Do jobs run in families?

How much of our choice of profession depends on who our parents are? Parents pass on their genes, set an example, provide opportunities, and give advice to either aim for or steer clear of their own lines of work. In the end, do their children end up in the same type of job? Do siblings choose the same occupation? And is this more or less true for different professions?

To study these questions, we analyzed in aggregate two related sets of de-identified Facebook data: one a sample of siblings' choices of profession, and the other of parent-child choices. The sample included those pairs of individuals in English-speaking locales who specified a sibling or parent-child relationship on Facebook, along with filling in their occupations. The occupations were mapped to major occupation categories 1. The military occupation category is over-represented because it is mapped based on both employer and stated occupation and past military service, whereas other job categories were mapped based on stated occupation only. Since the data excludes those not specifying an occupation on Facebook, it may not be representative of the population overall, but is interesting to study nonetheless.

Do children share their parents' occupations? Mouse over the interactive visualizations to see what professions children choose.

The interactive visualization shows that the probability of a child's occupation falling into any given category does vary by occupation. Using a sample of 5.6M parent-child pairs from English-speaking locales both listing an occupation, we first calculated the probability of a child having an occupation, given their father's occupation, e.g. a lawyer-father having a doctor-son (5%). We also then calculated how elevated this probability is relative to the overall proportion of doctors among sons. In this case a son of someone in the legal profession is 4.6 times as likely to practice medicine than sons in general.

In this network visualization, each node represents an occupation-gender pair, e.g. women who are scientists. Labels are abbreviated. Edges denote how much more likely a child of a parent in one profession is to choose another profession vs. someone from the general population. Only edges with a multiplier of at least 2.5 are shown.

Another way of looking at connections between parent-child occupations is through a network visualization, shown above. For example, fathers in the military are more likely than average to have a son in protective service, and a line is drawn showing that relationship between parent and child occupations. By laying out the network of such connections, using a force-directed network layout algorithm to place occupations with unusually high inter-generational links closer together, we see professions clustering slightly into ones requiring a secondary degree (college and higher) or not.

Even though relatively speaking, a child may be much more likely to follow in his or her parents' footsteps, the absolute percentage may still be quite low. A son who has a father in the military is 5 times more likely to enter the military, but just 1 in 4 sons of a military professional does so. For fathers in the dataset who work in farming, fishing and forestry, only 3% of their sons stay in the profession, but this probability is 7.6 times the overall rate. 20% of daughters of mothers who work in office and administrative support choose the same career, but this is only 2x the usual rate. On the other hand 8.5% of daughters of mothers in nursing also choose a career in nursing, and this is 3.75x the overall rate. We also see substantial cross-gender occupation “inheritance”, e.g. scientist fathers have scientist daughters at 3.9 the overall rate, while mothers working in law have sons choosing a legal profession at 6.6 times the overall rate. Note that a negative relationship, where a child is less likely to enter a profession due to their family background, is typically very small. For example for lawyer-fathers, the probability of their sons entering construction or maintenance or repair is about 85% of the overall likelihood.

Next we look at whether siblings choose the same occupation. Siblings not only share the same parent, but sometimes, as in the case of identical twins, they share the same genes (identical twins are for the most part* genetically identical, fraternal twins are genetically as different as siblings are). Whether identical or fraternal 2, twins are also more likely to be raised in a similar environment — parenting styles may differ as a parents add more children to their brood, but twins will likely be exposed to a similar parenting style.

For the analysis of sibling occupations we considered a sample of 2.37M same-gender siblings in the US who had filled in their occupations in their profiles. We compared the rates at which same-gender twins shared an occupation against same-gender siblings no more than two years apart, and then again any two people of same gender in our sample who are no more than two years apart in age.

15% of siblings share an occupation, which is higher than the 8.6% rate for any two same-gender, same-age individuals in the population. Twins' tendency to choose the same occupation, at 24.7%, is even more striking.

The interactive plot above shows a breakdown of this effect by occupation and gender (purple for female and blue for male pairs). For each bar triplet, the most opaque bar is the baseline expected occupation overlap. The lighter bar is the observed overlap between non-twin same-gender siblings. The lightest bar is the observed overlap among same-gender twins. By default the occupations are ordered from top-down according to the observed twin overlap for each gender separately. Clicking on 'Siblings' or 'Population' changes the ordering.

To conclude, we see that people within a family are proportionally more likely to eventually also choose the same occupation, and this is especially true of twins. However, in absolute terms the vast majority of kids strike their own path and choose a profession different than that of their parents or their siblings.

  1. See for how these are defined.
  2. Due to increased use of fertility treatments, which prompts the release and/or implantation of multiple eggs, the number of fraternal twins born has risen from being roughly equal to the number of identical twin births to outnumbering them 2:1.

An earlier version of this post had an interactive visualization of father-daughter pairs instead of mother-daughter pairs.

Related Blog Posts

KDD Brings Together the top Data Science Minds
by Kelly BerschauerAug 15, 2016
Cat People, Dog People
by Moira Burke, Lada Adamic, Amaç Herdağdelen, Dirk NeumannAug 7, 2016
Imperfect Treatment Assignments
by Dominic Coey, Michael BaileyJun 21, 2016
Facebook Researchers Focusing on the Future of Human Computer Interaction at CHI 2016
by Mike Massimi, Vanessa Callison-Burch, Alex Dow, Lada Adamic, Vasanth Rajendran, Margaret Gould Stewart, Sean KellerMay 6, 2016

Related Publications

Compressing Graphs and Indexes with Recursive Graph Bisection
by Laxman Dhulipala, Igor Kabiljo, Brian Karrer, Giuseppe Ottaviano, Sergey Pupyrev, Alon ShalitaKDDAug 13, 2016
Towards Optimal Cardinality Estimation of Unions and Intersections with Sketches
by Daniel TingACM Conference on Knowledge Discovery and Data Mining (KDD 2016)Aug 12, 2016
The Relationship between Facebook Use and Well-Being Depends on Communication Type and Tie Strength
by Moira Burke, Robert KrautJournal of Computer-Mediated Communication (JCMC)Jul 27, 2016
The Social Ties of Immigrant Communities in the United States
by Amaç Herdağdelen, Bogdan State, Lada Adamic, Winter MasonWEBSCIMay 22, 2016

Join Us

Do you want to help more than a billion people all over the world connect and share?

View Open Positions


Learn about our open source tools and technologies, our challenging scaling experiences, and more.

Go to Facebook Code