"Predict", "Concurrently Predict", or "Associated"? I have heard different things about the appropriate verb to use for a research question for a non-experimental design that utilizes a multiple regression analysis. Three main ones: 1. Predict, 2. concurrently predict, 3. associated with. Any justifications for one of these. As an example: "Does height and education level predict depression?" or "Do height and education level concurrently predict depression", or "Are height and education level associated with depression". 
A journal instructed one of my colleagues to use "concurrent predictors". I welcome your thoughts.
 A: I would use a very simple test, namely: was there a prediction involved? As in, were models with and without height/education fitted to some data and used to predict depression in an independent sample, and did the models with height/depression perform better than those without?
If yes, talk about "prediction". If there were two or more predictors, talk about "concurrent prediction".
If there was no prediction involved, only, e.g., in-sample NHST and p-values, then you should only talk about "association". Because there simply was no prediction involved.
From my experience in psychology, I strongly suspect the latter case. I regularly argue as here in papers I am involved with.
A: The fact that they are using a multiple linear regression model does mean that the two variables can be concurrently used to predict the response. However, I don't think that this means that it is incorrect to say  "Does height and education level predict depression?". It is not implicit in the latter that the two covariates (height & education) are being used independently of one another, although the former makes it explicit that the two are being used together. If the goal of the exercise is to build predictive models for determining someones likelihood of depression then either of the above is perfectly acceptable, as far as I know.
However, if the goal of the exercise is inference about which variables might correlate with depression then I would approach it from a different angle. Prediction is an outcome of the model rather than what the model is actually doing, which is estimating the correlation of the two variables from the sample you have collected. For this reason I would usually make a statement like: "Are height and education level correlated with depression?". 
Similarly, using the word associated I would place more in the camp of inference rather than prediction. I am less inclined to use associated simply because it carries the conotation that the two covariates are actually causally linked to the response. In the case of the above example I somehow suspect that the researchers did not experimentally manipulate individuals height or education level but rather conducted a observational study, which typically cannot prove causation; at least not in isolation. That is not to say that the two are not causally linked, but rather that you do not have evidence to support that using just a multiple linear regression. 
In spite of the above arguments I think that the whole debate is largely semantics though and the reviewer who suggested the change is being a little over-zealous and dogmatic.
