Children make very unreliable witnesses because their memories are less accurate than adults' memories and they are susceptible to suggestion when trying to recollect past events. Additionally, children often tell adults what the adults want to hear. Because current research suggests that children usually do not lie about sexual abuse, a presumption of accuracy exists about children's allegations. Current methods of interviewing children about sexual abuse and theories used to diagnose child sexual abuse are open to question. The case of Jordan, Minnesota illustrates how a child sex abuse investigation can go awry. Once a child sex abuse case comes to trial, the defendant's ability to cross examine a child witness is severely inhibited by Rule 608 of the Federal Rules of Evidence. Testimony of children should be excluded at trial under Federal and State rules of evidence because their testimony is likely to be the product of suggestion, and securing a criminal conviction through testimony tainted by suggestion violates the due process clause of the U.S. Constitution. Finally, the system of investigating child sex abuse cases should be reorganized to provide safeguards to protect the defendants. 190 footnotes.