“Smaller Share of Women Ages 65 and Older Are Living Alone,” before and after age adjusment

After noticing this from a recent Pew Research report:

ST_2016-02-18_older-adult_0-01

Ben Hanowell wrote:

This made me [Hanowell] think of your critique of Case and Deaton’s finding about non-Hispanic mortality.

I wonder how much these results are driven by the fact that the population of adults aged 65 and older has gotten older with increasing lifespans, etc etc.

My collaborator Jonathan Auerbach looked into this and found that, in this case, age adjustment doesn’t make a qualitative difference. Here are some graphs:

Percent of adults over age 65 who live alone, over time, for women and men, with solid lines showing raw data and dotted lines showing populations adjusted to the age x sex composition from 2014:

adjustment_type

The adjustment doesn’t change much. To get more understanding, let’s break things up by individual ages. Here are the raw data, with each subgraph showing the curves for the 5 years in its age category:

adjusted_ind

Some interesting patterns here. At the beginning of the last century, a bit less than 10% of elders were living alone, with that percentage not varying by sex or age. Then big changes in recent years.

We learn a lot from these individual-age curves than we did from the simple aggregate.

In this case, age adjustment did not do much, but age disaggregation was useful.

Jonathan then broke down the data by age and ethnicity (non-Hispanic Black, Hispanic, and non-Hispanic White; I guess there weren’t enough data on Other, going back to 1900):

ethnicity

To see people blogging about it in real time — that’s not the way science really gets done. . . .

Seriously, though, it’s cool to see how much can be learned from very basic techniques of statistical adjustment and graphics.

And maybe we made some mistakes. If so, putting our results out there in clear graphical form should make it easier for others to find problems with what we’ve done And that’s what it’s all about. We only spent a few hours on this—we didn’t spend a year, sweating out every number, sweating out over what we were doing—but we’d still welcome criticism and feedback, from any source. That’s a way to move science forward, to move the research forward.

P.S. I sent this to demographer Philip Cohen who wrote:

There are two issues with 65+: 1 is they are getting older, 2 is they are getting younger since the baby boomers suddenly hit 65.

Here’s a post by Cohen on this from a couple months ago.

21 thoughts on ““Smaller Share of Women Ages 65 and Older Are Living Alone,” before and after age adjusment

  1. I wonder why they didn’t use the 2010 decennial census data. Apparently the definition of “living alone” needs to be given for each different study cited:

    “In the analysis for the 2014 and 2009 Pew Research Center surveys, older adults who live alone may be in living in their own home or in an assisted living facility by themselves. Older adults living with others may be living with their spouse, children, other family members or non-family members”

    So without some assurances that the definitions are comparable from year to year I would be very wary about the changing data source. However, even if the data was all from the US census, the question may change from year to year. Maybe the explanation of the “living alone” term became harder/easier to understand for whatever reason. It definitely took me a few minutes to satisfy myself that someone living with family members was indeed “living alone” for this data, I wonder how many people actually understood the question.

    • I need to take another look at these definitions. Living in an assisted living facility should not be considered as living alone even if you have a private room. You eat in a communal setting and take part in regular communal activities. You aren’t living with a family member but you certainly are living alone.

      • Well not always, it really depends on the facility and level of care. The younger people there often have their own kitchens and don’t buy the meal plans (just based on my in laws).

        • That is true. But the point is that when you categorize all people living in a private unit in an assisted living facility as living alone, you get a mix of people, many of whom (perhaps more than half of whom) are in a residential living situation that is far removed from living alone.

        • So I think the problem with any definition is that once you apply it you realize the exceptions. In NYC lots of old people live in apartment buildings. THey may go to a senior center every day for lunch or get meals on wheels. Or they may just meet at McDonalds or a coffee shop every day (you see them if you walk around). Some place like Co-op City which is an NORC (naturally occurring retirement community) http://www.nyc.gov/html/dfta/html/services/retirement.shtml will have many services available.
          But at the same time, it’s important that “living alone” does not ever mean “being alone all the time” or “having no support structure.” That’s like when people think that never marrieds who are childless must be lonely and sad, and many are happy and have lots of friends.

    • Pew chalks this pattern up to increased life expectancy among men, which causes fewer women than before to be categorized as living alone. I would also like to know the extent to which the increased use of assisted living facilities explains the trend. Probably not as much as life expectancy changes but maybe there is something there.

    • ACS is considered vastly better than the census. The census is just the way it is for redistricting not because anyone thinks it is the best estimate of anything.

    • Anoneuoid said: “It definitely took me a few minutes to satisfy myself that someone living with family members was indeed “living alone” for this data”

      Are you sure this is the correct interpretation? The problem seems to lie in the interpretation of the phrase, “older adults who live alone may be in living in their own home or in an assisted living facility by themselves”. I see this as ambiguous, with two possible interpretations:

      Possible interpretation 1: “older adults who live alone” may be either (a) living in their own home by themselves or (b) living in an assisted living facility by themselves.

      Possible interpretation 2: “older adults who live alone” may be either (a) living in their own home or (b) living in an assisted living facility by themselves”

      • >’Are you sure this is the correct interpretation? The problem seems to lie in the interpretation of the phrase, “older adults who live alone may be in living in their own home or in an assisted living facility by themselves”.’

        Actually, it isn’t clear to me whether that quote is referring to the same or different data source used for the current report. It sounds like it refers to a different data source, because from the pdf:

        ‘“Living alone” refers to people who live in their own homes, not in group quarters. Group quarters are places where people live in a group arrangement, which is owned or managed by an organization that provides housing and/or services for their residents. Group quarters include nursing homes, in-patient hospice facilities, mental hospitals, group homes and other institutional and non-institutional living arrangements. Generally, assisted-living facilities are not defined as group quarters. Therefore, individuals who live in an assisted-living facility unit may be counted as living alone if there are no other residents in their unit. For more details see:
        https://www.census.gov/popest/about/terms/housing.html

        http://www.pewsocialtrends.org/files/2016/02/ST_2016-02-18_older-adult-FINAL.pdf

        • The quote from the pdf suggests the interpretation I labeled (1) above. But neither quote is ideally written — the one from the pdf appears to be written from a historical perspective, which does have the advantage of more detail.

        • >”The quote from the pdf suggests the interpretation I labeled (1) above. But neither quote is ideally written”

          Thanks for taking a look, it is this type of stuff that used to make me think I was crazy. Maybe I missed it, or they just do not clearly explain what they did, because “100 – (% of people living in group quarters)” seems like a pretty misleading definition of “% of people living alone”. However, I did try pretty hard to figure out what I think should be considered one of the most important parts of the report (how was the primary outcome measured) and had a lot of trouble.

          Still, even my “pretty hard” was only skimming these documents. Even so, as best as I can say it is unclear whether or not “living alone” includes living in your own home with family members. Just to put it all in one place, here is what I have read:

          “Older adults who live alone feel more financially strapped than older adults who live with others. 2”
          http://www.pewsocialtrends.org/2016/02/18/smaller-share-of-women-ages-65-and-older-are-living-alone/

          Where I expect 2 to be a citation, but instead it was a footnote reading:

          “In the analysis for the 2014 and 2009 Pew Research Center surveys, older adults who live alone may be in living in their own home or in an assisted living facility by themselves. Older adults living with others may be living with their spouse, children, other family members or non-family members.”
          http://www.pewsocialtrends.org/2016/02/18/smaller-share-of-women-ages-65-and-older-are-living-alone/#fn-21393-2

          Now, that footnote refers to 2014 and 2009 Pew Research Center surveys, while the headline figure (shown at the top of this blog post) includes the caption:

          “Source: Pew Research Center Analysis of 1900-2000 decennial censuses and 2010 and 2014 American Community Surveys”
          http://www.pewsocialtrends.org/2016/02/18/smaller-share-of-women-ages-65-and-older-are-living-alone/st_2016-02-18_older-adult_0-01/

          So are the 2009 and 2014 surveys the same as the 2010 and 2014 surveys? I don’t know for sure but I think not. Anyway, they also have a section called “Terminology” on that same page:

          ‘“Living alone” refers to people who live in their own homes, not in group quarters. Group quarters are places where people live in a group arrangement, which is owned or managed by an organization that provides housing and/or services for their residents. Group quarters include nursing homes, in-patient hospice facilities, mental hospitals, group homes and other institutional and non-institutional living arrangements. Generally, assisted-living facilities are not defined as group quarters. Therefore, individuals who live in an assisted-living facility unit may be counted as living alone if there are no other residents in their unit. For more details see: https://www.census.gov/popest/about/terms/housing.html

          Once again, it is unclear to me if, let alone how, they determined whether people were living at home with family members. Then I saw a link to the “Complete Report PDF” at the top of the page and reasoned that may contain more detailed information. There the “Terminology” section quoted above is repeated on page 3, which as mentioned already, I found unclear. However there is also a “Methodology” section on page 23 where it says:

          “The analysis in this report is based on two nationally representative telephone surveys conducted in the United States in 2014 and 2009…
          The trend analysis on living arrangements of older adults utilizes the 1990-2000 decennial U.S. censuses and the 2010 and 2014 and the American Community Surveys (ACS) provided by the Integrated Public-Use Microdata Series (IPUMS). The IPUMS assigns uniform codes, to the extent possible, to data collected in the censuses and ACS over the years. Figures for 1900-2000 use the 1% samples, including the 1970 Form 1 metro sample. More information about the IPUMS, including variable definition and sampling error, is available at https://usa.ipums.org/usa/doc.shtml.”
          http://www.pewsocialtrends.org/files/2016/02/ST_2016-02-18_older-adult-FINAL.pdf

          So it looks to me like that “#2 footnote” refers to a different source of data than that shown in the headline figure. Also, they say to follow that “usa.ipums” link for more info. Unfortunately this only leads to a general page, they don’t cite exactly what sections being referred to. However, I clicked “Variables”, hovered over “Household” and clicked “Group Quarters”. There I saw a few options, with the one called “GQ” labeled “preselected. So I clicked on that GQ and read:

          “There are three slightly different definitions of group quarters in the IPUMS. For the period 1940-1970 (excluding the 1940 100% dataset), group quarters are housing units with five or more individuals unrelated to the householder. Before 1940 and in 1980-1990, units with 10 or more individuals unrelated to the householder are considered group quarters. In the 2000 census, 2010 census, the ACS and the PRCS, no threshold was applied; for a household to be considered group quarters, it had to be on a list of group quarters that is continuously maintained by the Census Bureau. In earlier years, a similar list was used, with the unrelated-persons rule imposed as a safeguard.”
          https://usa.ipums.org/usa-action/variables/GQ

          Still, it is not clear to me if, let alone how, they determined whether people were living at home with family members. I think I gave a good faith effort to find a clear, straight-forward explanation of how they measured the primary outcome, and was unable to do so under the assumption that they somehow attempted to determine whether the person lived with family members. They do very clearly state that ‘“Living alone” refers to people who live in their own homes, not in group quarters,’ and link to outside information describing group quarters. So I figured this must be the extent of what they did.

          If it was me, I would attach the actual questionnaire for each year and show examples determined to be “living alone” and “not living alone”…

        • This discussion illustrates a couple of common problems with measures:

          1. Their definitions are often not given (or not clearly given) when results are quoted.

          2. The definitions are not always what the names might suggest.

          Re (1):
          a. Yes, your idea to “attach the actual questionnaire for each year and show examples determined to be “living alone” and “not living alone”…” would be much better than what was done.

          b. I can only speculate on possible reasons why such sloppy definitions are given. Some possibilities: The people writing don’t really know what they are talking about, or are poor writers, or are working on an unreasonable deadline to get the results out; the culture doesn’t stress writing or having someone critique work before publication; or the strictures put on from above (e.g., by Congress, for government agencies) don’t allow the time and money needed to do a good job.

          Re (2): An example I often give is the official U.S. Unemployment Rate, which includes as people who are employed “unpaid family workers, which includes any person who worked without pay for 15 hours or more per week in a business or farm operated by a family member with whom they live.”(http://www.bls.gov/cps/cps_htgm.htm#concepts)

  2. Apparently the definition was changed for the 2000 and later data (now “living alone” means you were not living in a building on some list of group quarters):

    “There are three slightly different definitions of group quarters in the IPUMS. For the period 1940-1970 (excluding the 1940 100% dataset), group quarters are housing units with five or more individuals unrelated to the householder. Before 1940 and in 1980-1990, units with 10 or more individuals unrelated to the householder are considered group quarters. In the 2000 census, 2010 census, the ACS and the PRCS, no threshold was applied; for a household to be considered group quarters, it had to be on a list of group quarters that is continuously maintained by the Census Bureau. In earlier years, a similar list was used, with the unrelated-persons rule imposed as a safeguard.”
    https://usa.ipums.org/usa-action/variables/GQ

  3. My feeling is that if you are just using 65+ you are probably still a bit on the young side to see the impact of age adjustment due to baby boomers. You can see that with the age disaggregation where the real change is among the old old. The 65-70 year olds would be boomers. In the data for older people you are dealing also with the smaller depression cohorts rather than the larger baby boom ones, so the impact on age distribution within the 10 year blocks for the people in their 80s will be different.

    Of course what is interesting from my perspective is not so much why is it “going down” as why was it so consistently going up for all those years and why was the male/female gap so big for NH Whites in all those years, while now maybe we are seeing a trend back toward what was true before and is more the pattern for other groups, with little difference by gender.

    I can think of a few stories.

    Higher proportion of war widows than of war widowers among the WWII generation. And the smaller impact of war in later cohorts.

    I think the whole issue of grandchildren is a bit of an under-analyzed one, given how many kids in their 20s are still with their parents, I wonder how many are with their grandparents (who may have more space). Anecdotally I can say that I’ve known a number of unemployed 20-somethings who have taken on the live in care-taker jobs since their parents are still working.

    Now unlike the life expectancy example (where we know … total death is only increasing over time) living alone is a bit of a conceptual issue too. As you see in the note, it excludes people living in nursing homes or assisted living facilities. Maybe single baby boomers are more open to living in group quarters.

    • “You can see that with the age disaggregation where the real change is among the old old.”

      I’m bad at precisely mapping aggregate chart positions to numbers, but isn’t that partially an illusion of the graphs having the same vertical scale? For the youngest age group, the peak is around 30% compared to something like 25% now, so 1.2x, and the second-oldest age group it’s like 64% versus maybe 48%, so 1.3x, a little bigger of an effect.

      I mean, you might be right, but I’m just not good enough to tell from the graphs.

  4. Could you say what the “adjustment” is, how to interpret it, how to perform it, and why you feel it is appropriate for estimating something in particular?

  5. Is 65+ a reasonable way to present things over several decades? Presumably, life spans became greater and health status for people of fixed age is changed. Now, if we have a program like Social Security or whatever with hard age cut off (AFAIK, it is not, but it’s close) and leaving alone changes your status in this program it might be what you need, but as a general societal interest, why take this approach. Or is it just easier to deal with and it is not all that different from settling on a less arbitrary choice?

Leave a Reply

Your email address will not be published. Required fields are marked *