Anecdotally, I have seen ACT scores improve with work and time. Most people at my high school took the ACT not the SAT, so possibly the effect is not the same. But, as I'm sure you would agree, it's a terrible argument to say a given study must be wrong because I observed differently in a unscientific sample from my own personal life. I will concede that changing your score is difficult, and possibly much more so than my experience shows.
That still does not change my central assertion that the ACT or the SAT are not good tests of intelligence. Or that IQ tests are not good tests of intelligence. I don't deny that high scoring test takers of such tests are usually intelligent but as the famous line goes, correlation does not imply causation.
Intelligence is a very tricky thing to define. People typically try to quantify intelligence as one's ability to solve tricky problems, solve problems quickly, retain hordes of knowledge, etc; but these things leave out rather important aspects of intelligence such as creativity, so-called "emotional intelligence", and the value of personal experiences. This is not an exhaustive list of things that are missed by these tests. Suggested reading here is Malcolm Gladwell's excellent book Outliers.
I'm moving the goal posts a bit on this conversation, but an even broader point to make about intelligence is that if you have a certain amount (what ever that means, again I'm not claiming that you can measure this), magically gaining more intelligence would have relatively little impact on your ability to have a successful life. It is my opinion that hard work, social connections, and even luck are more important than intelligence. Again, anecdotally, I have seen brilliant friends struggle to gain any sort of measure of success in life while the friends I judged as less intelligent are blowing them out of the water. Successful people are intelligent people, but they are rarely geniuses.
And this is where my argument begins to loop back around to the point the original article makes: Outcomes improve for people from poor and underprivileged backgrounds when they are integrated into a social groups with people who make them feel like they belong in higher education and encourage them to learn from bad test scores or rough semesters. So what does it matter if this girl has only slightly above average intelligence (as measured by tests that are flawed indicators of intelligence anyway)? If participating in these programs helps her become a productive member of society and become successful against her personal goals, isn't that a huge win?