R: why does stargazer return incorrect values for min and max with this dataframe?

问题

I was using stargazer today to make some basic summary tables, and noticed some strange behavior:

library(stargazer)

my_df <- data.frame(speed = c(3.00,3.20,3.00,3.20,3.10,2.90),
                    age = c(22,24,24,28,19,25))

stargazer(my_df, type="text",
          summary.stat = c("min", "mean", "max"),
          digits = 2)

Output is:

=======================
Statistic Min Mean  Max
-----------------------
speed      3  3.07   3 
age       19  23.67 28 
-----------------------

Clearly min should be 2.90 and max should be 3.20, as I've specified digits=2...am I losing my marbles? I'm using stargazer 5.2.2 and R 3.6.3 if it matters.

回答1:

That looks suspiciously like an undocumented behaviour (aka bug).

I may be wrong, but it seems that the function is inspecting the first value of each variable to determine how to present the min/max results, ignoring the digits argument if it finds anything that looks like an integer. For example, change the first value from 3.00 to 3.10 and you get the expected output.

my_df <- data.frame(speed = c(3.10,3.20,3.00,3.20,3.10,2.90),
                    age = c(22,24,24,28,19,25))

stargazer(my_df, type="text",
          summary.stat = c("min", "mean", "max"),
          digits = 2)

=========================
Statistic Min  Mean  Max 
-------------------------
speed     2.90 3.08  3.20
age        19  23.67  28 
-------------------------

I tried this with the mtcars dataset and got the same behaviour.

stargazer(mtcars[,c("mpg","drat","wt","qsec")], type='text', digits=2)

=========================================================
Statistic N  Mean  St. Dev.  Min  Pctl(25) Pctl(75)  Max 
---------------------------------------------------------
mpg       32 20.09   6.03    10     15.4     22.8    34  
drat      32 3.60    0.53   2.76    3.08     3.92   4.93 
wt        32 3.22    0.98   1.51    2.58     3.61   5.42 
qsec      32 17.85   1.79   14.50  16.89    18.90   22.90
---------------------------------------------------------

head(mtcars[,c("hp","mpg","drat","wt","qsec")])
#                    hp  mpg drat    wt  qsec
# Mazda RX4         110 21.0 3.90 2.620 16.46
# Mazda RX4 Wag     110 21.0 3.90 2.875 17.02
# Datsun 710         93 22.8 3.85 2.320 18.61
# Hornet 4 Drive    110 21.4 3.08 3.215 19.44
# Hornet Sportabout 175 18.7 3.15 3.440 17.02
# Valiant           105 18.1 2.76 3.460 20.22

So, I don't think you are losing your marbles. But you should report this to the package author and see what he (Marek Hlavac) says.

来源：https://stackoverflow.com/questions/61553731/r-why-does-stargazer-return-incorrect-values-for-min-and-max-with-this-datafram

标签

stargazer