jamovi

I love teaching with Jamovi...except for the part where I have to explain scales of measurement. The textbook that I use, like many, lists scales of measurement as nominal, ordinal, interval, or ratio. Jamovi lumps interval and ratio together into a single category, which in and of itself is totally fine (other apps do the same thing). The problem is the name of that category - "continuous." Most textbooks define "continuous" as the opposite of "discrete." That is, continuous implies that you can divide a variable an infinite number of times into fractional parts. That's very different from saying it's "interval or ratio." Consider, for example, the number of pregnancies a person has had. That variable is on a ratio scale, but it's also discrete (as opposed to the typical definition of continuous).

In Jamovi, could the label "continuous" be replaced with something else? For example, one software package simply calls interval/ratio data "scale" data. Another option would be to just call it "interval/ratio" or something like that. I hate to pick at such a seemingly minor point, but it does cause some confusion when teaching undergrads, and I'd love for more instructors to adopt Jamovi.

hi anthony,

yup, there's lots of different ways to describe these things.

is there a problem with the definition of continuous being the opposite of discrete?

kind regards

jonathon

i think for all intents and purposes, number of pregnancies can be considered ordinal?

At least by the definition in Gravetter and Wallnau, it's ratio because it has a true 0 point (no one answers less than 0 pregnancies), and you can interpret ratios from two scores ( a person with 6 pregnancies had twice as many as a person with 3 pregnancies). I'm starting to wonder if some textbooks are defining it differently. I have access to several other textbooks on the Amazon top 10 bestsellers lists for stats. If I took the time to look for some consensus (if any) on scale of measurement, would that be useful to you, or are you trying to get at some other distinction when users do the Data setup?

hey,

thanks for bringing this up, you have made me ponder this a bit more.

i think there is a bit of a disconnect between the "levels of measurement", and what's required in analyses.

so number of births is a ratio, but for many (most?) analyses you'd *treat it* as ordinal. or if we were dealing with the number of babies that spiders have (1000s), the we'd *treat it* as continuous.

that's the way i've internalised it anyway. the exact 'type' isn't so important, so much as the way it is treated.

but maybe we need to revisit this. tbh, you're the first person to bring this up (and i've been doing this for a while!) when i began jasp, i didn't like 'scale' for interval/ratio, because the word scale was most closely associated in my mind with likert scales! and i haven't really thought about it since.

cheers

jonathon

If we were to replace "continuous" with "scale," then we'd be in the awkward position of saying:

Question: On what sort of scale is the DV being measured?
Answer: It's being measured on a "scale" scale.

I think that as far as the software is concerned, if you want to do something like a t-test, the DV has to approximate a continuous variable (and it will never be more than approximately continuous, if for no other reason than the need to round/truncate), and the levels need to be non-dependent on one another. I don't know that there's a perfect, concise label for that. But the label "continuous" seems fine to me. Given that we're outside the realm of mathematical abstraction and in the word of actual data, it should be understood that there can't really be a continuous data scale, and that we're only approximating continuity (often, a very rough approximation).

I use all these terms when teaching so I had to check my notes to see which I used most often under various contexts: continuous.

This is an old post now... but it really bothers me too, and very much confuses my students.

Why not use "numerical' then?

The idea of using "ordinal" to describe the number of pregnancies seems a very odd suggestion. That implies a qualitative variable. But it isn't.

I cannot see any reason why the description cannot be changed to "numerical".

At least, it cannot cause more confusion than "continuous" does now. Selecting "continuous" and then "integer" just sounds... so, so wrong!

But thanks for all your work. I love jamovi.

P.

I don't think "numerical," works, because whatever term is used, it can't subsume "ordinal." Ordinal values are numerical (Ranks 1, 2, 3, 4, etc). So the selection of a term to substitute for "continuous" has to refer to a special sub-type of numerical variable that is non-ordinal. In other words, the term would need to define the set of scales consisting of ratio and interval scales. The only think I can thin of is "Interval-or-ratio,"

What about "quantitative" then?

"Continuous" surely is wrong.

P.