Which types of analysis? (continuous DV ranging from 0 to 1)

[Chanhum Yoon]

Hi,

My dependent variable is score, continuous variable ranging from 0 to 1 (e.g., 0.1, 0.23, 0.34).

What I need to do is to predict who has a lower score (i.e., who is closer to 0).

However, I was wondering whether the value of 0 can be predicted? Is there any suggestion such as variable transformation?

Also, I would appreciate if you could recommend the types of regression I can run.

Thanks in advance.

[Carlo Lazzaro]

Chanhum:
I would consider -regress-.

[Chanhum Yoon]

Hi Carlo, I don't think I can run a normal regression here? My DV ranges from 0 to 1 as a proportion/continuous variable and I need to set the direction to predict who has low score (i.e., heading zero, not one.)

[Carlo Lazzaro]

Chanhum:
could you please share and example/excerpt of your data via -dataex-? Thanks.

[Nick Cox]

A continuous response or outcome between 0 and 1 could be tackled, respecting the range, with a generalised linear model -- say link(logit) family(binomial) vce(robust) -- or with beta regression.

However "who has a lower score" needs more explanation here: does this mean a person? or do you mean just which observation?

[Phil Bromiley]

It may be in essence the same as what Nick suggests, but you might look at fracreg,