I don’t see exactly what you’re talking about?

In the first article, if you mean 0.09, it’s of course because precision isn’t about decimal place, but about number of significant digits. The source(s) seem to be indicated.

In the second article, I don’t see any high-precision numbers. In this article, I don’t see clear links to sources, though they are mentioned in the text.

@Zaku: I meant, to the smallest unit.

It is my understanding that the larger the dataset the more significant digits they can calculate the values to. Also there is generally an error range associated with statistical calculations. Again, the larger the data set, the smaller the error envelope becomes.

See: Law of large numbers

They have to be, otherwise their results wouldn’t be worth a fart in church.
We had a major road construction here several years ago and one of the road blueprints had an error of about ¼ of an inch at one spot. The result was that the road would have a concrete pillar about ½ way into one lane of traffic. That’s why they have to be so precise.

There is error in collecting any data. Always a plus or minus a certain amount.

Data can be very precise and not be very exact.

Ideally exactness and precision come close together.

Good answers above. Adding to that, nobody thinks the numbers are precisely representing the real world to the exact number (211,597 deaths at this moment.). They are gathering information as best as they can.

For example, in the US, the counties tally up death certificates and the states compile those numbers and WorldOMeters collects those figures.

There will be mistakes. Regardless of the precise number, the trend can be seen clearly.

Yeah, if you have an exact number of known or reported cases, that’s large like 211,597, and you’re presenting a table of data, even though you know the actual number is unknown and may be likely to vary by 30%, it would not improve but reduce the value of the data to round the reported number to some round number – that’s adding an error to the reported number.

Rounding numbers is useful for broad statements and for communicating to people whose eyes will glaze over if you say too many digits, but the actual reported numbers still communicate more accurate information even if there is a broad margin or error about the actual number.

98.43234% of all respondents to this question say that statistics are, at best, indicators, and are not to be taken as precise measures.

Sometimes they make it up so precisely that it’s unbelievable! ;)

_ say that statistics are, at best, indicators, and are not to be taken as precise measures._

Nobody is saying that. These particular ones are not precise. Sometimes statistics are precise.