Well, we all want to engage Millennials more and better, and Extension is no different. So let’s get a sense of their geographic distribution in the more urban parts of the state.

Nugget2The Nugget:

In most of Oregon’s metropolitan areas, you’ve got slightly higher odds of bumping into a Millennial in the central cities than in the suburbs, small towns, or rural areas. Only in Salem and Bend are you more likely to find a Millennial in the small towns of those metro areas than you are in the central cities, suburbs, or rural areas.

Evidence for the Nugget:

According to data for the 2008-2012 period from the American Community Survey, Millennials were concentrated at these percentages in the central cities, suburbs, small towns, and rural areas of Oregon’s major metropolitan areas:

Albany-Corvallis Millenials

Eugene-Salem Millennials

The data show that across all these metro areas, Millennials represented 1 in 5 to 2 in 5 members of the population in the central cities. Though this means they were still in the minority in these central cities, that representation is higher than the roughly 1 out of 7 people Millennials represented in the rural parts of these metro areas.

Of course, please note that in some of these metro areas there wasn’t much difference between the representation of Millennials in central cities and their representation in suburbs or small towns. In those areas, take that into consideration when you think about targeting and consider using a multi-geography approach.

Click here:  UrbanSuburbanRuralProfile_Methods_Etuk_20140905 for more information about how these geographies were defined and the estimates calculated.

Take-awayTake Aways:

Recruitment or outreach efforts directed at Millennials have to meet that population where it is, from a relevance perspective and from a geographic perspective. These data suggest directing those efforts, geographically, at the central cities of the state’s metropolitan areas. In these areas, any effort you make is more likely to reach a Millennial than an effort you put forth in a rural part of these metro areas.

When you meet a Millennial in one of these metro areas, ask them if they think they’ll move later on in life and where they’re hoping to move. Is it going to be the suburbs or rural areas eventually? This will help you get a sense of the future location of this age group. The fact that they’re now concentrated a bit more in central cities may just be a function of this part of their life-course — as they age they may gravitate towards the suburbs and rural areas like their parents (the Baby Boomers).

If you are trying to recruit or engage Millennials in the rural parts of these metro areas, try using a “hidden population” strategy like snowball recruitment (referral-based), facility-based recruitment (places where they tend to be), or time-location based recruitment (places where they tend to be at times they’re likely to be there).

We don’t always know what to do with data when we see it. And sometimes there’s nothing we can do with it, except tuck it away for a rainy day.

What if we could strive to use data to inform decisions we make about programs, policies, and day to day actions? What would that world look like, and how would we use the data? Would we use other types of information to inform our decisions?

Here’s a model I found from the organization, StriveTogether . I’ve tweaked it a bit to make sure we’re aware that the decisions we can make using data are affected by outside forces; and I tweaked it to draw attention to the juiciest opportunity for a data-driven decision…



  1. What do you think about this model?
  2. What would it take for you to have the convergence opportunity available as often as you need it?
  3. What types of decisions do you face that you wish you had data, research, and community voice about?
  4. When have you used data, research, and community voice to inform a decision?


ruralUrbanOR Latino graph


We know that Latinos have a long history in our state, and Oregon is becoming increasingly diverse as a result of marked growth among this population. So now we need to pay attention to how Latinos are represented in various aspects of community life and democracy – so we can make sure all Oregon residents are fairly included.

Nugget2Learning Nugget:

Available data suggest that Latinos are significantly under-represented in Oregon leadership roles. Demographic causes of this under-representation include measurement limitations, the age structure of the Latino community, and the citizenship status of many residents. This under-representation means that decisions affecting this population are being made by people who have limited personal experience of Latino issues – a fact that conflicts with most of our values for a representative democracy.

Evidence to support the nugget:

  1. Latinos are under-represented among voters

According to sample data from the 2012 Current Population Survey, November Supplement, administered by the US Census Bureau, only about 3% of all 2012 Oregon voters were Latino. Compared to the 11.7% representation of Latinos in Oregon, Latino voters were under-represented among Oregon voters by almost nine percentage points.

  1. Latinos are under-represented among business owners:

According to data from the 2007 Survey of Business Owners, administered by the US Census Bureau, 3% of all Oregon businesses were owned by Latinos. Latinos were thus also under-represented among Oregon business owners by about nine percentage points.

  1. Latinos are under-represented in elected office:

According to data I compiled from, the Oregon Legislature, Association of Oregon Counties, Oregon School Board Association, and the League of Oregon Cities, and compared to data collected by the National Association of Latino Elected and Appointed Officials (NALEO) Educational Fund in Washington, DC, Latinos made up less than 1% (about .3%) of all elected office holders in Oregon in 2008 (serving at federal, state, county, city, judicial/law, and education board levels). Latinos were under-represented among Oregon elected officials by 11 percentage points.

Demographic reasons for under-representation

  • Age & citizen composition of Oregon Latinos:

Only a third of Oregon Latinos are citizens and over the age of 18 (see chart below). Thus, only a third of Oregon Latinos are eligible to vote, run for office, or likely to own businesses. Even if we account for the age structure, Latinos are still under-represented in leadership roles I was able to measure; Latino adults represent 9% of Oregon’s adult population.


  •  Measurement limitations:

There are many forms of leadership for which we don’t have public, reliable, and regularly available data about race and ethnicity – like legislative staffers, non-profit executive directors, government agency heads, neighborhood association leaders, and other informal local leadership positions. It is much easier to enter these types of leadership positions than elected office, so Latinos are more likely to be in these types of roles than formal elected offices. Being unable to measure the number of Latinos in these leadership positions means I am probably under-counting Latinos in leadership.

Take-awayThe Take-aways:

  • Even if we account for the age composition of Oregon Latinos, this group is still under-represented in leadership roles (proportional representation percentage would be 9%).
  • In Extension, we often build leadership skills among our clientele and create opportunities for residents to connect with each other, which can cultivate informal and formal leadership. As a result, we have a unique opportunity to help fix this problem of under-representation by:
          • Cultivating leadership among the many Latino youth in Oregon
          • Creating, cultivating, and supporting informal leadership opportunities for non-citizen Latinos
  • What are some of the social reasons Latinos may be under-represented in leadership roles in Oregon?


I recently got a question from a colleague wondering what the difference is between growth rate and percent change. Understanding concepts like these help us understand what’s going on among Oregon’s people, places, and society, so here’s a post that deals a bit with the technical details of demographic measurement.


Nugget2The Nugget:

Percent change represents the relative change in size between populations across a time period. Growth rate represents the average amount of change per year or per month across a time period.


Two Examples to Illustrate:

Today we’ll use change in 1) the population age 18-34 and 2) the Latino population in Oregon, 1990 to 2010, as examples to illustrate the difference between growth rates and percent change.

LatinoYoungAdult Change

This chart clearly shows that both populations have grown over the last 20 years, but how much have they grown? How can we talk about their rates of change and the magnitude of increases? Calculating percent change and growth rates allow us to do both.

Percent change represents the relative change in size between populations across a time period. The formula is:

So in our example, the Latino community in Oregon grew 273% between 1990 and 2010, because there was an absolute increase in the population from 112,707 to 420,195 people. I calculated that by taking 420,195 – 112,707 = 307,488 and dividing that by 112,707 to equal 2.73. Which, when multiplied by 100 to create the percentage, yields 273%.

The Oregon population age 18-34 grew 30% between 1990 and 2010, from 678,677 to 882,922 people. That’s 882,922 – 678,677 = 204,245, divided by 678,677, which is .3; 30%.


Growth rates are trickier and not everyone uses the term the way they are supposed to (even I sometimes get sloppy with the terms!). Often people use the term “growth rate” when they mean percent change, but technically a growth rate is the average amount of change per year or per month across a longer period. There are lots of formulas for calculating growth rates because there are different assumptions you can apply to the data – like assuming the growth rate in the period was linear, or exponential, or geometric. But if we assume linear growth, the formula for the annual growth rate is:


So, in our example the annual growth rate of the Latino population between 1990 and 2010 was: 13.64%, because 420,195 (people in 2010) – 112,707 (people in 1990) = 307,488 and 307,488/20 years = 15,374.4. Which divided by 112,707 (people in 1990) = .1364; equivalent to 13.64%.

And the annual growth rate of young adults age 18-34 between 1990 and 2010 was 1.5%. As this formula shows:


Important to note is the fact that neither the growth rates nor the percent changes were equal in the two decades for either population.





As the tables show, there were higher percentage increases and greater rates of growth among both these populations in the 1990s than in the 2000s. This illustrates the importance of breaking down time intervals of change into the smallest measureable units so you can see the way growth and change fluctuate over time.

Take-awayThe Take-Aways:

  • Percent change and growth rates are different measures and each communicates something unique about population change.
  • The Latino community and the population age 18-34 have grown significantly over the last 20 years.
  • The largest increases in these population groups happened in the 1990s, but there was still growth in the 2000s.

Sorry folks, I’m still on the 40-40-20 kick. I’m working on a project to put outcome based planning into action and 40-40-20 is the outcome of the day.

So if we want to achieve 40-40-20 (recall that’s 40% of Oregon adults with a Bachelor’s degree, 40% with an Associate’s degree or post-secondary credential, and 20% with High School), it would be good to know if there are any places in Oregon that are coming close to achieving that so we can use them as models for replication elsewhere.


Nugget2The nugget:

I looked at data from the American Community Survey (ACS) across all counties and towns in Oregon, for the 2007-2011 period and found only one that was at 40-40-20: Tetherow, OR at 48-52-0. A tiny (45 people), affluent, resort community outside of Bend, where all the employed adults work as management professionals in the education, health, and social service industry.

Tetherow may be a town to learn from, but it may be a bit of an extreme example of the conditions for 40-40-20. Let’s relax our demands and look for counties and towns that come close to achieving parts of the 40-40-20 goal.


Model Counties

Bachelor’s 40% goal

  • There is one county that is at or above 40% with a Bachelor’s degree or more: Benton County – it’s at 47-7-16. Go Beavs!Evidence

Associate’s 40% goal (I can only look at the percent with an Associate’s degree because the ACS does not provide estimates of the number of people who have any other type of less than 4-year, post-secondary credential)

  • There are no counties that come close to 40% of adults with an Associate’s degree, but Sherman County was the highest, at 14% — it’s at 16-14-28, followed by Josephine, Deschutes, and Gilliam. An interesting mix of non-metropolitan and newer metropolitan counties, all with higher than average employment in construction, extraction, and maintenance occupations.

High School 20% goal

  • Five metropolitan counties hover around the 20% goal for adults with high school education:
    • Benton (47-7-16)
    • Clackamas (31-8-24)
    • Deschutes (29-10-24)
    • Multnomah (38-7-21)
    • Washington (39-8-19)

Model Towns

Bachelor’s 40% goal

  • 42 (11%) of the 372 towns in Oregon recognized by the Census Bureau have 40% or more adults with a Bachelor’s degree or more. You can access that list here: Townsw40BS. These towns are our big cities, are very closely adjacent to our big cities, or are amenity destinations high in second-home ownership.

Associate’s 40% goal

  • There are two towns in which 40% or more of adults have an Associate’s degree: Tetherow and Wamic (which is also a very small resort community, in Wasco County)
  • There are six towns that are one standard deviation unit above average in the percent of adults with an Associate’s degree:
    1. Tetherow (48-52-0)
    2. Wallowa Lake (55-26-18)
    3. Sunriver (50-22-4)
    4. Neskowin (45-21-19)
    5. Adair Village (46-16-16)
    6. Camp Sherman (61-15-14)

These towns all share some characteristics as well. They’re small and either associated with high-value natural amenities and vacation rentals or adjacent to affluent communities.

High School 20% goal

  • There are 56 towns that hover around the 20% goal for adults with high school education (have between 16% and 24% of adults with high school). You can access that list by clicking here: Townsw20.

Take-awayThe take-aways:

  • These findings suggest that there are certain types of local conditions associated with the 40-40-20 educational outcomes: natural resource amenities, affluence, adjacency to metropolitan areas, and maybe others that you’ve thought of as you read the lists.
    • Though we can’t do much about the availability of natural resource amenities across all parts of the state there may be attributes of the economies or culture of these areas that can be replicated. How would implementing those conditions affect existing populations and their qualities of life? How might we play a role identifying or trying to create these conditions? What additional data might we need about these communities?
  • The findings also illustrate that there may be some difficulty ahead in achieving the 40% with an Associate’s degree or post-secondary credential goal.
    • Very few communities have attained it and we don’t have a reputable, consistent source of data about the number of Oregon adults with a short-duration post-secondary credential.  This demonstrates the importance of setting goals for program planning that are measureable and attainable – a key lesson for outcome-based, data-driven planning.


You may have heard that there’s now a 40-40-20 goal for the state of Oregon. The Legislative Assembly declared that the mission of all education beyond high school in Oregon includes achievement of the following by 2025 [ORS 351.009]:

  • Ensure that at least 40 percent of adult Oregonians have earned a bachelor’s degree or higher.
  • Ensure that at least 40 percent of adult Oregonians have earned an associate’s degree or post-secondary credential as their highest level of educational attainment.
  • Ensure that the remaining 20 percent or less of all adult Oregonians have earned a high school diploma, an extended or modified high school diploma or the equivalent of a high school diploma as their highest level of educational attainment.

According to the 2010-12 American Community Survey we’re at about 27-8-25 right now in Oregon. So clearly, we’ve got some work to do.


Nugget2The Nugget:

In order to attain 40-40-20 there are two groups of people we really need to pay attention to: adults with less than a high school education and adults with only some college, because they fall outside the Bachelor’s-Associate’s-high school group. Unfortunately for 40-40-20 goal attainment, there are a lot of Oregon adults, at different stages of the lifespan, who have less than high school education and who only have some college. So we have to figure out if we should encourage all of them to increase their levels of education and how we might go about doing that.


Evidence for the Nugget:

  1. There are a lot of adults who aren’t in the Bachelor’s-Associate’s-high school group of 40-40-20. Between 2010 and 2012, according to the American
    Community Survey (ACS),
    • 11% of Oregon adults age 18+ (about 338,000 people) had less than high school education and
    • 29% had some college (about 872,000 adults age 18+)
    • So a total of about 1.2 million Oregonians

That’s a lot of people to be outside the desired education groups! So chances are the only way we’ll be able to make a dent in this as a state is for all of us involved in education to work together. And as Extension, we definitely will because we interact with adults across the lifespan – and targeting adults, particularly those 45 and over, is going to be key to this.

  1. As the chart below shows, about 50% of people with less than high school and some college are people age 45+, and people age 65+ make up a significant proportion of the total. (Click on the chart image to open a new window where it displays larger)

  Age by Educ Percent 10-12

  1. As the chart below shows, just over 606,000 people age 45+ have less than high school or some college education, and in each education category we see that that there are significant numbers of adults in each age category. (Click on the chart image to open a new window where it displays larger)

Age by Educ Number 10-12


The Take-Away:

Sorry folks, but this time around I don’t have lots of neat take-aways for you – just lots of questions. This exploration into Oregon’s people, places, and society is a real head-scratcher for me. Maybe for you too?

  • How do we encourage adults between the ages of 25 and 64, who have less than high school or only some college education, to go back to school to get a GED or a post-secondary credential? Many of these folks are busy people – they work for pay outside the home, they work for no pay inside the home taking care of their children or other loved ones, they’ve got kids, and they’ve got established lives. Is increasing their level of education a priority to them? Should we encourage them to make it a priority? How?
  • If we decide we do want to encourage people across the working-age life span to go back to school, are there going to be different approaches needed to encourage them to do so? What will those look like?
  • Honestly, should we be striving to increase the educational attainment of adults age 65 and over? If so, how do we realistically encourage these older adults to go back to school and increase their level of education?
  • Maybe the only real take-away I can offer is that we’re going to have to put our heads together on this one to reach the 40-40-20 goal. It might be worth some really concentrated effort among us in Extension, precisely because we do interact with a lot of people across the adult lifespan. How can we, Extension faculty and staff, help achieve the 40-40-20 goal?

I was recently inspired to look across Oregon for examples of communities that have experienced positive changes in big, tough social areas. So I pulled together 2000 and 2007-2011 county-level data on poverty and unemployment to see if any counties stuck out. Lo and behold Hood River is the only county in Oregon where poverty and unemployment actually decreased between 2000 and 2007-11! So, what’s up with Hood River County and what’s their secret to success?!

HoodRiver PovertyGraph

HoodRiver UnemploymentGraph










Nugget2Key Learning Nugget:

Depending on which stats you look at, from which agency, it looks like Hood River County’s overall economic wellbeing either improved between 2000 and 2011 or degraded slightly/severely in this period. The discrepancies can at least partially be explained by measurement errors, but the overall story is essentially that Hood River County didn’t fair too poorly during this 11 year period, despite the Great Recession. This suggests two things: be careful which statistics you use to tell your community’s story of success or failure and there just may be some things we could learn from Hood River County about moving the needle on these tough areas.



Evidence for the Nugget:

1. According to statistics from the US Census Bureau, Hood River County experienced declines in poverty from 2000 to 2007-11 and improvements to employment from 2000 to 2007-11


According to data from the 2000 long-form of the US census and the 2007-11 American Community Survey (both representing samples of the population), in Hood River County:

  • Poverty declined from 14% in 2000 to 10% in 2007-11
  • Child poverty declined from 19% in 2000 to 10% in 2007-11
  • Moderate poverty (people whose incomes are 185% or less of the poverty line) declined from 34% to 33%
  • Extreme poverty (people whose incomes are 50% or less of the poverty line) declined from 5% in 2000 to 4% in 2007-11
  • Unemployment declined from 6.5% in 2000 to 5.5% in 2007-11 (07-11 contains four years of the recession, so the recession will have a large statistical effect on the average. If the 07-11 rate is close to the 2000 rate then it’s apparent that unemployment in Hood River County, on average, was not particularly affected by the recession)
  • Employment in the agriculture, forestry, fishing, and mining industry sector increased from 14% in 2000 to 17% in 2007-11

2. According to statistics from the Oregon Employment Department and Oregon Department of Education, however, Hood River County experienced increased unemployment and poverty across the 2000 to 2011 period.

  • According to the Oregon Employment Department, unemployment in Hood River County increased from 6.5% in 2000 to 6.8% in the 2007-11 period (which is comparable to the ACS period); but increased from 6.5% in 2000 to 7.9% in 2011
  • According to the Oregon Department of Education, the percentage of students in Hood River County public schools who qualified for the free or reduced price lunch program because they lived in households with income 185% or less of the poverty line increased from 43% in 2000 to 60% in 2011

3. There is a lot of measurement error in all of these estimates.

  • Data from the long-form of the census and from the American Community Survey all come from a sample of the population. Each statistic has a margin of error between +/- 1% to +/- 4% for Hood River County. When you bear these margins of error in mind, the only statistic for which we can be 95% sure actually differed between the two time periods in Hood River County was the child poverty rate — that decline of 9 percentage points surpassed the margins of error in both years.
  • Unemployment data from the Oregon Employment Department are estimates as well. See this article for an explanation of the methods they use to calculate Oregon’s unemployment rate: They estimate that the state’s unemployment rate in June 2013 had a margin of error of +/- .8%. County level rates will have slightly larger margins of error. Applying even this .8% margin of error to the Hood River County estimates of unemployment nullifies the increases reported by the Employment Department for the time period discussed above.
  • Data from the Oregon Department of Education about eligibility of the student body for the free or reduced lunch program are also estimates, but the error in this statistic isn’t a function of pure, random sampling it’s a function of self-selection of the sample. In order to qualify for free or reduced lunch, the parents of children enrolled in school have to voluntarily complete an application for the program on which they report household income and the number of household members. If there is any reason for parents to feel uncomfortable about this reporting, perhaps they’re embarrassed about their income, they’re earning income under the table, or they’re undocumented immigrants, they may forgo completing the application. If, however, they feel safe reporting this information then they are more likely to do so. This means that increases in the percentage of kids who qualify for free or reduced lunch could be due to actual increases in the number of kids in low-income households or increases in the number of parents who feel comfortable reporting their low-income status.

Given all the measurement error, it’s probably safest to say that the overall economic wellbeing of Hood River County residents didn’t change much between 2000 and 2007-11, but that it did improve for some kids whose families were able to move out of poverty and into moderate poverty.

Now here’s the story, because in Oregon overall we definitely have seen statistically significant declines in the economic wellbeing of people between 2000 and 2007-11, and other Oregon counties experienced significant declines. So despite the less-than-perfect data, there does seem to be something positive going on in Hood River County worth further investigation and possible replication! Does anyone have any ideas about why the county fared so well in this period?

Take-awayA Few Take-Aways:

  • We should investigate what’s going right in Hood River County. Is it their industry composition; the relationship between local institutions like the school system and vulnerable populations like Latinos and low-income families; is it something about their community culture; or something else? If we identify some factors, we need to understand why they contribute to the well-being of Hood River County. Then we can consider if it might be possible to replicate in our communities. If we do try to replicate we have to monitor that it does work in our community, and modify or scrap it if it doesn’t.
  • Many of us orient ourselves toward developing programs that will “move the needle” on some significant economic and social outcomes that are measured using population estimates. So what does it take to really move the needle on these outcomes? The discussion today suggests that, statistically, it means we should aim to make an impact that results in a 5% to 10% improvement in the thing we’re trying to affect. If we don’t aim that big, we might not see any change at all because smaller changes won’t be statistically significant.
  • Aiming big for most of us will probably mean aiming for multiple, long-term, incremental improvements. It’s doubtful that Hood River County found a silver bullet that took six months to implement. Have a long-range plan for improving population level outcomes that are measured with estimates, and make sure you track your short-term wins by evaluating your program’s impact on the people it has directly served. Don’t forget that it will take a long time for the needle to move in the population. This also means, don’t pull the plug on something that’s aimed at affecting population level outcomes, that’s only been around for a few years and you haven’t seen the population change (yet). If you’re seeing the desired changes in the participants, these changes in them persist over time, and these observable changes are DIRECTLY related to the population level change you want to see, then keep faith in the program for the long-haul.

What else do you take away from this discussion of Hood River?




There’s a common belief out there that small towns are in decline. I was contacted by someone from the media a couple months ago who was looking for data that would prove just this point. What I shared surprised her quite a bit! Maybe it will surprise you too.

Nugget2Learning Nuggets:

  1. The vast majority of small towns in Oregon have increased in population over the last couple decades.
  2. Small towns that have declined are not spread evenly across the state, but small towns that increased population are present across the whole state.
  3. The demographic reason small towns haven’t declined is because as Oregon’s population increases we also see more and more people living in towns, large and small, as opposed to the outlying country-side.
  4. There are differences in the settlement patterns across counties; in some counties the vast majority of folks live in towns and in others, the majority of the population lives out in the county.

Evidence to Support the Nuggets:

  1. In 2010, out of the 245 small towns (places with fewer than 2,500 people) in Oregon, only 31 (13%) had declined in population since 1990.

    • There were about 144,000 people living in small towns in 1990 and about 186,000 people living in small towns in 2010
  2. The small towns that declined in population since 1990 were located in 18 counties (only half of all counties) in Oregon, and all counties but one had small towns that  grew.Small Town Pop Change in OR Counties - 1990 to 2010

    • All counties in Oregon, except for Crook, have small towns that increased in population between 1990 and 2010. Crook is the exception because it only has one town that’s recognized by the Census Bureau and it’s larger than 2,500 people (Prineville is the town and its 2010 population was 9,200).
  3. The population of Oregon increased by about 1 million people between 1990 and 2010. At the same time, the percentage of population living in towns went from 70% in 1990 to 79% in 2010.

    • In other words, in 1990 30% of Oregonians lived in the “country-side” (villages and areas outside of town limits) and now in 2010 only about 20% of Oregonians do.
    • The concept of a populated country-side is “dying,” not small-town life. About 52,000 fewer people lived outside of towns in 2010 than did in 1990, while the population living in small towns grew by about 42,000 people.
    • Why do you think fewer Oregonians are living in the country-side?
  4. BUT, the population living in the rural country-side isn’t gone and it isn’t dying everywhere in Oregon!

    • In Lake County, Crook County, and Polk County fewer than 50% of the population lived in towns in 2010. So this means that the majority of people in these counties live out in the county, outside of town limits.
    • Four counties in Oregon actually saw increases in the percentage of people living outside of towns recognized by the Census Bureau. Can you guess which four? I’ll give you a hint; they’re all non-metropolitan counties…

 A Few Take-aways:

  • Take-awayWe should be planning our programs anticipating modest growth in small-towns. People don’t just move to our big cities in Oregon.
  • We should recognize that our rural populations, though still rural, are increasingly living in closer proximity to one another and our programs should reflect the needs that come along with small town life as opposed to life in the country-side.
  • There are counties where the bulk of the population doesn’t live in towns. In those cases we should plan to invest significant resources in reaching those across the county, outside of the town centers. Also, we need to bear the lifestyle (longer travel times to work and services, more place-bound activities) and the values (perhaps related to a desire not to be tied to city ordinances, taxes, and rules) of this population in mind when we design our programs.
  • What else do these data suggest to you about how we should be thinking about Extension or other programs?


It’s well-known that young people typically leave rural areas in search of higher education and work opportunities. There’s also a common belief out there that this out-migration of young people represents a “brain drain.” Recently, a University of Minnesota Extension study showed that though young people (18-29) might be leaving rural areas, between 1990 and 2010 rural areas across the US were attracting people age 30-49. These authors argue that this is evidence of a rural “brain gain.” So the question for us is, is this true for Oregon? Is this really a brain gain?

Nugget2Key Learning Nugget:

Between 1990 and 2010 rural Oregon did attract people age 30-49, but these in-flows of people weren’t necessarily associated with gains in the overall education levels of the population – we actually can’t be sure if this in-migration is synonymous with “brain gain.”

Evidence to Support the Nugget:

  • The charts below show that the 1990s and 2000s saw net in-migration of 30-49 year olds in non-metro Oregon


Positive values indicate net in-migration and negative values indicate net out-migration of people in the age group.
Positive values indicate net in-migration and negative values indicate net out-migration of people in the age group. Source:


Positive values indicate net in-migration and negative values indicate net out-migration of people in the age group.
Positive values indicate net in-migration and negative values indicate net out-migration of people in the age group. Source:
  • Net in-migration of 30-49 year olds is not associated with gains in the level of education among the population.
    • Correlation of non-metro net migration rates with change in the percent of adults age 25+ with a Bachelor’s or more actually reveals a negative relationship in the 1990s and 2000s (rho = -.47 for 1990s, rho = -.08 for 2000s). This means that counties with higher net in-migration rates of people age 30 to 49 had lower growth in the percent of people with a Bachelor’s or more in these two decades. In other words, counties with high rates of in-migration of 30-49 year olds saw low growth in educational attainment, while counties with low rates of in-migration of 30-49 year olds saw high growth in educational attainment.
    • According to linear regression, these negative correlations were statistically significant for the 1990s, but very small (b = -.0057), and non-existent for the 2000s. This ultimately means that in the 2000s net in-migration of 30 to 49 years had nothing to do with changes in educational attainment among the population, and in the 1990s net in-migration had only a small amount to do with changes in educational attainment, and they were negative.
    • There may be a few reasons for this finding:
      1. The education levels of rural 30-49 year old in-migrants in the 1990s were actually relatively low.
      2. We aren’t accurately measuring brain gain. Instead of using overall educational attainment in counties perhaps we need to be measuring the “brainy-ness” of the in-migrants themselves. Unfortunately, we don’t know the education levels of these in-migrants because the data don’t exist.
      3. The effect of in-migrants age 30-49 on the overall education level in the non-metro counties may be muted by the presence of other age groups and their education levels.

Take-awayThe Take-Aways:

  • New-comers, age 30 to 49, are a reality in our rural communities. This means we can think about and talk about rural communities in Oregon as places of growth in this respect. Out-migration of youth can happen at the same time as in-migration of middle-aged adults. It also means that we shouldn’t forget to include these new-comers in our programs. They may have some cool ideas about new programs or ways of offering current programs and they’ll likely benefit greatly from being involved!
  • The data also show us that we can’t infer that the in-migration of middle-aged adults to rural areas represents a brain gain. If we want to find out about the education levels of this new population, we need to gather better data.
  • What else do you take away from these data?