printp(x) output is inconsistent when x has values greater & smaller than .001

[mvuorre]

Currently, if you run printp(x) on x that contains values both above and below the threshold of 0.001, you get inconsistent output values, some of which contain an operator and some don't. This is especially problematic if you're e.g. constructing a variable that has p values from several tests that you'd like to report together (in a table or in text).

library(papaja)

ps <- c(1, .05, .00001, NA)
printp(ps)
#> [1] "> .999" ".050"   "< .001" ""

# Reporting in text can result in two operators
paste0("Test ", 1:4, ": p = ", printp(ps))
#> [1] "Test 1: p = > .999" "Test 2: p = .050"   "Test 3: p = < .001"
#> [4] "Test 4: p = "
# or some without equals sign
paste0("Test ", 1:4, ": p ", printp(ps))
#> [1] "Test 1: p > .999" "Test 2: p .050"   "Test 3: p < .001"
#> [4] "Test 4: p "

# New feature is suggested in my fork's devel branch
# remotes::install_github("mvuorre/papaja@devel")
printp(ps, add_eq = TRUE)
#> [1] "> .999" "= .050" "< .001" ""

# Default is add_eq = FALSE, which gives old behavior
printp(ps)
#> [1] "> .999" ".050"   "< .001" ""

^{Created on 2019-09-05 by the reprex package (v0.3.0.9000)}

See the commit here.

An additional benefit of including those extra two lines of code in printp() is that you could remove quite a bit of code that does the same in thing in various apa_print* methods (e.g. here) and instead replace them with printp(p, add_eq = TRUE)

[crsh]

Hi Matti, thanks for the suggestion. That makes sense to me. I think, it would even make sense to add this to printnum() since printp() is just a shortcut. Addition of an equal symbol could be useful in printnum() as well when zero = FALSE or gt1 = FALSE, right?

@mariusbarth, what do you think about this? This would be a more general alternative to your proposed add_equals() in #310, right?

[mvuorre]

Great, yes it looks like you'd want to put it in printnum().

[crsh]

@mariusbarth, you worked on this right? Could you summarize how you solved this (maybe in reference to the above example)?

[mariusbarth]

Sure! On the development branch, I basically implemented what Matti suggested: Both printp() and the printnum.numeric() method now have an additional argument add_equals which defaults to FALSE. Therefore, both functions by default show their previous behavior. If, instead, you want a consistent output which is always prepended by either =, <, >, then you choose add_equals = TRUE. Hope this is exactly what you wished for.