I don’t think my reasoning falls neatly into any one of the categories you listed, so I’ll post it as its own comment. I don’t give to “multiplier” charities mainly because I think a huge percentage of the good that they do probably comes from running great websites, but the fixed costs that were necessary to get these websites built and online have already been paid, basically, and while I believe that initial investment probably had a large multiplier, I’m far less convinced that subsequent expenditures by these organizations (other than maintaining their websites) will have such a large multiplier (and big donors would happily step in—or the multiplier charities would tell us—if maintenance costs could not be met).

Furthermore, in the exceptional cases when subsequent expenditures would likely have large multipliers, my sense is that usually, those expenditures require atypically substantial amounts of funding, without which the investments in question cannot happen. I am not a large donor, and it just isn’t clear to me that if I give a few thousand dollars to a multiplier charity instead of to, say, the GiveWell Maximum Impact Fund, that few thousand dollars will enable anything particularly high-impact to occur that otherwise wouldn’t have. By my mental model—which may be mistaken—for each additional dollar I give to GiveWell’s Maximum Impact Fund, my impact rises by some smooth function that probably isn’t far off linear. In contrast, I think that the value of additional dollars given to a multiplier charity probably follows some kind of a step function. I understand that my donations might increase the probability of the multiplier organization being able to “go up a step” sooner, but I suspect that if the step were truly likely to have an extraordinarily high charitable return, large donors, like foundations or ultra-high net worth individuals, would fund it no matter what, and the fact that I’d chipped in a few thousand on the margin wouldn’t change their calculus on that one bit. I’m just not the limiting factor here.

Finally, multiplier charities seem like sort of obvious breeding grounds for conflicts of interest in the community, and I’m quite wary about that because 1) I think the community has had a poor track record on managing conflicts of interest historically (though this has unquestionably improved), and 2) there is effectively no oversight of multiplier charities. They don’t have to go through anywhere close to the level of scrutiny or provide nearly as much transparency as GiveWell’s top charities, so I’m much more reluctant to take many of their claims to impact at face value.

Ultimately, I feel that my giving to multiplier charities would be troublingly analogous to the fact that around a quarter of foreign aid by OECD countries never leaves the donor country because it gets spent on consultants, auditors, and evaluators domestically. There is obviously a plausible case that the consulting, auditing, and evaluation in question increases the value of the foreign aid so much that it pays for itself, but doesn’t it seem more likely that these firms get retained for bad reasons (they lobby governments, have friends in high places, employ voters, tell unrepresentative horror stories about the misuse of aid, etc.) than for good reasons? I don’t mean to implicate multiplier charities in an unsavory comparison... but for the fact that the unsavory comparison is actually a meaningful reason why I don’t give to them. I just have no idea how I could tell with confidence that the way they would use my marginal dollars would actually beat other opportunities.

I don’t believe the multipliers that fundraising organizations report (e.g. because they don’t appropriately adjust for money that would have been donated counterfactually, rely on aggressive assumptions, or ignore the opportunity cost of having people working at the multiplier organization)

There's a good talk on this - Considerations for fundraising in Effective Altruism.

I'm skeptical of multiplier organizations relative effectiveness because the EA community doesn't seem that excited about them. 

(P.S.: This is actually probably my #1 reason, as someone who hasn't spent much time thinking about where people should donate.  I suspect a lot of people are wary of seeming too enthusiastic because they don't want EA to look like a pyramid scheme.)

I think multiplier organizations have provided leverage in the past, but think that going forward the marginal multiplier will be lower than the average multiplier

Multiplier organizations typically raise funds for a lot of different charities, and I only care about money that’s raised for the charity with the highest absolute impact

I’m generally skeptical of the multiplier model because it seems too good to be true

There aren’t multiplier organizations available in the cause areas I care about

I have very strong opinions on which organization in the cause area I care about is doing the most effective work, and I don't think that the relevant evaluating organization is any better equipped to opine on it than I am. I sometimes compare evaluating charities to doing a fermi estimation. If there are a sufficient number of steps to estimating a fermi problem or if you are sufficiently off on your intermediate steps, and you have a some idea about the general magnitude of the target that your are estimating, it becomes better to just directly take a guess on the end-estimate rather than attempting intermediate guesses to guide you.  It seems to me that even though evaluators are often very thorough and transparent, they end up making a ton of assumptions on top of high-error estimates (because one has to in order to make any progress, not because they don't do a great job given what they are working with). I am not at all suggesting I could do a better job of rigorously estimating which organization is better, but I don't think that is the right approach given the complexity of some of these problems. In other words, I'm far more convinced by the arguments and track record of the direct work charity that they are doing the most efficient work than I am by the evaluator charity that they should be trusted to make that evaluation.

I think multiplier organizations are significantly riskier than organizations doing direct work

I feel an emotional “warm glow” when I give to charities that do direct work, but not when I give to multiplier organizations