I’ll make this a quick post today, but I come across this issue so often I figured I would post it for everyone’s benefit. You might have wondered how you should spell this statistical test. Let’s do a little multiple choice. Which of the following spellings is correct?
A. t test
C. t test
Okay, so I gave you a trick question. They are all correct and incorrect, depending on what style you are writing in and how the term is used in the text. Here is how to use the spellings correctly.
Option A is the standard spelling for the noun form in APA (see APA 6th ed., 4.44). The t is in italics because APA sets common, non-Greek statistical symbols in italics (APA 4.45).
Option B is only correct in APA when the term is used as an adjective (see the examples on p. 100 of the APA Manual).
Options C and D are never correct in APA because the t must be in italic. Turning to Chicago style, the sole reference to “t test” or “t-test” in the thousand-page Chicago Manual of Style (16th ed.) is in Table 12.3, which lists “t-test” (a hyphen, even in the noun form, with no italics). So in Chicago style it seems that we would use this spelling in both examples, like this:
The t-test results show a significant difference.
There are a couple of problems with this though. First, letters are often used as variables in equations in papers in various fields, so when a paper has both equations with a variable t (which might signify a time point) and statistical t values, some confusion could result.
Second, the letter t (no italics) is used as a standard abbreviation for metric tons (see CMOS 10.52). So when the author writes t = 1, it could technically be unclear if that refers to a value from an equation variable, a t-test result, or a measurement in tons. It probably does not happen very often that researchers are doing t-tests that involve metric tons, but it is quite common to have a variable with the letter t.
Also, t is not the only letter that can cause a problem; for instance, M is often used to refer to a mean but can also refer to molar or metal, and N can refer to a population sample and also newton or number). Plus, in scientific writing it is best to be as specific, clear, and consistent as possible in the writing. Getting in good habits prevents writing sentences that might seem okay at first glance but that can be confusing to other readers who are seeing it for the first time.
CMOS appears to lean toward Greek letters for common statistical symbols (see 12.58), but the guide does not appear to use italics for the Arabic abbreviations. Of course, the wording of the sentence and the surrounding context usually clarify, but setting statistical abbreviations in italics and nonstatistical abbreviations in regular (roman) type helps create consistency and clarity for the reader. It’s also just easier to read in my opinion, but perhaps that’s because I’m used to seeing the italics.
Probably for these reasons, many academic journals tend to italicize statistical symbols, so in most cases you will often be using Option A for the noun and Option B for the adjective, as described above.