There are two aspects to this.
The first - Did the votes cast for Perot on Election Day swing the results?
The answer appears, as mentioned in the Wikipedia article cited by OP, and in the video and podcast from Five Thirty-Eight linked in the comments by Bobson, to be "no." Five Thirty-Eight is generally known for their very analytical approach to looking at on going polling and election results. In the linked sources, they paid particular attention to the fact that Perot was in, then he suspended his campaign, and then he jumped back in. If Perot was peeling off a particular partisan side of the equation, then you'd see his moves having a particularly significant impact on that side. That was not found to be the case.
Specifically, exit polling consistently showed that Perot voters, when asked what they would have done if not for Perot, and it was 38% each for Bush and Clinton, with 24% staying home. They also emphasized that when Perot got out of the campaign in July, it was because Clinton had surged from third to first after the Democratic convention, and that while Perot was out and after Perot got back in, Clinton was leading in the polls throughout.
Video - Five Thirty-Eight: The Ross Perot Myth: Deep voodoo, chicken feathers and the 1992 election (about 11 mins run time)
Cornell University also did a review of primary polling, looking at results when Perot was not yet an option and when he was in the race:
About 1/3 of the Republican voters in the primaries were not happy with Bush. Pre-Perot - Among those who were satisfied, he has 95% support. Among that 1/3, when Clinton was the nominee (as opposed to a "generic" Democrat), about 1/3 of them said they would vote for Clinton. About 1/2 said Bush would still get their vote, and about 1/6 said they'd stay home.
Once Perot was an option, he became the favorite among the unhappy GOP primary voters - he had up 40 to 50%, Bush stayed around 40 to 50%, so about a 5% average disaffected voter hit, and Clinton faded to about 10% as an option (about a 23% hit). Perot clearly picked up a good portion of the "stay home" voters.
Among the 2/3 happy with Bush's performance, he slight drop in support.
In terms of overall GOP base/primary support, Bush saw, roughly, an overall 8% drop, Clinton lost 8% of the overall GOP pool, that he previously had, to Perot. A pretty close split, though opinion polling that far ahead of elections is much less reliable than actual election day results. Clearly, there was an indication that a portion of the GOP voters (about 18% by this polling) was already going to defect from Bush or stay at home. By those estimates, Perot was peeling off about 16% of the GOP voters.
The Democratic impact was much harder to gauge, since Clinton ran in a contested primary, but the majority of Brown and Tsongas voters didn't have a great opinion of Clinton's worthiness for the White House, and among those dissatisfied with Clinton in New York, for example, the preference was evenly split between Bush and Perot.
exit polling '92: on the perot factor, and "who votes?" - Roper Center
Now, something that the GOP and conservative pundits, when faced with those facts, say is "you can't tell me he didn't have a bigger impact on Bush." This might seem like digging in the heels, but it speaks to a different matter than just counting where his votes came from. How did his candidacy shape perspective on President Bush?
Perot's persona and reputation was one of a straight-talking, no bs (albeit a bit quirky), no-nonsense person. He also didn't have a particular party history, platform and political base or hierarchy that he had to work his way through before becoming a candidate. As such, his views and utterances were perceived as being less influenced by party or ideological doctrine than the main party candidates.
So, there is the possibility that his criticisms would not be as easy to dismiss, out of hand, as knee-jerk partisan political utterances. Who did he attack more? Definitely Bush.
As the non-mainstream candidate, looking to topple the party politics cart, so to speak, he's going to attack the status quo of federal government. Any non-fringe third party candidate is generally going to take that tact. From a strategic political point of view, is he going to make a strong case as the outsider option to right a nation going in the wrong direction by attacking the incumbent and his policies, or by attacking the former governor of a state with the population the size of Atlanta, a median income ranked 47th and education rank of 49th? I mean, he might easily make a case that such a person is not qualified, but, unless that candidate is way ahead, it's only going to be about fighting for second place. To ascend to the top office, the case to be made is against the status quo, generally, and incumbency in any federal election is a huge advantage.
The big disadvantage of incumbency, and another reason why it makes for a target, is because there is an established track record to attack, much moreso than a small state office holder.
The strategic targeting of the status quo/incumbent, the track record to go after (especially with an economy struggling under massive, for that time, record deficits, and a recession that hit hard well after most of the "big name" Democratic possible contenders sat out because they didn't think incumbent war-president Bush was beatable) and the added effect of it basically being two out of three candidates going after the third could very well have hurt Bush's reputation and standing with voters more than if it was just Clinton.
However, generally, a third party candidate doesn't mount a campaign unless there is a weakness or public dissatisfaction with the incumbent. Was Bush weakened because Perot was in the race, or was Perot in the race because there was a weakness there?
The second of the two factors (votes cast vs public perception) has a possibility of being an impact on the results, but that's a much less tangible and more nebulous discussion. In terms of the actual votes cast and where they came from, it's pretty much accepted by those with objective expertise that Perot's disparate impact on Bush is a "myth."
As far as the OP's edit, where a single poll of like/dislike from a single state is cited, that seems to be a bit of a non sequitur. A like/dislike poll says nothing about whether that like/dislike translates to actual support for Perot over any particular candidate. I can say "I like Perot" while liking Bush or Clinton more, or just feeling they are better or more practical choices for candidate. OP's contention that his impressions are somehow verified by the historical record, with that poll snapshot is not accurate. It's a shame that OP didn't give the link to the original source of the image, so we can see which "statewide" impressions that came from.
Also, the polling consistently showed that in states where Clinton won, Perot peeled off more Clinton voters than Bush voters (the ratios pretty much matched the differential), and vice versa in states that Bush won. So, if this one statewide poll was from a state that Bush won, it matches the polling data, and Perot having a 2 to 1 former Bush to former Clinton voter margin, in a state that Bush won, would not exactly bolster the argument that this hurt Bush. It would be useful to know what exact poll is being referred to.