How do you decide whether or not to collaborate with an acquaintance?

It is good to be wary if you don't have much experience with the person and their basic ethical patterns. On the other hand, collaboration is a good thing generally if people contribute on an equitable (not necessarily equal) basis.

It depends on how much risk you are willing to take here and how important this particular project is to you. Once you agree, and the other person contributes anything, it will be difficult to back out. If success in this project seems vital to your future, I'd suggest not taking on much risk here. If it is "just another paper among others" then you can afford to take a chance.

But, I'd suggest that you work out the issue at the beginning of the project and not hope for the best. In particular: Will you be co-authors of the work? Who is the first/primary author of the resulting paper, assuming that is a consideration? Or, will one of you be the author and the other get appropriate acknowledgement? How much effort is each willing to put into the project. If it is a major effort for one party but just a minor (hobby) project for the other, then misunderstandings are possible. Are you willing to do the work yourself if the other party quits? Will you still be OK with co-authorship in that case?

Think about the worst case. Is that acceptable? But think about the most likely case, given what you do know about the other person.

Lots of questions on this site arise because participants didn't set up parameters and understandings at the start. The writers are mostly unhappy with the outcome and wonder how to recover. There are usually comments that indicate these things should have been worked out early and it is too late, at the end, for a happy outcome.

However, I think that most collaborations are more successful and can lead to long term productive relationships. Valuable, but with both upside and downside potential.

Buffy's answer already makes a key point: discuss the potentially thorny issues at the start, not at the end when you feel the sting.

I'd like to add to that: put some work into getting to know this person. Rather than agree to a collaboration right away, gather some information first;

• Look up this person's previous publications. What's the quality like?
• Look at their academic profile - how productive is this person? Where do they work? What previous topics have they worked on?
• Find out if you have any shared acquaintances that can tell you a bit about the person. Has this person done any collaborations with people you know? Does your supervisor or any of the senior academics in your department know them, or know something about their institution?
• Get to know them in person a bit before saying yes. Are you planning both to go to any conferences? Are they living nearby enough to go for coffee?