Experimenter effects are the potential faultiness of data due to the fact that people are funny.

When humans know that they are being watched or their interviews recorded, sometimes they do not behave like themselves. This can potentially skew our data.

If I were to be asked, How many MoonPies did you eat this week? I could honestly say, just two. That is because I ran out of them this week. If I were to be asked next week, you probably would have a harder time getting the truth out of me, because I would be really embarrassed to discuss it. For this very reason, it is so very hard to do nutritious/dietary studies well.