MIGCOUNTY1, you should expect a number that is quite different from the current population of the county as the universe for MIGCOUNTY1 is persons age 1+ who lived in different house 1 year ago. So the population count you have of 97,079 is the number of people who lived in different house 1 year ago whose county of residence (1 year ago) was DuPage County.

I hope this helps.

Hello, I am interested in researching something similar to OP but for a whole state, not just a county.

I understand that the MIGRATE1 variable will tell me about the respondent’s migratory status based on whatever the year before meant for them (I am using the ACS 2017 5-year sample, so their answer to this question could refer to, as I understand it, a year as early as 2011).

I’m getting lost in how to determine where the respondent moved to, and, more importantly, how to focus my extract on just people moving from a specific state.

Before, I was restricting my extract to the FIPS state code I was interested in, but it now seems that this will just capture the 2017 5-year ACS residents of that state, and not all of the ones who may have lived there (or in a PUMA in that state) the “year” before.

Could you offer any pointers for how I can go about answering my question of where people from a specific state moved to? Ideally, my extract would include more demographic information about the movers (race, income, occupation). And if I should post this as a new question, let me know.


It sounds like the variable you ought to be using is the MIGPLAC1 variable. This variable identifies, for residents who lived in a different residence 1 year before the survey data, the US states where the respondent lived 1 year ago. You can either use the Select Cases tool to limit your data extract to only specific observations based on MIGPLAC1 values or perform these manipulations once you have downloaded your data extract.

1 Like

That sounds really promising! Thank you.

One follow up question: I’m using the MIGPLAC1 variable for one state, as suggested, and out of curiosity, I got a breakdown of MIGRATE1 outcomes using the sda.usa.ipums.org.

I’m getting that 7% of people stayed in the same residence a year before. How can this be if filtering by MIGPLAC1 I thought would narrow it down to people who reported having moved?

I agree that this observation sounds strange. However, I am not able to replicate this observation. For all of the samples I’ve looked into, conditioning on MIGPLAC1 > 0 leads to only values of MIGRATE1 > 1. Could you let me know what samples you are using?

I’m not sure, to be honest. Looks like all of the United States 1850-2017? I certainly would only want to look at the last 5-year ACS, though. Here’s what I see:

Okay, it looks like this is entirely due to the 1950 sample. You can see this if you include “year” in the “column” field using the 1850-2017 SDA file. This observation is driven by the fact that in 1950 the universe for MIGPLAC1 does not only include those who lived in a different house 1 year ago. It includes all “sample-line” persons who are over the age of one.

If you only want to use a specific ACS 5-year file on SDA, scroll down to the “use data from a single sample” table on the SDA page. In this table you will find links to the default samples for each year and all of the ACS 5-year files.

Ah-ha! Thank you! Also, I am trying to use the migpuma1 5-code variable also, using this equivalence file.

As I understand it, I should also provide a migplac1 code to differentiate from MIGPUMAs in other states?

Yes, you are correct. MIGPUMA1 codes are state-dependent and must be read in combination with a state ID variable.