When nine miners were pulled unharmed from a collapsed Pennsylvania mineshaft in 2002, a representative placard read: "Thank you God, 9 for 9."....When 12 miners were killed in a West Virginia mine explosion in January 2006, no one posted a sign saying: "For God’s sake, please explain: Why 1 for 13?"
This was part of an argument for the conclusion that theists when talking about God use, in her words, "double standards of a kind that would make even affirmative action look just". But the natural response would be that Mac Donald is being very selective in her evidence. The most plausible reason why people have signs like "Thank you God, 9 for 9" and not "Why 1 for 13?" is that it is normal for people to celebrate with placards, and not normal for people to grieve with them. The problem of evil is not something believers have just ignored through the ages; the questions Mac Donald thinks should be asked are asked. They just are not thought to override all other considerations. In any case, this did, as I said, seem to be a tangent. In the newer article, because of some critics, the tangent has taken center stage. It is handled better, but is still problematic.
One of the problems with Mac Donald's new argument is that she assumes without argument that they do override all other considerations, a common rookie mistake in dealing with the problem of evil. She argues:
Let me take a banal example. As I write this, the Los Angeles Times has a small item on a thoroughly unremarkable traffic accident. A 27-year-old man in Los Angeles misread a traffic signal, and drove his car into an oncoming Blue Line Metro Train. He and his sister were killed; his 7-year-old son and his grandmother were seriously injured.
Now imagine that a human father had behaved towards the occupants of the car as our Divine Father did. That is: a) He knew that his children would be mowed down by a train; b) he had the capacity to avert the disaster through any number of, for him, quite simple means; and c) he chose to do nothing. No one would call this father’s deliberate and possibly criminal passivity “love.” Instead, we would deem such a father a monster and banish him from our midst. Yet when God behaves in just this way, we remain firm in our conviction that he loved the occupants of that car, and that each was “precious” in his eyes.
Of course, the whole argument hinges on the assumption, which most believers would not grant her without clarification, that "God behaves in just this way" -- to 'behave in just this way' requires one to admit that God 'chose to do nothing'; whereas, at most, most theists would concede that it's a case where God chose to do something other than what a human father would do -- which, given that God is not a human father, is not obviously problematic, and needs to be ruled as unacceptable by some kind of argument. To point out just one obvious example, one would need to rule out that God permits death in order to 'welcome his children into the perpetual bliss of the saints' -- i.e., would need to argue that this is either not a possibility, or is not what actually happens, or is just as bad as choosing to do nothing. Whether they are right or wrong, many people regard themselves as having reason to believe that God does, indeed, give people bliss at death, so when dealing with real theists Mac Donald cannot ignore possibilities like this (i.e., the various things God might be doing besides 'nothing'). Otherwise what she calls 'objective evidence' is not really objective evidence, but a tendentious and controversial characterization of it.
But that's a relatively minor issue. The chief problem with Mac Donald's reasoning is the chief problem that infects most reasoning about conflict between a good God and evil, namely, that it's a design inference, and what is more, it is one of the weakest kinds of design inferences: it's an argument from an alleged type of design to the character of the designer. This is only possible if the type of design is (1) accurately characterized; and (2) of the sort that it could not reasonably be treated as the result of a designer of a certain character-type. Despite her verbal appeal to 'objective evidence', Mac Donald does none of the evidentiary analysis that would be required to establish either (1) or (2). And, as a matter of fact, her reasoning depends on the assumption that the only relevant evidence in drawing a conclusion is the evidence of problem-of-evil cases themselves.
To see that Mac Donald's argument requires this assumption, note how she proceeds in her reasoning:
Perhaps when believers speak of God’s “love,” they use the term in a way that has nothing to do with ordinary usage. Novak maybe implies as much when he states: “What is difficult to believe is that any one of us . . . knows more than God does about His love for every individual.” God’s “love” is different from human love; it includes the capacity to foresee and watch the destruction of one’s children and not intervene. But then why not use a different word entirely — “callousness,” say. At the very least, if we are going to continue to use ordinary words in counterintuitive ways to refer to God, we should give them some sort of diacritical marker to let listeners know that the words they are hearing don’t mean what they ordinarily mean. One could speak of “G-love,” for instance, to distinguish it from ordinary human love.
But this can only be drawn from the reasoning so far if we assume that there is no other evidential basis for talking about 'love' or the like except for the problem-of-evil type of case. If there is another evidential foundation -- e.g., moral arguments, or religious experiences -- that provide a basis for the application of the term 'love' to God, Mac Donald could not conclude that 'love' in this context is used "in a way that has nothing to do with ordinary usage". Nothing to do with ordinary usage is a strong claim, much stronger than Mac Donald is entitled to make, unless she is assuming that all the relevant evidence is contained in the problem-of-evil type of case, which provides on its own no basis for using the term 'love'.
The assumption also clearly comes out in her discussion of the 'readability of the divine will'. If someone, let's call her A, has a good reason, or what she thinks is a good reason, for thinking B thoroughly trustworthy, there is no problem whatsoever with her supposing that, when apparently contrary evidence suggest otherwise, that this is merely due to her not knowing all the facts. In other words, the original good reason becomes a reason for not regarding the apparently contrary evidence as really contrary. This is an entirely rational move, although it can get slippery; and rational people make this move in sorting evidence all the time. We have to sort our evidences; and this is one of the most rational ways we do it. What the apparently contrary evidence has to do in order to be recognized as really contrary is to undermine the original reason for thinking B trustworthy. Merely pointing out that it is apparently contrary is not good enough, because there is apparently good reason to think that the evidence is not really contrary. What Mac Donald has done is simply point out something apparently contrary; she has not obviously done anything to undermine what the theist thinks is good reason for considering this apparently contrary evidence not to be really contrary. But when she discusses the 'readability of God's will', her argument can only work on the assumption that the theist has no (standing) good reasons for taking the apparently contrary evidence (the bad situations) to be really contrary to the thesis (that God loves us all and works for our good). It depends on the claim that the theist is merely assuming that the good is imputable to God and the evil is not. But she hasn't shown this, unless we make the assumption that the problem-of-evil type of case is the whole of the relevant evidence.
With this assumption, however, it may be said of Mac Donald's argument, as she says of Novak's, that it is "more conclusory than evidentiary." That is, it follows not from the 'objective evidence', as Mac Donald suggests, but from a controvertible assumption about what can be admitted as the complete evidence, added to a controvertible characterization of this evidence when we take everything into consideration. This is perhaps not surprising. Contrary to the way it is sometimes treated, the problem of evil cannot be treated in isolation; how you handle it will depend on your views about other things -- about the evidential value of religious experiences, about the nature of morality, about the proper way to understand evil and suffering, about the good, the beautiful, the true, and the like. Cleanthes, in Hume's Dialogues Concerning Natural Religion, would be troubled by her argument, because Cleanthes would agree with her assumption about the narrow field of evidence, and would be committed to her characterization of it. But most theists are not.
I should say that it is possible to have an argument of the sort Mac Donald is gesturing at, and some atheists make an effort to provide it; but it requires a rather sophisticated conceptual analysis of terms like 'love' and 'justice' in this type of context, one which Mac Donald certainly does not provide. Since there are several mutually exclusive ways one might go about doing this, and we don't know what Mac Donald would prefer, there's not much to say in response, beyond the very general and vague points noted above. There is no possible response to the particulars of an argument that is not made. Nonetheless, the argument, despite its flaws, is interesting enough to be worth a read.