Distinguish opinion from suggestion while reviewing a manuscript?

Question

First time reviewer here. My purpose as a reviewer is to make recommendations on how the manuscript can be improved, right? In this effort, should I distinguish my solid expert opinion from mere suggestions?

For example, if I see an error, then I believe I should professionally but firmly point it out:

"X is not a valid conclusion because of y. It should be removed."

However, what if the issue is not a clear-cut error but rather a matter that, in my opinion, would strengthen the quality of the manuscript?

Should I use the same firmness? (Option A)

"X should be added because of y."

or distinguish this as an opinion (Option B):

"Perhaps it would be beneficial for X to be added. Consider including x due to y."

As an author on the other side of things, I would appreciate the distinction so that I understand what either needs to be changed in a revision (or needs to be strongly justified why the reviewer was wrong) versus what I can freely disagree with the reviewer about without having to make a strong justification about why I disagree.

This issue is giving me pause because I've rarely or never seen Option B in reviews of my own manuscripts. All comments by reviewers are similar to Option A.

My apologies if my examples are too vague. I tried to make the applicable to the whole community here which requires ambiguity unfortunately.

Answer 1

You have space. If you have a scheme of what to do, just follow it. Read further if you don't have any scheme:

Your purpose as a reviewer is typically NOT to make recommendations on how the manuscript can be improved. You are NOT asked for suggestions how to improve. You are asked for an evaluation. I suggest the following scheme:

[

Summary:

...

High-level issues:

X is not a valid conclusion because of y. It is wrong to include x into the paper.

...

Low-level issues:

...

Improvement suggestions:

Perhaps it would be beneficial for Z to be added. Consider including Z due to W.

]

Include "improvement suggestions" only if you really wish to or if you are asked for that explicitly. If in addition you are asked for "nondisclosed comments to the editor / program-committee" or something else, then include such parts, too.