Karl Ulrich Mayer: A Lifecourse Observatory – no fantasy!

Download this episode
In Episode 5 of the DIAL Podcast, Professor Karl Ulrich Mayer of Yale University and the Max Planck Institute of Human Development discusses life course research, longitudinal studies and how they can help develop develop effective social policy. He also discusses what he calls his “just one wish data set” and why he believes we are close to having a Lifecourse Observatory.

Karl Ulrich is a keynote speaker at the DIAL Mid-Term Conference 2019.

Useful links


In Episode 5 of the DIAL Podcast, Professor Karl Ulrich Mayer of Yale University and the Max Planck Institute of Human Development discusses life course research, longitudinal studies and how they can help develop effective social policy. He also discusses what he calls his “just one wish data set” and why he believes we are close to having a Lifecourse Observatory.

Karl Ulrich is a keynote speaker at the DIAL Mid-Term Conference 2019.

Christine Garrington  0:00  

Welcome to DIAL a podcast where we tune in to evidence on inequality over the lifecourse. In today’s episode, we’re discussing how lifecourse research and longitudinal studies are providing a better understanding of social change and the contribution they make to the development of effective social policy. Our guest is Professor Karl Ulrich Mayer from Yale University at the Max Planck Institute of Human Development, he starts by explaining what we mean when we talk about a lifecourse approach to research.

Karl Ulrich Mayer  0:28  

The idea is that we look at a considerable stretch of lifetime, not just small periods that we are interesting in the time being of events and transitions. And we assume that the lifecourse is a kind of endogenous causal system that in a sense, early conditions, have immediate and long term consequences. This causality is brought about, on the one hand, of course, by the individual actor, but also and more interesting for sociologists is by certain institutional structures. 

Christine Garrington 1:14

When it comes to our day-to-day lives, what sorts of things might we be thinking about when it comes to talking about the lifecourse?

Karl Ulrich Mayer  1:21  

We usually think of the interrelationship between education, training, work life and retirement. So work would be the overall dimension here, the other dimension is family and social demographic kind of life cycle. Of course, people also look at this in relationship to health and other aspects of wellbeing. And there is one another property which should be taken into consideration here, that is the idea of linked lives that we look at individual person. So it’s a kind of a micro perspective, persons in their social context. But also, of course, we look at their relationship to other persons, we proceed from the individual level, but then we also in a way, are partially interested in the sub individual; biological, psychological process on the one hand, in other ways, it is a super individual institutions, micro kind of level, we make observations on the individual level, but we look at it for whole populations.

Christine Garrington  2:36  

Now a number of countries over the last few decades have been collecting what we talk about as longitudinal data following the lives of those same people that you’ve been talking about their households and their relations over time, and then be making that data available to researchers like you. What difference is it making to our ability to better understand some of the pressing social and economic issues of our time? Would you say?

Karl Ulrich Mayer  3:04  

They all started with the British Cohort Studies in 1946 was the first one following infants through their lives. The other addition is that addition of household surveys. In Italy originated by economists like the Panel Study of Income Dynamics 1964, in the United States, also socio economic panel in Germany or Understanding the Society in Britain. And then you have longitudinal studies, which at least initially concentrate on certain life phases, you have the National Educational Panel with the cohort study in Germany, for instance, and you have a Share Study of Retirement for a number of, also in the US, in a number of European countries, these kinds of studies. Now, in a way form what we might call a gold standard of empirical social research.

Christine Garrington  4:08  

And why are they so important for things like social policy?

Karl Ulrich Mayer  4:12  

It allow us to first of all look at the impact of early conditions in life on later outcomes and if such relationship strong relationship exists, then of course, it becomes very important to look at what social policy interventions at which time of life, what kind of effect they have in later life. Let me give you just one example. We now have a good number of kindergarten studies or longitudinal studies which originated with a kindergarten sample. And then we can look at what are the cognitive benefits and social competence benefits of children who go to kindergarten and children who do not go to kindergarten, in how long these effects are actually noticeable even longer in life. There has been a recent study where people look at what kind of impact the working life has on working after retirement, these researchers found that people work part-time and self-employed tend to be working much longer than other people, which conditions at which times in life are most important. How consequential are they? And what difference does it make if social policy intervenes as these kind of periods?

Christine Garrington  5:43  

So what would you say are the benefits of collecting data in this way? What does it make possible that isn’t possible when we simply collect a snapshot of people’s lives and how their lives are going and moment in time as many other surveys and studies do?

Karl Ulrich Mayer  5:59  

First of all, in cross sector service, you cannot differentiate with this anything that comes from that you are in a diet or old age, or whether it’s a consequence of that they were born in a given period and lived your life in, in a given historical period. Many people in many countries talk about the shrinking middle class. So they would look at some definition of the middle class in terms of income categories. And then look at a later cross section, this cross sectional data some years later, say look, whether the middle class so defined as expenses or income. And now many times in, many countries they shrunk. Meaning that part of the people who used to be in the income middle are now in lower income or in in higher income. What we don’t know from perceptions is whether this is a change in composition, that in a way new migrants came into this country, or whether this is due to actually income mobility that people were either downwardly or upwardly mobile. So this is just one example of what you can do with longitudinal studies. What you cannot do is perception study, I give you another example of causal studies, this perceptual data, you never know, they might always be other factors responsible for these outcomes. Now, if you take longitudinal studies, the individual which you observe over time, in many of it’s his or her characteristics actually stays the same. And you can control the one factor or the one condition which changes over time. So in terms of man experimental causes, studies, longitudinal studies are very important.

Christine Garrington  7:57  

That’s fascinating. Now, you’re particularly passionate about how using this data can help to address social inequalities that can occur even before we’re born. I wonder if you can give us an example of how research can help or help in this way? 

Karl Ulrich Mayer  8:13  

Clearly, social inequalities unfold over time, all the outcomes we experience in our lives with ourselves with our families, are the outcomes of our prior kind of histories. Jim Heckman, the Nobel Laureate in economics has become very famous with a whole stretch of studies and secondary analysis, where he shows that cognitive development and the development of non-cognitive abilities things like self efficacy, are actually being developed very early in life, basically, during the kindergarten stage, and then it becomes much more difficult to change the older children now. So he actually produces some, some graphs, where you can see that it might take 5 or 10 or 15, for higher investments to change competencies and behaviours in adolescence, in comparison, with early childhood. Another very important example is that historical periods when people achieve, have to achieve their education and when they have to achieve labour market entry. And these might be historical conditions both in terms of overall societal development, but or they might be extraordinary conditions in terms of business cycle, or war or things like this.

Christine Garrington  9:50  

And that’s something that you yourself have looked at. Can you tell us a bit about that?

Karl Ulrich Mayer  9:54  

We looked at the consequences of war on the lives of Men. And our assumption was that having served many years in service in war. And being then also, quite a number of years prisoners of war would have said the biggest impact on civil security is of these men in Germany. In fact, that was not the case. They actually reintegrated into occupation lives quite well, because they could rely on their initial occupational training and experiences. The men who were with most was the one who had to make the transition from schooling to an apprenticeship in at the end of World War Two, or it’s the very beginning of, of the post war period. So there are many examples where there are decisive conditions, at the time when you have to either get your educational qualifications, or to get into the labour market. And for these cohorts, we could show that they never made it. Even during the time of the German economic miracle in the 50s and the 60s, where they had all the opportunities, they never really get out. A similar situation, which applies in many countries is the difference between baby boom cohorts and small cohorts. For baby boom cohorts, we could show that for them, it was much more difficult to get good labour market chances. But here, we found that actually, they could be able to make it up and partially they were able to make it up because they had a fairly good education credentials.

Christine Garrington  11:46  

I want to turn now to some of your most recent research, where you’ve tried to unravel the impact of societal change in West Germany over a period of more than 60 years, on educational outcomes for people tell us what you did there.

Karl Ulrich Mayer  12:00  

We looked at people born between 1919 and 1986. And looked at their education trajectories. We then looked at it on the one hand individual level factors like social class, and we looked at socio economic conditions. And then we looked at what is the impact of these micro conditions on the educational outcomes?

Christine Garrington  12:29  

What did you anticipate finding? And what did you actually find?

Karl Ulrich Mayer  12:34  

Our general assumption was that people had their increase education, just in order to keep up the status of the class of their parents, we found that we first had a worsening of social inequality of opportunity just for these post-war cohorts. And then before the cohorts born up to 1955-60, there was a markable increase in opportunities. And after that, it’s more or less stagnating or oscillating. So that’s our general outcome in terms of inequality of opportunity. And then we looked at how would macro conditions actually impact on these outcomes. And one of the mechanisms in which we tested and proved was that actually educational policy is a sense of provision of institutions, number of schools, number of universities, actually has a market impact on these kind of outcomes.

Christine Garrington  13:43  

Now, you talked a little bit about what you found from the study, but I wonder if you can tell us how it sort of helped to provide a better understanding of change? In that changing social, political and cultural context over those 16 years?

Karl Ulrich Mayer  13:57 

These were birth cohorts, which really experience quite different kind of conditions. Now the people born around 1920 well many of them were even drafted into the war. Then we had the people who were most hit in their educational and occupational careers were born around 1930. And then you had people who benefited a lot from the social reforms and educated reforms of the 50s and 60s, and increasing economic efforts. And then you had some kind of people born around 55-60, who did much less well, and so period of the 70s and 80s. And then you have the baby boom groups, which again, had some initial difficulties. So you have different cohort conditions where it’s interesting to see whether the these are lasting effects, which we found for some of these first cohorts, and cohorts which actually could make it up across their lifetime. Then there is, of course, a question of whether certain reforms, social policies have some impacts. Some recent study is in Germany since 1996, early 2000s, we massively expanded kindergarten, right for kids three years in older to have a kindergarten place. And after 2003 years, there is also corresponding route for 1, 2 and 3 year olds. And the question is, what kind of consequences would this have for employment and careers of mothers? And actually, my current show that it had remarkable impacts, both in terms of the labour market participation also in terms of the income they earn, and also the relative quality of earnings within the family.

Christine Garrington  16:06  

Fantastic to see that journey from research into policy into real benefits for people’s lives. Now, I want to move on and ask you that, in a recent article, you talked about your just one wish data set and design and I wonder if you can tell us what that looks like and why it features so largely for you?

Karl Ulrich Mayer  16:26  

It used to be that we had studies for small periods of life for a given country now we have more and more longitudinal studies, which address quite long part of life. And we have this across many historically, across many of those cohorts. And for many countries, so in a way you can imagine a kind of data cube or a kind of data set involving observations of individuals of representative populations, across their whole life, and for many birth cohorts across historical time, across many countries. And this would allow us to then ask such questions is there are now real big claims about changes in working lives, that working lives are less stable, are more complex, more disruptive. That’s one of the major social science claims and similar claims you have in regard to family life. So we can piece together observations over individual lives. So relatively long stretches, so we can look at what factors lead to situation when people retire, what kind of economic situation and social situation in their old age, etc. So that’s one thing, if we now already now combine my job is to study this initial educational panel. Now close have people born over a stretch of 100 years. And then data across countries, which now many studies have shown that actually, differences between countries with for instance, between central Saxon countries, continental countries like France and Germany and Scandinavian countries, it actually that makes a big difference in the kind of ways people live their lives.

Christine Garrington  18:35  

So this just one wish data set? How realistic is it? Do you think? And what are the challenges that lie ahead in terms of making it a sort of a reality as it were?

Karl Ulrich Mayer  18:46  

I think it already comes about because, of course, it’s already more or less realized if you have administrative data, which is the data like in Scandinavia, you already can do it. And in other countries, for instance, with the maturing of the household panels, you get more and more of such data. So it is not a fantasy.

Christine Garrington  19:12  

Karl Ulrich Mayer is one of the keynote speakers at the Dynamics of Inequality Across the Lifecourse midterm in June 2019. More information is available on the DIAL website at www.dynamicsofinequality.org. Thanks for listening to this episode of our podcast, which is presented and produced by Chris Garrington and edited by Elina Kilpi-Jakonen.