21) D

Out of all the options, D is the correct one.

A – Wrong

In 1990s, total number of crashes = 235

Crashes due to pilot error = 140.

Proportion = ^{140}/_{235} ≈ 0.6

You can see that in any decade, proportion of crashes due to pilot error is approximately 0.6

So proportion of crashes due to pilot error was **not** unusually high in the 1990s.

B – Wrong

In 2010s, total number of crashes = 66

Crashes due to mechanical error = 11.

Proportion = ^{11}/_{66} ≈ 0.17

You can see that in any decade, proportion of crashes due to pilot error is approximately 0.17

So proportion of crashes due to mechanical error was **not** unusually low in the 2010s.

C – Wrong

Proportion of crashes due to bad weather in 1960s = 14/248 ≈ 0.06

Proportion of crashes due to bad weather in 1970s = 13/238 ≈ 0.06

Proportion of crashes due to bad weather in 1980s = 11/205 ≈ 0.05

Proportion of crashes due to bad weather in 1990s = 13/235 ≈ 0.06

Proportion of crashes due to bad weather in 2000s = 7/112 ≈ 0.06

Proportion of crashes due to bad weather in 2010s = 5/66 ≈ 0.08

So proportion of crashes due to bad weather did not steadily decrease over the period.

D – Correct

Proportion of crashes due to sabotage in 1960s = 12/248 ≈ 0.05

Proportion of crashes due to sabotage in 1970s = 25/238 ≈ 0.10

Proportion of crashes due to sabotage in 1980s = 23/203 ≈ 0.11

Proportion of crashes due to sabotage in 1990s = 19/235 ≈ 0.08

Proportion of crashes due to sabotage in 2000s = 10/112 ≈ 0.09

Proportion of crashes due to sabotage in 2010s = 6/66 ≈ 0.09

So proportion of crashes due to sabotage was unusually low in the 1960s.

E – Wrong

If you work out the estimate proportions, you will realise that the proportions of crashes due to other causes increased steadily over the period.