The situation is IMHO even worse than you describe, in that the person originally doing the research for a duplicate does not get to review the question as they already have voted the question to be closed.
Finding a duplicate, which I agree is one of the harder tasks if you don't happen to remember to have seen the question before, is thereby not rewarded towards the reviewers queue at all.
It would be more fair if e.g. the initiator of a "close" would get two review points on actual closing—when the other necessary votes have been cast. And maybe a minus one on the review point, if the question is left open with an otherwise unanimous cast of "Leave Open" votes (to discourage frivolous initiation of closing).
In general there will always be different, however difficult to judge, efforts in the system:
- making a suggested edit (rep <2000) that only involves adding a tag, or removing thanks versus multiple reformatting, with manual adding of extra spaces etc., to make things look good¹.
- reviewing small suggested edits, versus large ones
- finding posts on multiple sites, the more difficult if the OP has posted anonymously, or with multiple accounts
However as an alternative to differentiating the award for normal reviews I would rather suggest to have the system assist reviewers in the more difficult review activities:
- When you type a new question (or its title) you get a list of possible duplicates. A reviewer could be presented such a list as well (optionally), based on title, tags and content.
- Link to other activities of the same OP within a certain time frame, or if the same/similar title is used on another SO/SE site.
- Being able to see the (initial parts of the) original question and a duplicate at the same time, the switching means you have to read through one completely, and memorize it, then switch and read; or alternatively go back and forth multiple times and loose where you had scrolled down to.
- While reviewing reopen votes, see the original reason for closing (which disappears when looking at changes, forcing you to go back and forth more often than necessary).
I also tend to open the original post when reviewing, to get more context than an answer to be reviewed and the original question can give. Sometimes I look for originality of the answer (not a offence if not, but I tend to comment on answers that contribute nothing new). The system could help there as well.
Of course with any change to the system it is important not to nudge the reviewer to much into the direction of a particular decision.
¹ That judgement is subjective. In reviewing suggested edits by the same person, removal of "Thank in advance" statements (which annoy me to no end) is something I almost always Accept, whereas adding one tag will I will Accept the first 2-3 times and then Reject as "Too Minor"