According to this chart [1] showing the inflation adjusted minimum wage over the years, it should be closer to $11 to match the dollar value it typically had most of the years it was in place.

[1] https://upload.wikimedia.org/wikipedia/commons/thumb/3/30/Ti...

A maximum is not necessarily - or even normally - an outlier.

Sure, but we’re not making general statements. This one is.

Sure, in general, that would be something to take into consideration - but in this specific case, the maximum is not an outlier.

Apparently (based on the comment I was replying to), labcomputer does seem to think the maximum is an outlier, and that this false generality applies here.

> Apparently (based on the comment I was replying to), labcomputer does seem to think the maximum is an outlier

Visually, 1968 sure looks like an outlier to me: https://fred.stlouisfed.org/graph/?g=EZn4

But let's check my intuition. The federal minimum wage was raised 10 times from 1960 through 1980. The raises have the following indexed values:

* 1961: 137

* 1963: 145

* 1967: 151

* 1968: 167

* 1974: 146

* 1975: 143

* 1976: 147

* 1978: 150

* 1979: 151

* 1980: 141

So mean = 148, st.dev = 8.1. That puts 1968 at 2.4 standard deviations above the mean.

If you just want to make the pedantic point that the maximum is not in general an outlier, we are in agreement. But, in this particular case, the maximum is clearly an outlier.

I appreciate the work you put into this, but when I look at the data, I see a nonstationary (humped) time series, where I am not sure that 2.4 standard deviations above the mean is a reliable identifier of outliers.

Furthermore, I think the data still supports janandonly's point (though not quite as starkly) even if we throw out the maximum.

It would indeed be pedantic of me to now make an issue of the maximum not in general being an outlier, but at the time I wrote that, I was replying to a post in which the only justification given for considering 1968 to be an outlier was the fact that it was the maximum.