Data Dive · Rung 4 · What Isn't There · 4.3

Earthquakes are declining

Stand in 2022 and it looks true.

Runs in 25 minutes · everything you need is on this page · no network, no accounts, nothing for the students to open

4.3→ next · ← back · esc leaves0 / 4

The claim on the table

“Southern California earthquakes are declining.”

It is 2022. The last few years read 6, 2, 2, and one recent year recorded none at all.

4.3 · Earthquakes are declining→ next · ← back · esc leaves1 / 4

The evidence

M4 and above, per year

Earthquakes of magnitude 4 and above, per yearSouthern California, 1995-2024 - the mean year is 9.1earthquakes recordedyearearthquakes recordedthe years that look like a decline33.5670795796129789867996008017021103504140550690780890911103117126138145156160174183196202212224231524

1995-2024. Marked: 2017, 2021, 2022.

9.1
the mean year
0
in 2017
15
in 2024

Then 2024 returns fifteen.

4.3 · Earthquakes are declining→ next · ← back · esc leaves2 / 4

The trap inside the trap

The obvious noise band is the wrong model

A first-year calculation puts the expected range at roughly 3 to 15, which makes 2021 and 2022 look significant.

Earthquakes are not independent events. They arrive in families — 1999's 67 is one magnitude 7.1 and its aftershocks, not sixty-seven separate facts. When events cluster, the real spread is far wider than that arithmetic predicts, and a band computed the easy way calls ordinary years extraordinary.

Notice2017 is written as 0, not left out. A gap in a file reads as nobody looked; a zero reads as we looked and found none. Those are different claims.
The three questions
  1. Where does this series start and end, and who chose that?
  2. Does the evidence reach the claim?
  3. What would convince you the decline is real -- and can thirty numbers ever do it?
4.3 · Earthquakes are declining→ next · ← back · esc leaves3 / 4

After the dive

The shapes on this page, in the library’s own words

is this real, or noise?

Shuffle the data many times and see whether your result still stands out. If it does not, it was luck.

What it leaves outIt tells you a result is unlikely by chance, not that it matters. A tiny difference can clear the bar and mean nothing.
the bar

Bar length is the value, so it only means anything measured from zero. Start the axis anywhere else and the lengths stop matching the numbers.

What it leaves outOrder is a choice. Sorted by size it is a ranking; sorted by name it is a lookup; and the same bars say different things in each.

library · ladder · dives · the pool · this dive’s worksheet

4.3→ next · ← back · esc leaves4 / 4