The End of Average

How We Succeed in a World That Values Sameness

By: Todd Rose

Published: 2016

Read: 2020


Minimizing errors. Businesses, schools and governments assess people based on their relationship to the average. Using the group average to determine the norm (type), they offer opportunities to individuals based on their deviation from that group average (rank).

Maximizing efficiency. Systems designed around type and rank have historically worked well and have contributed to a stable and prosperous democracy by leveling the playing field (less discrimination, cronyism and nepotism -> equal opportunities) and driving higher efficiency (standardization -> lower prices).

The error of averages. But problems arise when we rely on the group average (tests and scores) to make predictions about individuals. We assume that groups are ergodic: we assume that what we know about the group is true for the individual members of that group. Because of this, we are unable to identify, understand and appreciate individual patterns (they get averaged out).

Biases. To make matters worse, when we “score” individuals against a group average, we often focus on only one dimension, we assume that this dimension is stable over time regardless of context and we believe that the average path to success in this dimension is the best (and only) path.

Negative outcomes. Rigid social systems designed along these lines lead to excess standardization (only one right way to get things done), narrow expectations about what it takes to succeed (in school, career and life) and limitations of individual potential (conformity: people strive to be like everyone else, only better).

So what to do? This book outlines three principles to better understand and harness individuality.

The jaggedness principle: we tend to believe that things we care about (talent, intelligence) can be measured in a single score. However, they are usually multi-dimensional and those dimensions are often only weakly correlated. So when the stakes are sufficiently high, don’t rely on one single score (jeans: one size measure is okay; wedding dress: custom fit).

The context principle: we tend to believe that most personality traits are stable and behavior is predictable regardless of context. However, traits are only stable within certain situations. So if you want to be more productive, figure out the environments where you do well (and where you don’t do well). And when you don’t like someone else’s behavior, understand that this is only one situation that doesn’t define the other person.

The pathways principle: we tend to believe that the average pathway is the best or even only pathway. However, there are often many ways to get to the same outcome. Pacing and sequencing differ for everyone. So figure out what path works best for you, based on your (jagged) profile and context.

Applying these principles in businesses and schools would allow for a better fit between opportunities and individuals (as opposed to often ineffective concepts of equal access) and offer individuals a better chance to succeed on their own terms (as opposed to conformity and climbing the standard ladder).

Worth Reading

The most enjoyable and convincing section of the book is the chapter on what “the science of individuality” means for higher education. Where “Prepared” aims to provide a road map for improving high-school education and helping kids to better prepare for what comes next, the education chapter in this book does the same for post-high-school education. It diagnoses similar issues as “Prepared” (such as the push for students to be the same as everyone else, just better) and proposes similar solutions (create a system that allows students to figure out what they like, what they are good at, and what is the best way to pursue these interests). Its recommendations (credentials and competencies instead of diplomas and grades, etc.) should help close the gap between employers’ needs and students’ capabilities and experiences, as well as help students make better and more informed career choices. In contrast, the business section of the book feels a bit thin and doesn’t get too far beyond providing a handful of real-life case studies.

Given the message of the book (individuality matters), it makes sense and is an interesting touch that the central narrative is woven around the historic contributions of specific individuals (Quetelet, Galton, Taylor, etc. – see below).

Perhaps at times the pendulum swings a little too hard in the direction of “anything goes”. For instance, the pathway principle could be developed more using ergodic concepts like path dependency and individual versus group growth rates (see “Ergodicity”). Yes, it’s good to appreciate that there is more than one average pathway and you need to understand which pathway best fits your individual profile and context. But things don’t stop there. Sometimes the group average pathway is actually misleading, incredibly rare and difficult to achieve (when the group average is skewed by a few very large individual outcomes), so the individual pathway is unlikely to look anything like the group average pathway. Or there are many potential pathways, but some choices are very clearly better or worse than others. There is a lot more to say and explore on these topics.

Key Takeaways

  • Analyze then aggregate versus aggregate then analyze.
    • Look for patterns in individuals, as opposed to group patterns.
    • Combine individual patterns, as opposed to combining individuals.
    • Look for collective insights, as opposed to individual predictions.
    • This requires a lot more data.
  • When the stakes are sufficiently high, don’t rely on one single score
  • If you want to be more productive, figure out in what situations you do well (and don’t do well).
  • When you don’t like someone’s behavior, there is more to the other person than just that particular context.
  • Figure out your path (pace, sequence, fit), don’t just climb the ladder .

Key Concepts

Minimizing error: type and rank

  • Two ideas serve as the organizing principles behind education and talent assessment.
    • Type.
    • Rank.
  • Quetelet: the type or the average man.
    • The average is a reliable index of normality.
      • The average is the ideal.
      • The individual is the error.
    • Science: research mostly about determining the average, the type.
      • Put a group of people into an experimental condition.
      • Determine the average response to the condition.
      • Use the average to formulate a general conclusion about all people.
      • Make predictions about any individual by knowing the traits of the average member.
      • Assumes all members of a group act according to a set of shared characteristics.
      • The average person is a prototypical representative of a group.
    • Corresponds to natural urge to stereotype
      • Type-A personalities, leader types, micro-managers, etc.
  • Galton: use deviation from the norm to rank.
    • Individual cases are evaluated against the average.
      • What matters most about an individual is how they compare to the average.
      • Deviation from the average = error.
    • Leads to ranking.
      • Indexing against the average (= the norm).
    • If a person is talented at one thing, likely to be talented at most other things.

Maximizing inefficiency: standardization and scientific management

  • Taylor: takes Quetelet’s average -> standardization, one best way.
    • The system comes first.
      • One best way to standardize the system.
    • Individuality does not matter.
      • Hire average employees to fit the standardized system.
    • Creates management and working class.
      • Planners to figure out the one best way (system designing).
      • Workers to do the actual work (system conforming).
    • Businesses prospered.
      • Increased efficiency -> cheaper products -> higher wages.
    • Education adjusted accordingly.
      • Schools provide a standard education for an average student.
      • Instead of trying to foster greatness.
      • Organize and teach children to become workers that perform tasks in “a perfect way”.
      • Standardize everything around the average.
        • Grouped by age, fixed class hours, etc.
    • Opportunity: each student treated as an average student.
      • Provide each one with the same standardized education.
      • Regardless of background, abilities, or interests.
      • Less nepotism, cronyism.
    • Criticisms.
      • Reduces individuals to the same safe level.
      • Breeds and trains a standardized citizenry.
      • Puts down dissent and originality.
  • Thorndike: takes Galton’s rank -> identify and separate superior students from inferior ones.
    • Schools should sort people according to innate level of talent.
      • Not educate all students to the same level.
      • Efficiently rank students in order to assign them to their proper place in life.
    • Quality is more important than equality.
      • Fast learners = success in school = success in life.
    • Educational resources should be allocated accordingly.
      • Special needs, honors students etc.
    • Use of grades and standardized tests for sorting.
      • Standardized tests.

The error of averages

  • Some amount of error always occurs in any one test session.
    • Distraction, etc.
  • Repeated testing is required to determine the “true score”.
    • Apply the same test to the same person many times.
    • Average of multiple tests = true score.
  • But humans learn.
    • So you can’t administer independent test results.
  • Instead of testing one person many times, test many persons one time.
    • Substitute a group distribution of scores for an individual’s distribution of scores.
  • This assumption serves as the basis for research in many fields of science.
    • The average of 1,000 people’s scores = average of 1 person’s 1,000 scores.
  • Ergodic theory.
    • Investigates the relationship between groups and individuals (in math).
    • Use information about a group to draw conclusions about members of the group when:
      • Every member of the group is identical.
      • Every member of the group will remain the same in the future.
    • When these conditions are met, the group is “ergodic”.
  • The “ergodic switch”:
    • Substituting knowledge about the group for knowledge about the individual.
      • In a group of people, fast typing is correlated with less errors (expert typists).
      • That doesn’t mean: you should type faster, because you’ll make less errors.
  • But if you can’t use averages, then what?
    • Averages provide a stable, transparent, and streamlined process for making decisions quickly.
  • Science of individuality.
    • Individuality matters.
    • Human qualities that matter most can’t be reduced to a single score.
    • Dynamic systems versus statistics:
      • Changing, non-linear, dynamic values.
    • Analyze then aggregate versus aggregate then analyze:
      • Average: combine many people, look for patterns in the group, use patterns to analyze and model individuals.
      • Individuals: look for patterns in individuals, combine individual patterns into collective insights.
      • This requires a lot more data.
        • Data we didn’t have until recently.


  • Mental barrier: one-dimensional thinking.
    • Tendency to use a one-dimensional scale to think about complex human traits.
      • Size, intelligence, character, or talent.
  • Most of the time the trade-off is productive.
    • When efficiency matters (mass production).
    • When it’s not costly to make individual mistakes (because “on average” you’ll do better than random choice).
  • When stakes are higher, one-dimensional thinking may become too costly.
    • Mistakes are more costly (custom fit a wedding dress versus one size for a pair of jeans).
  • A quality is “jagged” if it meets two criteria.
    • Consists of multiple dimensions.
    • These dimensions are weakly related (weak correlations).
      • If correlations are strong, cost of focusing on one dimension is low.
  • Most human qualities that we care about are jagged.
    • Talent, intelligence, character, etc.


  • Mental barrier: essentialist thinking.
    • Traits and behavior are stable and predictable, regardless of context.
  • Traits versus situation:
    • Traits:
      • Behavior is determined by well-defined, durable personality traits.
      • Introversion, extroversion, etc.
    • Situation:
      • The environment drives our behavior more than our traits.
      • Culture and immediate circumstances.
  • Trait theorists on average do a better job of predicting the behavior of any given individual.
    • Personality tests correspond well to the way we think about ourselves and others.
    • Essentialist thinking: people are wired a certain way regardless of external conditions.
  • But when it comes to predicting actual individual behavior, traits do a poor job.
    • Weak correlations between personality and specific behavior.
  • Ignores the context.
    • Our traits are consistent within a certain context.
    • Traits and situation interact.
  • Context principle:
    • Individual behavior cannot be explained or predicted apart from a particular situation.
    • The influence of a situation cannot be specified without reference to the individual experiencing it.
    • Behavior emerges out of the unique interaction of traits and context.
  • If-then signatures:
    • If [context] then [behavior].
    • Implication: moral behavior is specific and conditioned to a degree by external situation.
      • For instance, self-control is not an essential trait, it is contextual (folly of marshmallow study).
  • Difficult to accept traits change with circumstances.
    • We tend to believe our own traits are stable.
      • Our brains automatically adjust for context.
      • We compare behavior in one situation to our behavior in a similar situation.
    • We tend to believe other people’s traits are stable.
      • We tend to interact with others in a narrow range of contexts.
      • We are often a key part of the other’s context.
      • We don’t see the other’s behavior under a variety of contexts.
  • Context principle.
    • Helps us to figure out the environment where we do well and don’t do well.
      • Identify situational factors that may lead us to productive or unproductive behavior.
    • Deal more productively with negative behavior of others.
      • Not: why do they behave in that way?
      • But: Why are they behaving that way in that context?
      • There is more to the other person than just that particular context.


  • Mental barrier: normative thinking.
    • The right pathway is the one followed by the average person.
  • We believe that average patterns of behavior are proof that something is innate and universal.
    • But, our biology does not compel us to follow a predetermined blueprint.
  • Pathways principle:
    • There is not a single, normal pathway for any type of human development.
      • Biological, mental, moral or professional.
    • There are many, equally valid ways to reach the same outcome (equifinality).
    • The particular pathway that is optimal for you depends on your own individuality.
  • Individuals vary naturally in their pace of progress and the sequences they take.
    • Speed does not equal ability (faster is not better is not smarter).
    • There are no universally fast or slow learners.
  • Pace: self-paced learning.
    • Different students learn at different paces at different times for different material.
  • Sequences: web of development.
    • In personal development, each step opens up a range of new possibilities.
    • No fixed sequences or one way.
    • No predestined climbing up a ladder rung by rung.
  • How to determine the “right path”?
    • Understanding our jagged profile and if-then signatures.
    • Judging how the path fits our individuality.

Application to businesses

  • Abandoning one-dimensional thinking, essentialist thinking, and normative thinking.
  • Create highly engaged and competitive workforces.
  • Embrace individual freedom, initiative and responsibility,

Application to education

  • Higher education = career preparation
    • Prepare students for their self-chosen careers at an affordable price.
  • But system is not working.
    • Many graduates can’t find a job in their field.
    • Many employers think graduates are not equipped for the job.
  • System focused on one-dimensional ranking, conformity.
    • Compels every student to do exactly the same things the average student does.
    • Be the same as everyone else, only better.
    • Hiding individuality, rather than developing it.
    • Quality keeps going down, costs keep going up.
  • Individualized system of education.
    • Grant credentials, not diplomas.
      • Emphasizes awarding credit for the smallest meaningful unit of learning.
      • Flexible and detailed certification of skills, abilities, and knowledge.
      • Can be combined and stacked.
    • Replace grades with competencies.
      • Grades:
        • One-dimensional and poorly reflect complex skills.
        • Difficult to interpret and translate into skills.
      • Competencies:
        • Demonstrated ability (pass or fail).
        • Institution-agnostic.
        • Professionally aligned.
      • Let students determine their pathway.
        • More options.
        • Institution independent.
        • Adjust as needed.
        • Fewer penalties for switching, experimenting.
  • Aligned with three principles of individuality:
    • Jaggedness:
      • Allows students to figure out what they like, what they are good at, and what is the best way to pursue these interests.
    • Context:
      • Evaluates students’ competency in a context as close as possible to the professional environment where they will actually perform.
    • Pathway:
      • Allows each student to learn at their own pace, and follow a sequence that is right for them.
  • Solves issues:
    • High cost:
      • Pay for what you want (credentials) and nothing more.
      • Differentiated offerings (only pay if you complete, etc.)
    • Better match:
      • Value and availability of credentials adjusts faster to market demands.
      • Content of credentials better aligned with market demands.

Better outcomes

  • Opportunities: pursue equal fit rather than equal access.
    • Equal access: best way address fairness in a standardized world.
      • Ensuring that everyone has access to the same experiences.
      • Addresses discrimination, cronyism, nepotism.
      • Maximizes individual opportunity “on average”.
    • But standardized system is not automatically a good fit for everyone.
      • If there is no average person, there can’t be equal opportunity “on average”.
    • Only equal fit creates equal opportunity.
      • Best fit creates new opportunities.
  • Priorities, meaning of success: seek personal fulfillment rather than beating others.
    • Personal fulfillment.
      • Opportunity to live your life to its fullest potential.
      • Being appreciated as an individual.
    • Be better than people around you.
      • Conformity: do the same as everyone else, just better.
      • Climbing the ladder: artificial, arbitrary, meaningless rungs.

Leave a Reply