Yes. We don’t even need to prove it, it’s common sense:

If a spammer can programmatically submit any number of requests before PoW, and then after PoW they have to spend even just 5% more compute resources to send the same amount of spam, their costs increase by at least 5%. If they can’t afford the costs, the spam is reduced. 

Other rate limiting techniques generally incur more costs on the host than the spammer (captcha, tracking, forensic analysis)… thats the wrong direction and eventually leads to centralization (only those with money run public relays). 

A relay can _combine_ techniques, if they can afford to do so, for best results. 

The goal isn’t elimination, it’s to incur costs on the spammer. We don’t need to outrun the bear, just the slowest runner.