Overview of Science Friday and WNYC Studios
This episode explores a mathematician’s analysis of more than 35,000 women’s fashion images to test the long-held idea that fashion trends repeat in roughly 20-year cycles. Dr. Emma Zajdela of Princeton University explains how she used mathematical modeling to measure changes in neckline, waistline, and hemline over time, finding real cyclical patterns in fashion history—and a later shift toward faster, more diverse trend cycles.
Key Findings
- Fashion really does cycle, at least in the dataset studied.
- The clearest pattern was a roughly 20-year trend loop in women’s dress styles.
- From the 1920s through the 1980s, shorter dress styles followed a pattern that looked very much like a sine wave.
- Starting in the 1980s, the pattern became more complex, with multiple styles coexisting:
- mini skirts
- mid-length skirts/dresses
- floor-length styles
How the Study Worked
Data sources
The analysis combined two major image sets:
- Commercial Pattern Archive (COPA): sewing pattern archives from 1869 to 2015
- Vogue runway images: used to fill in more recent years and broaden the sample
What was measured
To simplify fashion into something mathematically tractable, the study focused on the vertical axis of dress design:
- Neckline
- Waistline
- Hemline
This gave the researchers a way to quantify style changes over time in a consistent way.
Why Fashion Cycles Repeat
Zajdela’s model is based on optimal distinctiveness, a psychology idea that says people and trends succeed when they are:
- different enough to feel new
- but not so different that they feel weird
The model suggests three forces drive fashion cycles:
- Trends must differ from the recent past
- People want to stand out from others, but only within social limits
- Physical constraints limit how far a garment can go in either direction
Together, those pressures create oscillation: once a trend reaches its limit, it eventually swings back the other way.
Why the Pattern Shifted After the 1980s
The study suggests two major reasons fashion became more varied:
- Trends accelerated: styles changed faster and faster
- Fashion became more diverse: there was less conformity and more acceptable variation in women’s clothing
This helps explain why the older single “cycle” gave way to multiple overlapping style clusters.
Broader Implications
The conversation goes beyond fashion and shows how math can help explain complex human systems.
Complex systems
The episode highlights that individual choices can produce large-scale patterns, such as:
- fashion trends
- bird flocks
- fish schools
- spread of disease
- other social behaviors
The core idea: the whole is greater than the sum of its parts.
Practical insight
The same modeling approach could potentially help researchers understand:
- cultural change
- consumer behavior
- public health dynamics
- other trend-driven social systems
Takeaway on Style
Zajdela’s practical fashion advice is basically:
- aim to be “different, but not too different”
- use accessories to add personality without going into the “weird zone”
Examples she gives include:
- jeans and a white shirt with a bright bucket hat
- a blazer with a vintage pin or bold earrings
Notable Insight
“Innovations need to be different, but not too different.”
That idea captures both the math and the fashion lesson of the episode: trends become successful by balancing novelty with familiarity.