A recent example is an adaptation of a story by Daniel Pinkwater. The test panel changed the rabbit to a hare, the eggplant to a pineapple, added some commentary from other animals, and changed the moral of the story.
It has this summary:
A talking pineapple challenges a hare to a race. The other animals wager on the immobile pineapple winning — and ponder whether it’s tricking them.The grade school students were given these multiple-choice questions:
When the pineapple fails to move at all and the rabbit [hare] wins, the animals dine on the pineapple.
These questions received a lot of criticism for their ambiguity, but the problem is worse. The questions do not even try to test what the story actually said. They try to test mindreading of cartoon-like characters. Instead of asking what the animals said, it asks about their feelings, motivations, and intents.
Beginning with paragraph 4, in what order are the events in the story told?
The animals ate the pineapple most likely because they were...
Which animal spoke the wisest words?
Before the race, how did the animals feel toward the pineapple?
What would have happened if the animals had decided to cheer for the hare?
When the moose said that the pineapple has some trick up its sleeve, he means that the pineapple...
Some people say that it is important to test the emotional understanding and empathy of children, and they apparently believe that a child should be able to feel how a moose or a crow might be annoyed at a talking pineapple for being tricked into disbelieving that a rabbit can win a marathon race.
This seems dubious to me. I think that the supposedly-correct answers are wrong. One does not eat a pineapple because he is annoyed by the pineapple. Not even a moose.
Regardless of the answers, a reading comprehension test should test reading comprehension of the actual story, not test for psychological feelings about a moose anthropomorphizing a fruit.