Had the question been to seat the 10 people around a circular table with two specific people seated together, the solution would match your suggestion:
- Seat the first of the two specific people in 1 way (as all positions at an empty circular table are identical under rotation)
- Seat the second of the two specific people in 2 ways (in one of the two seats adjacent to the first specific person)
- There are 8 people remaining and 8 seats, so there are 8! arrangements.
- Thus, total number of arrangements is 2 x 8! = 80,640
However, the question has two groups of 5 people (students and parents) and we need
at least two students together. We could do this by counting up the cases:
- with exactly two students together (and the other three students not together with the first two, but possibly together with each other)
- then, with exactly three students together (and the other two together or separate but not together with the first three)
- then, with four students together
- then, with all five students together
but that is long and involved.
We are better off looking to the complementary event, which is the single case with no students together. This can only be done if the seating alternates parent - student - parent - student etc.
- Total number of arrangements of 10 people at a circular table is 9! = 362,880
- Seating alternately, we seat a student first in 1 way, then the other four students have only four possible seats, so seat them in 4! ways. The remaining five seats go to the five parents in 5! ways
- So, seating alternatively can be done in 1 x 4! x 5! = 2,880 ways
- So, at least two students together has 9! - 4! x 5! = 360,000
Intuitively, the likelihood of there being at least one pair of students together if seated randomly is quite high as the only way that it doesn't happen is if the result is an alternate seating pattern - and alternate seating is unlikely. Two students together is also less likely if it is a specific pair who must be together. Looking at the above answers, we have:
This fits the pattern that is expected, and shows that the specific two together interpretation that you are thinking about has a probability of around 22% compared to over 99% for at least two of the students, not specifying which, are together.