OPINION

ROBERTSON, Senior Judge.

STATEMENT OF THE CASE

Defendant-Appellant Michael J. Sigler (Sigler) appeals his conviction of robbery, a Class B felony. Ind. Code § 35-42-5-1.

We reverse and remand for a new trial.

ISSUE

Sigler presents two issues for our review, one of which is dispositive: whether the trial court erred by limiting Sigler's cross-examination of certain witnesses.

FACTS AND PROCEDURAL HISTORY

In the early morning hours of December 1, 1997, Matthew Powell (Powell) and his friend, John Honabarger (Honabarger) were in Powell's apartment watching television. After hearing a knock on the door, Powell answered the door, and an individual asked if Craig was home. Powell responded that the person had the wrong apartment and closed the door. Later there was another knock at the door, and, when Powell opened the door, five men entered Powell's apartment and began beating Powell with a baseball bat. Honabarger was also hit over the head with a baseball bat. While Powell was being beaten, others in the group rummaged through Powell's apartment. The intruders took pain medication and marijuana and left. Powell and Honabarger went to the hospital where they were treated and released. Following this incident, Sigler was charged with and convicted of robbery. He now appeals that conviction.

DISCUSSION AND DECISION

Sigler contends that the trial court erred by limiting his cross-examination of certain witnesses. Specifically, three of Sigler's co-defendants testified against him on behalf of the State in exchange for beneficial plea agreements. The trial court limited the extent to which Sigler could question his accomplices about their agreements with the State, and Sigler asserts that this limitation violated his constitutional right to cross-examine the witnesses against him.

Our supreme court has long held that beneficial agreements between an accomplice and the State must be revealed to the jury. Morrison v. State, 686 N.E.2d 817, 818 (Ind. 1997). Although a criminal defendant's Sixth Amendment right to confront witnesses is subject to reasonable limitations based upon the discretion of the trial court, the bias or prejudice of a witness can be explored at trial and is always relevant as discrediting the witness and affecting the weight of his testimony. Standifer v. State, 718 N.E.2d 1107, 1110 (Ind. 1999). Thus, this rule serves to help the jury better assess the reliability and honesty of the felon-witness. Morrison, 686 N.E.2d at 819.

Although a defendant may be denied the opportunity to fully cross-examine witnesses about their bias toward the State, a conviction will not be reversed if the State can demonstrate that beyond a reasonable doubt the error complained of did not contribute to the verdict obtained. Standifer, 718 N.E.2d at 1110. This is known as the harmless error analysis. Id. at 1110-11. Whether the trial court's error is harmless depends on several factors, including: (1) the importance of the witness' testimony in the prosecution's case, (2) whether the testimony was cumulative, (3) the presence or absence of evidence corroborating or contradicting the testimony of the witness on material points, (4) the extent of cross-examination otherwise permitted, and, (5) the overall strength of the prosecution's case. Id. at 1111.

In the present case, defense counsel wished to question Sigler's accomplices about their terms of imprisonment and the benefits they received from the State by agreeing to plead guilty in their respective cases. The trial court sustained the State's objection to defense counsel posing questions about the specific number of years of imprisonment that Sigler's accomplices would receive in exchange for their pleas of guilty, as well as the specific number of years of imprisonment that they were exposed to had they not received the beneficial plea agreement from the State. Both the State and the trial court expressed concern about the jury discovering the potential penalties to which Sigler would be exposed if found guilty.

However, defense counsel did not contest the State's objection but instead presented an alternative method to alert the jury of the witness' bias and prejudice. Defense counsel suggested, and the trial court and the State agreed, that he question the State's witnesses about their plea agreements, and, when referring to the reduction in prison terms, he use percentages rather than a specific number of years of imprisonment. For instance, defense counsel was allowed to ask the witnesses if their sentences had been reduced by a specific percent, so as to give the jury an idea of the extent of the benefit the witness received from the State while at the same time concealing the specific number of years they would be imprisoned, or the specific number of years to which they could have been exposed had they not pled guilty. Further, defense counsel was permitted to question the witnesses as to their pleas of guilty to felonies that are a class lower than the one with which they were originally charged.

Our supreme court has stated:

While sentencing information may not be relevant to the jury's duty in a criminal case, we do not perceive it to present a substantial risk of prejudice to the State. To the contrary, however, significant harm results when the jury is prevented from learning the extent of benefit received by a witness in exchange for his testimony. It would be obviously relevant and proper for a jury to consider the amount of compensation a witness expects to receive for his testimony. It is equally proper for this jury to know the quantity of benefit to accusing witnesses. It is quite relevant whether they are thereby avoiding imprisonment of ten days, ten weeks, or ten years.

Jarrett v. State, 498 N.E.2d 967, 968 (Ind. 1986) (emphasis added). Here, the trial court improperly disallowed defense counsel's questioning of the State's witnesses regarding their specific terms, in years, of imprisonment. Next, we consider whether the error was harmless.

Although defense counsel artfully informed the jury of each witness' benefit derived from their agreement with the State, he was not able to pose questions regarding the specific amount of time, in years, that each witness would be imprisoned. For example:

Q [Defense Counsel]: You were originally charged with a class felony that's higher than the one you actually pled to; correct?

A [State's Witness]: Yes.

Q: And I think as the deputy prosecutor stated, by pleading guilty you actually reduced your sentence by, it was actually 16 percent I think was the figure of what it actually could have been.

A: Yes.

Q: That's a substantial cut in time; isn't it?

A: Yes, it is.

Q: That's a fairly powerful motivation for testifying?

A: Yes.

Q: And in the plea I think part of it was that you testify truthfully. Do you know who determines whether or not you're testifying truthfully?

A: I know I am. I don't know who determines it.

Q: Who's your plea agreement with?

A: The State.

(R. 706-07). The statement that the witness' term of imprisonment was reduced by sixteen percent has much less of an impact than saying, for example, that the witness avoided ten years of imprisonment.

Additionally, in evaluating the four factors set forth by our supreme court, we conclude that the trial court's error was not harmless. Although the testimony of the accomplices was cumulative to that of Powell and Honabarger, the State's case hinged on the testimony of the three accomplices. Further, although Powell identified Sigler as one of the men standing outside his apartment door, the testimony of each of the three accomplices provided further testimony as to Sigler's specific involvement in the planning and execution of this robbery. No other witnesses besides Sigler's accomplices could provide this evidence on these material points. Moreover, defense counsel was permitted to perform otherwise extensive cross-examination; however, as evidenced by the colloquy set forth above, this specific topic had a diminished effect upon the trier of fact when the trial court unreasonably restricted counsel's ability to elicit the information via full cross-examination. We cannot say that, without a reasonable doubt, the error of the trial court did not contribute to the verdict in this case. Thus, we hold that there is no sufficient alternative to allowing a criminal defendant to question a felon-witness regarding the specific number of years of imprisonment he or she has avoided by agreement with the State.

Additionally, we note that retrial of Sigler is permissible on remand. See Grundy v. State, 695 N.E.2d 167, 170 (Ind. Ct. App. 1998) (stating that although double jeopardy bars retrial in cases of reversal for insufficient evidence, it does not bar retrial in cases of reversal for trial error).

CONCLUSION

Based upon the foregoing, we conclude that the trial court improperly limited Sigler's cross-examination of the State's witnesses, and the error was not harmless.

Reversed and remanded for a new trial.

BARNES, J., and BAKER, J., concur.