Criminal charges against an accused rapist were dropped Tuesday due to concerns about the credibility of the victim as a witness, the prosecution said.

Milton Thomas, 58, a parolee charged last year with raping a 70-year-old widow, figured prominently in a recent Associated Press investigation of a national movement to use surveys to predict which prisoners or ex-inmates will commit future crimes. Thomas, who was quoted in the AP investigation saying he was innocent, was assessed three times with three different outcomes.

States are trying to reduce prison populations with these secretive assessments, which supporters said can help reduce prison crowding and save billions of dollars because it is far less expensive for governments to supervise convicts on parole than keep them behind bars.

In the rape case, the assessments were unrelated to the decision to drop the charges against Thomas, a parolee who has been in and out of Arkansas jails since 2008 for non-violent crimes, including check fraud.

The deputy prosecuting attorney in the 6th Judicial District, Tonia Acker, said the rape case against Thomas was based entirely on the statement of the victim.

The AP doesn't identify victims of sexual assault, but Diana Miller, now 71, agreed to be identified by her middle and married names because she said it was important for her story to be told.

Acker said Miller stands by her account, but she would not share her phone records. Based on evidence prosecutors have now, "it did not comport with the allegations in the past."

Thomas' public defender, Lisa Thompson, said the contact between Miller and Thomas -- including how long they talked to each other on the phone -- was a factor.

Miller said Thomas raped her on a hot summer day last July. In letters to the AP, one of which the AP published, Thomas said he was innocent. He said he and Miller had a previous sexual relationship, which Miller denied.

Thomas will remain incarcerated for at least another month because his parole was revoked for not paying supervision fees, the parole board said.

Thomas kissed his wife Tuesday and told her he had to go back to jail until the parole board makes a decision about whether he should be released.

"This is crazy," Thomas said.

Since June 2013, Thomas was assessed three times to calculate the likelihood he might commit another crime when released on parole. All three assessments produced different results. When Thomas was up for parole in 2013 after serving time for theft, the Arkansas Parole Board assessed him as a high risk to re-offend.

He was released in November of that year, and the state's community supervision agency assessed him again. This one determined Thomas was a low risk, Thomas said, and required the minimum amount of supervision and no rehabilitative programs. After Thomas was arrested on the rape charge last July, the parole board assessed him again and downgraded his risk from high to moderate.

The Arkansas Parole Board said the system worked, and the reason for lowering Thomas' risk was varying accounts of Thomas' age when he was first arrested.