**Garbage in, garbage out.** If most of the underlying studies are poorly designed, combining them doesn't magically produce good evidence. - **Publication bias.** Studies that find an effect are more likely to be published than studies that don't. Meta-analysts try to account for this (using techni