The main differences between the electrodes are the materials, but that will not affect the overall scale of the data. As you can see, the forehead electrodes are made of metal, and the ear electrodes are made of conductive rubber. The gain is the same for each channel, though.
Generally, the data from the ear channels will look different than the data from the forehead channels, but this is because they are receiving very different signals (they are in different locations, after all), not because the hardware is different.
You are correct about the order of the channels from MuseIO as recorded by MuseLab. So in that table, the order should be as you’ve stated: left ear, left forehead, right forehead, right ear.
The fact that one of the channels has different-looking data on it just means that that particular electrode is receiving a different-looking signal, not necessarily that something is wrong with the headband itself. From that very small snippet in the table, it looks like it is fluctuating quite a bit, so perhaps it is not well connected to the head, or there is some large disturbance in the signal such as muscle movement.
Are you observing a real-time graph of the data in MuseLab while you’re recording it? That can be helpful for making sure you’re getting reasonably good data. Normally I would suggest using the signal quality indicator in the top right corner or MuseLab, but there is a bug in the current SDK which disables it. We’ll release a fix for this soon.