Yours is a good description of publication bias. I think it's worth mentioning that funnel plots can give you a visual sense of whether there is publication bias, and tools such as Egger regression can help to try to tease out a less biased estimate in the face of publication bias.