This joint CPC-CG/CRA Webinar took place on Thursday 21 March 2024 at 12:00 UK Time. Klara Raiber of Radboud University presented, “Future pathways of unpaid caregiving research: Conceptual and methodological challenges”.
Title: Future pathways of unpaid caregiving research: Conceptual and methodological challenges
Abstract: More and more research focuses on unpaid caregiving, which is here defined as care towards friends or family members with health issues. In my talk, I will highlight the existing literature on the topic of unpaid care and the consequences thereof, including my own work. Building on this, I will sketch missing links and where I see future pathways both theoretically and methodologically.
Speaker: Klara Raiber is a Ph.D. candidate in Sociology at Radboud University and the SCOOP program. She holds a master’s degree in Sociology from the University of Mannheim in Germany. Her Ph.D. project is on long-term employment consequences of caregiving for family and friends with health issues from a life-course perspective.
Find out about more webinars and events at www.cpc.ac.uk.
Uh start with my typical thank you very much I’m really so glad to be here again also because I just said it nothing has much change and sometimes I really like if there’s not much that changes so I’m really glad to feel a bit of the old feelings and they’re very positive ones
For being here um yeah I thought I’m very Smart in picking this title because I thought I can put all my research in there and then give you idea of what comes next but when I Was preparing I was like why did I do that now I have to
Figure out what’s next so it’s more about the past and challenges and then some ideas of future research now I’m will be able to click yes perfect thank you so the content will be basically I will give you some definitions what is the unpaid care I’m
Looking at and how I approach it in my research then I will go a bit into the conceptual challenges some about methodology and the future Pathways with for question mark I have two slides about future Pathways um but yeah sometimes also great to not give you all
My ideas away um but I also have to say with the future Pathways um I’m very thankful that I’m not in this field alone so also there I will highlight who is also having these ideas or maybe someone also had these ideas before me and I’m jumping in on that and find it
Very important to highlight that so you will know some names throughout the presentations you will hear some more often that are more important to how I was shaped as well so a bit of um the unpaid caregiving so my definition is basically I understand in my research
Care unpaid care as the care for family and friends with health issues so I normally exclude the regular child care and this is also often known as informal Care Family Care I’m saying now that I’m use all of these terms because now I have sometimes res say don’t use
Informal care then I use unpaid care and now res says I should not use both of them and should go to Family Care which I normally don’t do because in the Netherlands there is also a lot of neighbors that get help and care and also friends so there’s quite a higher
Percentage if you would compare to the UK um in the Netherlands there’s like about 20% uh really not to family directly so that’s why I normally don’t use the firm I don’t have to explain I think the relevance of this all in this room but just a very short recap there is aging
Everywhere um it’s a global phenomen and there will be the problem that there will be not enough care both from paid care as well from unpaid care so there will be UNM Nets in the future so there is relevance to do that better the research that we’re doing now but also
An example why it’s a very broad definition I’m handling so this is a a very long Sur question we ask in ad that’s questionnaire but I want to highlight how diverse it basically can be so here we say okay really about health issues it can be also of old age it could be
Partner family member friend a neighbor and acqu with a colleague um that really needs because of their physical limitations or the mental limitations they need help help and uh it can be really just going to the supermarket to really keeping company or medical and practical care so it’s really a lot of
This and that makes our research sometimes very complicated because sometimes you’re really interested one type of care like to a partner or you’re more interested in really different kind of um things you can do in this care like sometimes really about the medical and practical help but sometimes it’s
Also really important to understand the emotional we know the loneliness studies in this room so this is also a very important part what is also interesting also in this definition we put it it’s also sometimes very short periods but it can be also a very long period so if you
See such a broad definition that’s basically the conceptual issue I’m seeing because it is so diverse in a sense and um here you see the the people I’m also thinking along with um that are also interested in also these terms uh we need definitions in every survey it’s
Bit differently uh ask sometimes you know for how many hours someone cared sometimes you know for whom care was but sometimes you don’t have this information and it’s quite important to have this information and what I also told you before is UNC can be very diverse there can be very different
Forms so these are the upper ones are both call you from I noted it down because oh no I don’t have my I think it’s the German Center for gerontology in Berlin and yes those are also looking at employment consequences and is now at the University of constance and we were
Thinking about um joining a network together uh a German based Network and I’m the person from the Netherlands but I’m always will stay connected to the German Academia also because uh yeah I’m now even living in Germany again over the Border not exactly close to Berlin
Or anywhere but yeah I was always stay connected so with those people we are think thinking about that we need clearer definitions and terms and the life course perspective was there I think from the beginning of my time on my PhD it was even in the
Title at the end it didn’t make it in the final title anymore but it was really the perspective that I was looking into um and I think even in the first year of my PhD I made these slides I think I made them for the live course
Course we are teaching but just to show how diverse it is it can be a very different kind of duration can be very short uh when for instance your partner breaks their leg then they might need have it can be for longer but it can be
Also lifelong for instance if you have a child with a disability that can be really um going on and really until the end basically and I always then also show employment that’s not only the focus today but their employment is then hardly intersecting there it’s also the
Timing when caregiving starts it can be in early age often it’s in middle age but not only so for especially if we have surveys that are only focusing on older age we miss kind of all this young age care and that of course then inct also different with your career building
Stases or your retirement and depending on that and the last one is also just frequen see we know that some people I I think I can also already skip to the next slide because that’s a research we did where we were looking at different kind of patterns that you can see if you
Have like complete histories of caregiving then you can see some patterns so the original idea is from Canada and we replicated that in the Netherlands and we also see that there is special kind of trajectories there is for instance the onetime partner care so a long time you don’t care and then your
Partner gets sick and um then only you care for your partner that’s often men who care for their wies and but there’s also some people that are really like serial care and lifelong care that’s often women that an early age started to care then adapted their employment and because they adapted their employment
Are always the ones that get asked so we also know that there’s really patterns going on but yeah it’s very important to take all these patterns into account because they what I already now mentioned are intersecting with gender they’re intersecting with work they’re intersecting with their own health but
It’s something to really take into account so that’s a one of the challenges to really understand how caregiving to very different people in very different stages in your life uh has impact on others but when I’m talking about impact on others or other things then I also mean that it’s quite
Difficult in a methodological sense because there is a bit of an issue it’s the selection issue and with selection issue I don’t mean um that we are not maybe not always able to sample correctly but even if you sample correctly and you have a represent to
Samples then the issue is that there is a selection into those that take up the care and so it’s not a random selection into those caring and those not caring so if you’re like me interested in all these outcomes of caregiving it is quite challenging to figure out what is going
On because they’re very different to compare with caregivers and non- caregivers I really love it that in all languages it’s another fruit but apparently it’s apples and oranges in English that you shouldn’t compare apples and oranges I don’t recall what even in German anymore but I
I love that that this is aars Apple and bears too yeah which are more close actually so that that makes even more yeah so um where was so yeah that it’s very different groups to compare with so but we want to know what the caregiving has effect what kind of
Effects it has on someone’s life um so we kind of have to take that into account if you want to do both qualitative and quantitative research but I’m a very quantitative kind of person so now it will be very quantitative also my solution so some examples where you really like
See because the issues really when both the selection into the caregiving also influences your outcome so for instance if there is a person that has lower working hours this person is more likely to start caregiving because we know you need time to care so there’s a selection of especially people for
Instance of part-time that select themselves into caregiving so if you’re later interested in the wage of the caregiver you have to take into account that they had lower working hours and probably due to that also lower wages beforehand and have a very different kind of directory in their wages to
Begin with so you have really to account also for the lower working hours and wait before the caregiving even started another example is occupations that certain occupations have more flexibility and then if have an occupation that is quite flexible you’re then the one that always gets called like okay there’s a medical appointment
Could you be the person to drive the person to the medical appointment so also here the selection is um that your occupation makes you more likely to be a caregiver or not that can also be the case so if you’re again interested in the wages then we also know that the
Flexibility or the the occupation that are more flexible often are the ones that are also less page so um there again you have to kind of also account for the equation there’s also an example on health um that will also come back later um because if you have parents that have
For instance dementia you you need to have someone to care for to select into caregiving then you’re more likely to select into caregiving if your parents have dementia but if you’re also interested in a caregivers cognitive abilities you have to take into account that there is some genetic overlap in
Dementia so you have to take into account that this person might have the same prositions so that is also a selection that is maybe not that social but also something to you have to take into or you don’t have to take into account but if you really want to know
If the caregiving is decreasing your cognitive abilities or increasing you will see later um then you really have to also take it into account so how how do I do that normally um the one first step is of course you have to know what is going on
What is the selection process going on because only the things you know you can later account for and so here the the example is the inform care model byun and Al and they made a very nice model that can help you if you’re interested in unpaid caregiving can help you to really
Understand who are the ones selecting into caregiving and in this case it’s really like you’re are in the context that’s important but what I’m always most interested in is also the disposition of the caregiver in more detail and they have these three questions do I want to
It’s more about affections do I really want to be a caregiver do I like the person I’m caring for then there do I have to that’s about normative beliefs and there I’m now doing really nice research with my markers she’s at the University in Amsterdam and she made a
Really nice survey experiment about uh what if your your neighbor has a broken leg under which circumstances would you help and then you have different scenarios and we could match that to our caregiving data so we are also in this study taking into account your experience from the past so if you have
Cared before but what P comes out and which is very clear that reciprocity is really the driving Factor we know that already from children and parents that you really feel that you have to because they took care of me when I was a child and now I
Have to feel the obligation to do that later on and we see that this even works with neighbors that if you have the feeling that this person would help you in the future or you know this person helped you before you’re more likely to have this person in the
Future and can I is also really about toce various to provide care here about distance for instance the one if you have siblings the one living closer to your parents might be the one taking care the one time we had discuss already if you work part time you’re more likely
To be the one and taking care so it’s really important to understand the selection process and models like this can help to do that there are probably many other ways but it’s really about Al the thinking process that’s behind and we also tested this with uh the pandemic because we
Thought it’s basically looking into the future um where you know kind of okay there’s less formal care there’s less time or time in a sense of also more barriers so we checked all these different kinds of um um characteristics and we saw all of them are really
Working out if you have this um phenomena of the pandemic as an examp to test this model but uh sometimes it’s really not that easy to measure everything we know that but also here there are some things like I highlight attitude Norm belief competence yeah that’s very difficult to
Measure so that’s very difficult so I have some tricks how to avoid it the first one uh um yeah the first one going I kind of want to go to the one that I like the most but it’s still not coming um but but I give you one case um where
I couldn’t do my my most favorite method but I came up with something else so what we did in March 2020 which was also the basis for most of my PhD chapters U was that we were able to ask um the complete sample of a representative sample in the Netherlands if they ever
Cared for someone in their life H yes I do still do so yes I did in the past and no I never did this and then the ones that said yes we could give a very long questionnaire where we also were able to ask like starting at end points and for
Some even more information that’s also how we made these nice crafts that you saw before with trajectories but that you really need to know the beginning and the end of all these episode in a person’s life but my PhD was actually about employment outcomes which were not
That many in the survey itself uh but I was like okay I have some information in Cent wages but I also want to take into account the life course um how do I do that I want to compare to non-caregivers but on the non-caregivers I only have like the classical information but of
Course nothing on the caregiving because they never cared so what would I do and I came up with the idea then of matching uh and you see it’s it’s published in advanc life course research if you want to read up on that but uh my solution was to match very similar
Caregivers to non-caregivers on split by sex the same sex similar educational level uh just having a partner children living in the household in same age so where you would say they normally would have the same kind of wages going on and then what I did I matched them
And then I calculated the wage difference and used the wage difference as an outcome variable so I explained the wage difference or wage penalty because in most cases it was a penalty for caregiving and then I could uh really also related to the life course uh factors that you saw before the
Duration of caregiving from how many people did they care before and also what the timing when was the first time they cared and uh so it was really nice way of uh of one solution is always matching but in this case also had the advantage that I could also include all
These life course factors into my analysis as well um if you want want to read the paper short if the more people you care for the bigger is the penalty I think that’s the most takeaway that I can give you from this paper um you can also only look at
Caregivers themselves and that helps you to ignore the whole process of selection into caregiving of course some very different kind of questions here I looked at different strategies of that caregivers can choose to adapt their work so I looked under which circumstances someone would reduce their working hours under which circumstance
Someone would choose to stop working completely under which circumstance someone changes um their jobs or become self-employed because those strategies are not looked much in the literature so far and we know from child care that those are also relevant but I can already tell you that not a lot of
Unpaid caregivers change their jobs or become self-employed because of unpaid care giving it’s about just the typical strategies of reducing working hours and stopping to work so it is a solution but maybe not for the questions you’re interested but now I think now I really that’s what I’m normally doing I’m
Normally calculating fixed effects that’s my panel regression that’ss my go-to strategy and what do you do when you do that so you need panel data so you need information over years and you really look at how changes in your variable um in changes in this case of caregiving so
Starting to care or even starting to intensively care how that relates to the outcome of Interest so how do you do that that that’s called demeaning so you kind of the mean you take the mean over all years of a person and each of the observations get
Distracted by the mean you don’t have to really go into that we don’t have to go into that in depth but what I want to give you away that’s is a solution where you can account for all the stable characteristics of a person and then if
You think back to the model there were quite some stable characteristics in there that you want to have accounted for like the normative beliefs we know are formed very much and during the first years of your life and then don’t change that much anymore but um maybe they change because of caregiving that’s
A question to still be answered um and what you also do you don’t compare caregivers to non-caregivers but you compare caregivers to their previous self so before they started so you can really see how the caregiving made have changed so that is a a solution I’m using quite frequently how my time was
25 25 what 25 perfect perfect because I have three examples of that the first one I did in a building there there with these beautiful people that sit here in the front um with j and Maria and there we were interested in to what extent is unpaid carrying related
To Jan and job satisfaction and we were also interested here in the care gaming intensity and duration and here you see already that you need quite high quality panel data to do this kind of analysis and understanding Society because I never can pronounce this it’s why I’m always saying understanding Society it’s
Really a good example of high quality data that you can use for these kind of analysis or you have information both unemployment outcomes as as well as um the caregiving um it is now a waiting recommendation so we will see what will happen with that but hopefully you can
See it out soon um to not wait that long and takeway is that um especially if you start non-intensive care it lowers your CH satisfaction then the CH satisfaction goes up again when you intensively care so there might be something positive going on there that’s the longer you
Care you might be really uh no not the longer in the sense but the more intensive you care there might be some going back to the normal standard again um but with caregiving duration we see the longer you care the less happy you are with your job so you really reduce your job
Satisfaction and the the second example I’m having here is working hours employment status that was with very good German labor market data where they have a lot of caregiving questions as well so I also recommend the German labor market and Social Security data even better I would say no camera but I
Will still say that I think it’s better quality than a sub the classical German data and also here really what you would accept expect is that um especially with high intensive caregiving working hours are reduced so here you see the working hours are on average lower if you’re an
Intensive caregiver and you also way more likely to stop working completely and the last one is the work in progress and I hope to next week um meet Patrick in Oxford or we have a one day planned on really doing the analysis so there we are interested in health and
Physical health and because you of course I love vious cycles and every presentation has to have one vious cycle so if we have more older people people with health issues but if caregiving also makes you foren your health then we have a circle of that is even more
Difficult than the the lack of care that is to be expected and there we look at two outcomes ctive functioning as I told you before which is has to be controlled the selection of althetic load I think Patrick would be way better in explaining anesthetic load but I
Understand as the Veer and teer of a um stress model so they take blood and in the blood they can see how stressed you are that’s my very bold way of saying that um and we use the English longit stud of Aging for that because they we
Have this information on blood sample as well as really good memory tests on cogn functioning so Al here really good high quality data that can be used and some preliminary results um but we also want to change the the analysis a bit but what you can see something positive that
Um all changes away from not having care before are associated with better cognitive functioning mental health we we looked if that’s moderating but that didn’t make a difference and with anesthetic load there we have L statistical power because there you have the issue that of course they don’t take
The blood of the respondents every here but every second so it’s a bit more blurry what’s going on there but you also see the starting intensive here it’s associated with lower anetic load so more stress and I have also another solution but there you need very detailed uh very very detailed
Information and let’s grow curve modeling there it’s really about how does something develop over time after and before the caregiving started and um it is also really then accounting for the starting position so if I will give you my example maybe then it’s getting clearer if we look at
Hourly wages then it’s about of course those that later might may care are the ones that are later Al are already having lower wages so I will go back before you sneak peek so it is really about taking into account the the different starting levels uh that was the last paper of my
Dissertation so it’s um also I have to do some some changes for a journal so maybe it’s also in a good way to to be out there soon and we also differentiated between intensity and duration and really wanted to see how unpaid caregiving is influence wage development and I had the opportunity
And possibility to match the data that I told you before where we know the beginning and the end of each car gaming episode to register data so monthly data on based on taxes so we really know what wages people had at that moment so it’s really high quality data over 19 years
So that’s a really nice data do you see so the the Gren one is um women and the black one is men but these you see if it’s two lines yeah yeah good and we were also the discussion if we should zoom in but there are there are different in
Significance but showing you like this gives you the hint that the effects are not that big but of course you know with wages so when you have some changes they build up through the career so it is something but what you would see if you would zoom in which I don’t wanted to
Because want to be open that there are pretty small effects but if you would zoom in you would see that those men that sto caregiving so when they stopped they were able to increase their wages more as if they wouldn’t have cared so there is like um similar to the
Literature on child care there is like a care benefit effect but only for men like the daddy bonus would you say in the child care literature and this Betty bonus or this caregiving bonus in this sense we also see for men that care intensively um so also there we for
Intensive care they can benefit while women are really disadvantaged by intensive caregiving there’s no uh duration effects to be found but uh I think that’s the most catchy result of my dissertation so anyway it will come out disertation will be so yeah that’s um another strategy so
That was kind of the ways I handle the selection issue and also the the different heterogenity uh in caregiving and now I want to highlight some future Pathways so what I wanted to highlight and I think that becomes clear is that longitudinal data is what you need to really take into
Account both the life course and also how diverse caregiving can be so with our long data I think it can be there very difficult I mean there are other options um but I would say that might be the gold standard and we have to support that yeah but longital data is of course
Very expensive there’s a lot of attrition people are dropping out and uh not every country has two very good longitudinal data sets that you can draw on like the UK like the tools that I was showing you so so far it’s really un some countries have a lot of really nice
Analyses we know from the US from Canada Germany Netherlands UK so there is a lot of really good research coming out sometimes also from Australia but we only know about single cases so we don’t know about the broader Spectrum so also it’s very eurocentric
So we it’s Global aging it’s not a not a western or european uh issue it’s really Global issue issue and as well if we also want to into take into account that maybe there’s some migration because of caregiving going on we really need this Global Perspective so I think that there
Should be really one pathway if I go it or or you go it someone would should really look into how can we look from a Global Perspective and but of course if you look at single country context studies there you have the advantage that you of course have one policy context and one
Norm cont context and but it would be also nice to to see how these different kind of contexts work differ and then we can learn also way more about policy um but there’s of course not that much country comparative data available share is one exception the survey of Aging
Retirement yeah I think everyone in here knows probably all everyone online knows what share is but of course they are the disadvantage is it’s 5050 plus so we miss the very early caregivers and of course if you want to are interested like me in employment then sometimes the
Employment is already at the end and not at the beginning where it might be more influential um so I would say the two fuge pathways are first we should solve the terminology issue because it’s just also annoying I think for everyone to have to change their their terms and
Have the same Contex with three different names in their in their publication list also just to to search research it’s quite gets quite complicated and it would be amazing if there’s a big comparative longitudal data set also globally there is the the global aging data it’s from the global
Gateway to aging so they already try to really get different kind of longitudinal data together and this is already amazing start so this is something that should be explored way more and then it would be of course great to also include policy and my my supervisor also with
Ken cler and um a German and another Dutch scholar they were already looking into what can you do what is important to look at and I think they were working on that like four or five years and they came up with these uh five different uh policy measures that you could use to
Compare also globally it’s a caregiving support um if there is some if there’s cash benefits for the caregiver if there have cash benefits for the care user and how many beds there are on in the long-term care and also how many formal care or paid care workers are there that
Could take care as an alternative to Family Care and by that I leave it because I as I said before of course I don’t want to you away all my my ideas but you also see these are quite challenging in itself and you would probably need forever to have really
Good longitudinal data set with all these policies um included so it’s also an invitation for collaboration because it takes time and money to do so but I’m really looking forward to your comments and questions