But, from their perspective, are Labor MPs making the right choice? Let's be completely reductive and assume that the leadership choice boils down to two things: who the MPs think is the best leader while they are in government, and who they think is the leader who is most likely to keep them in government.
In terms of who Labor MPs think is the best leader while they are in government, the evidence suggests Gillard, and by a considerable margin. The article by Kevin Rudd's former speechwriter in yesterday's Age is far from the first bit of anecdotal evidence of Rudd's unpopularity amongst his staffers and colleagues. The next election is in 21 months - despite the leadership changes to date, you would expect that whoever wins the leadership vote this week would be leader for most of those months; that is, there is a very strong chance that if the ALP vote for Kevin Rudd he will be leader for the next 21 months, and similarly if they vote for Gillard. Anecdotal evidence suggests that Labor MPs would put a large negative value on the first outcome.
The case for Rudd then rests on two things: the first of these being that he potentially gives the ALP a better chance of winning the next election and remaining in government. That may be true, although polls suggest that even with Rudd as leader the Coalition would still be considered more likely to win the next election. Also, the next election is 21 months away! - to put this in perspective it's been less time since the last election, and one can see how much the figures have changed since then. Of course things could get worse for the ALP over that time, but really that would likely only have an impact on the ALP's chance of winning elections even further away than the next one.
So, again keeping in mind that we're being totally reductive here, the choice of Rudd v Gillard for Labor MPs seems to boil down to the trade-off between the relative negative impact on Labor MPs (in their opinion) of having Rudd as their leader while they are in government compared to the increase in probability of winning future elections if Rudd is the leader. The first impact, being closer and more within the MPs control, is more certain to them than the second, and that could be seen as swaying them to retain Gillard as leader.
But wait... there's the second thing, which is that Rudd may resign his seat if he loses the vote, therefore bringing down the minority government, whereas Gillard will likely go to the backbench if she loses. Depending on how likely it is that Labor MPs think that Rudd will resign, this possibility could dramatically erode the certainty of having a Labor Prime Minister for the next 21 months. That is, it may be better for them to pick Rudd primarily based on the expectation that if he loses, a chain of events will occur whereby the government loses office. Well, it doesn't look like that's having a large influence on the expected numbers, but it's a good reminder of the power of "credible threats" (though to be fair not a threat that Rudd has made, to my knowledge) to change the outcome.