Yes and no. What I have seen is that when an o2 sensor is placed in a collector it will read a rough average of four cylinders.
I say rough average because it can change the reading you see based on where the sensor is placed. Place it where it is biased to a lean cylinder and it will bias the reading more toward lean. Place the o2 sensor closer to a rich cylinder and it will bias the reading as richer than the true average of the four cylinders.
In one collector you could have three very rich cylinders and one very lean cylinder. The o2 sensor doesn't know it is four seperate cylinders it sees the rough average of the four cylinders as being rich. Even if one of the cylinders is lean. In this way a lean misfire on a cylinder could be seen as rich.
I haven't done this yet but i think if you were to put a o2 sensor in the primary tube of the lean misfiring cylinder it would show it as being lean not rich.
What have you seen when testing that has shown a lean individual cylinder misfiring and causing the o2 sensor to read rich on that individual cylinder?