“Mobility increased” is not yet a complete result. It may mean that more adults earn more than their parents, that origin predicts destination less strongly, or that more people changed occupational categories. Those statements describe different outcomes and can move in different directions.
The AAMC content outline includes social stratification and patterns of mobility. The original data below are deliberately small so the comparison and denominator remain visible. They describe fictional populations, not estimates of any country or social group.
Identify the resource and the time comparison
Income is a flow received over a period. Wealth concerns accumulated assets minus liabilities. A person can have high current earnings and little net wealth, or modest earnings and substantial assets. Education, occupation, prestige and power provide other dimensions of social position. They should not be silently substituted for one another. OpenStax introduces these stratification distinctions.
Intergenerational mobility compares generations, such as adults with their parents. Intragenerational mobility tracks change within a person’s life. Moving to a different occupation at similar status can represent horizontal mobility; upward or downward movement in status is vertical mobility. Structural changes can alter the number and types of positions available. OpenStax discusses these mobility concepts.
For a passage, translate the term into a specific comparison: whose resource, measured at what age, over what period, against whom? An income at age 23 is not directly comparable to a parent’s peak-career income without accounting for the life-stage difference.
Worked example: everyone gains, yet rank persists
Suppose four families constitute an entire fictional cohort. Compare each parent’s annual family income at age 35 with their adult child’s family income at age 35. Use the same family definition and inflation-adjusted monetary units in both generations.
| Family pair | Parent income | Adult child income | Parent rank | Child rank |
|---|---|---|---|---|
| A | 20,000 | 30,000 | 1, lowest | 1, lowest |
| B | 40,000 | 60,000 | 2 | 2 |
| C | 60,000 | 90,000 | 3 | 3 |
| D | 80,000 | 120,000 | 4, highest | 4, highest |
Every child’s real income exceeds the corresponding parent’s. Under the stated definition, absolute upward income mobility is 4/4 = 100%.
Yet the rank ordering is unchanged. A remains lowest and D highest; no pair moves to a different rank. This constructed example has perfect rank persistence despite universal absolute gains. Relative mobility can be studied through the association between origins and destinations, rather than only whether money amounts increased.
Chetty and colleagues’ primary study of absolute income mobility distinguishes a child-versus-parent real-income measure from the joint distribution of their income ranks. The table here illustrates that measurement distinction; it does not reproduce the study’s data or estimate its findings.
The example also does not show that every aspect of well-being improved by 50%. Household needs, time worked, public services and security are not represented. Nor does unchanged rank by itself prove a particular causal mechanism. A descriptive pattern is a starting point for explanation.
A transition table needs a named denominator
A separate fictional study classifies parents and adult children into lower and higher income bands. Cutoffs are defined using external reference populations for each generation. They are not medians recalculated within this study sample, so the sample groups need not contain equal numbers.
| Parent band | Child lower band | Child higher band | Total |
|---|---|---|---|
| Lower | 60 | 20 | 80 |
| Higher | 30 | 90 | 120 |
| Total | 90 | 110 | 200 |
Among participants from the lower parent band, the fraction reaching the higher child band is 20/80 = 25%.
Among participants in the higher child band, the fraction originating in the lower parent band is 20/110 ≈ 18.2%.
These are different conditional probabilities. The first asks about destinations given an origin; the second asks about origins given a destination. Both use the same cell, but changing the denominator changes the question.
Meanwhile, 20/200 = 10% describes the fraction of the entire sample in that upward-transition cell. It is not the upward-transition percentage among lower-origin participants. Report the group after the word “among” before calculating.
This table concerns crossing band boundaries. It cannot reveal every within-band income gain or loss. A person can experience a substantial real-income increase while remaining below the cutoff, and another can cross it after a small increase near the boundary.
Selection can change the apparent mobility pattern
Imagine that the transition table includes only adults who remain in their childhood region. If moving away is associated with income opportunities, the observed origin–destination pattern may differ from that of the full original cohort.
The direction of the distortion cannot be inferred simply from the word “migration.” People might leave for high-paying employment, leave after losing work, or move for reasons related to neither. You need evidence about who is excluded and how exclusion relates to both origin and destination.
Similarly, a sample recruited through university alumni records may underrepresent people who never attended that institution. More observations from the same selected source do not automatically repair the selection problem. The target population must match the claim being made.
Transfer exercise: bound the missing outcomes
A longitudinal study begins with 100 adults from a lower-origin group. Researchers successfully follow 60; 18 of those move into the defined higher destination band. The other 40 have unknown outcomes.
What fraction moved upward among the observed participants? 18/60 = 30%.
What can be concluded about the original 100 without assumptions about the missing people? The minimum is 18/100 = 18%, if none of the missing 40 moved upward. The maximum is (18 + 40)/100 = 58%, if all did.
Thus the original-cohort proportion is bounded between 18% and 58%. These are missing-data bounds, not a confidence interval. They describe unresolved outcomes rather than sampling uncertainty.
The answer 30% for the full cohort requires an additional representativeness or missingness assumption. The answer “nothing can be learned” is also too strong: 18 outcomes are known to be upward transitions, and the unknown group has a finite size.
Separate mechanisms from judgments about people
An origin–destination association is compatible with several mechanisms: financial constraints, access to educational opportunities, social networks, discrimination, inheritance or differences in available employment. A study needs an appropriate design to distinguish their contributions. Naming a mechanism is not evidence that it caused the observed pattern.
Do not turn a group average into a claim about every member’s effort, ability or values. Likewise, showing a structural constraint does not imply identical outcomes for everyone facing it. The data’s level of analysis limits the conclusion.
For your next passage, write one sentence defining mobility, one naming the denominator, and one identifying who is missing. Then separate the measured pattern from the proposed explanation. That small discipline makes social-science reasoning more precise without reducing people to a category.
Sources and further reading
- AAMC: What Is on the MCAT Exam? Content Outline
- OpenStax Introduction to Sociology 3e: What Is Social Stratification?
- OpenStax Introduction to Sociology 3e: Social Stratification and Mobility in the United States
- Chetty et al. (2017): The fading American dream: Trends in absolute income mobility since 1940