Leo Schofield Jr. appeals the summary denial of his postconviction motion filed pursuant to Florida Rule of Criminal Procedure 3.850, in which he raised three grounds: newly discovered evidence (ground one) and loss and destruction of evidence (grounds two and three).1 We reverse and remand for an evidentiary hearing on all claims. For ground one, the court below erred in its legal conclusions made to justify its summary denial. On remand the court is instructed to carry out the full cumulative analysis of the evidence that is required to properly evaluate the claim. For grounds two and three, the record does not conclusively refute the allegation of bad faith on the part of the State in the loss and destruction of potentially exculpatory physical evidence. On remand the court must conduct an eviden-tiary hearing and make the findings necessary to properly evaluate these claims.
Schofield was convicted of the first-degree murder of his wife Michelle in 1989. Michelle Schofield left work at 8:15 p.m. on February 24, 1987. Her car was found on the side of 1-4 on February 25, 1987. The car was not in running condition as the flywheel had come off and there was only a small amount of water in the radiator. Her body was found on February 27, 1987, in a canal in an area near the intersection of SR33 and 1-4 in Lakeland. She had been stabbed twenty-six times. Fingerprints that did not match the defendant were found in the car. The newly discovered evidence at issue is a Florida Department of Law Enforcement (FDLE) report of December 26, 2004, revealing that the previously unidentified fingerprints were matched to a person named Jeremy Lynn Scott. An investigation into Scott, detailed in the postconviction motion, revealed that he has an extensive history of criminal convictions, he was not in jail at the time of the Schofield murder in 1987, and at that time he frequented the area where Michelle Schofield’s body was found. *92Since 1989 Scott has been imprisoned for the 1988 first-degree murder and armed robbery of Donald Moorhead in Lakeland.
Ground One: Newly Discovered Evidence
Schofield argued in his postconviction motion that he was entitled to a new trial based on the fingerprint analysis linking Jeremy Lynn Scott to the victim’s car and the subsequent investigation showing Scott’s violent history and his connection to the area where the body was found. In denying the claim, the postconviction court first stated that there were no disputed issues of material fact. The court then concluded that Schofield failed to show due diligence because the defense was aware of the existence of the fingerprints prior to trial — an FDLE report of 1987 noted that the prints recovered from the car did not match Schofield, his father, or Michelle. The postconviction court rejected Scho-field’s argument that the Scott fingerprint identification result is newly discovered evidence. The court concluded that the result was not newly discovered evidence because it was not “in existence” at the time of trial. Alternatively, the court reasoned that if the evidence was in existence, defense counsel failed to exercise due diligence in discovering it. Last, the court concluded that the new evidence would probably not produce an acquittal on retrial.
The supreme court has enunciated the proper standard of review of a summary denial of a rule 3.850 claim, including a claim of newly discovered evidence, as follows:
“To uphold the trial court’s summary denial of claims raised in a 3.850 motion, the claims must be either facially invalid or conclusively refuted by the record. Further, where no evidentiary hearing is held below, we must accept the defendant’s factual allegations to the extent they are not refuted by the record.”
McLin v. State, 827 So.2d 948, 954 (Fla.2002) (quoting Foster v. State, 810 So.2d 910, 914 (Fla.2002)). The analysis that governs a newly discovered evidence claim is set forth in Green v. State, 975 So.2d 1090-100 (Fla.2008) (citing Jones v. State, 709 So.2d 512, 521 (Fla.1998) {Jones II)). In Green, the supreme court reviewed a newly discovered evidence claim after an evidentiary hearing.
To obtain a new trial based on newly discovered evidence, a defendant must meet two requirements: First, the evidence must not have been known by the trial court, the party, or counsel at the time of trial, and it must appear that the defendant or defense counsel could not have known of it by the use of diligence. Second, the newly discovered evidence must be of such nature that it would probably produce an acquittal on retrial. See Jones v. State, 709 So.2d 512, 521 (Fla.1998) (Jones II). Newly discovered evidence satisfies the second prong of this test if it “weakens the case against [the defendant] so as to give rise to a reasonable doubt as to his culpability.” Id. at 526 (quoting Jones v. State, 678 So.2d 309, 315 (Fla.1996) (Jones I)). In determining whether the evidence compels a new trial, the trial court must “consider all newly discovered evidence which would be admissible,” and must “evaluate the weight of both the newly discovered evidence and the evidence which was introduced at the trial.” Jones v. State, 591 So.2d 911, 916 (Fla.1991). This determination includes
whether the evidence goes to the merits of the case or whether it constitutes impeachment evidence. The trial court should also determine whether the evidence is cumulative to other evidence in the case. The trial court should further consider the materiality and relevance of the evidence *93and any inconsistencies in the newly discovered evidence.
Jones II, 709 So.2d at 521 (citations omitted).
Green, 975 So.2d at 1099-100.
On appeal, Schofield first argues that the postconviction court erred in focusing on whether the newly discovered evidence was in existence at the time of trial. Schofield discusses the postconviction court’s rebanee on Kearse v. State, 969 So.2d 976, 987 (Fla.2007), distinguishes Kearse, and argues that the evidence only “must have been unknown” at the time of trial. Schofield contends that the “must have existed” passage in Kearse was incorrect and he notes decisions since Kearse which have not included the requirement that the newly discovered evidence was in existence at the time of trial.
In Kearse, the appellant alleged that newly discovered evidence could impeach the state’s mental health expert and show that the expert gave biased testimony in favor of the state at resentencing. Id. at 987. The new evidence consisted of evidence about the doctor’s conduct as an expert witness for the federal government in New Mexico. Id. In rejecting the claim, the court stated that the evidence must have existed but been unknown, and the court concluded that this impeachment evidence did not meet the test:
Both Dr. Martell’s actions in the federal criminal case and aüegations regarding his conduct postdated Kearse’s sentencing. Thus, the evidence did not exist at the time of the resentencing, and Kearse fails to meet the first prong of the test.
Id. at 987. This situation differs substantially from the fingerprint identification situation in the present case. Here, the fingerprint match was not known at the time of the original trial because the analysis had not been performed, but this evidence does not equate to a future event that had not yet occurred, as in the Kearse case. It cannot be said that the match evidence did not exist at the time of the Schofield trial. In other words, if the fingerprints in the car can now be matched to Scott’s fingerprints, then that was also true at the time of the Schofield trial. We reject the postconviction court’s assertion that Kearse justifies the summary denial of the claim.
Schofield also argues that the postcon-viction court ignored the facts and precedent when it found no due diligence on his part. He asserts that the postconviction court imposed a burden on Schofield and his trial lawyer that is contradicted by law. He also claims he could not have discovered Scott’s identification before the identification was made by FDLE and disclosed to him in December 2004 and March 2005.
The postconviction court’s due dibgence analysis concludes that Schofield’s trial counsel should have requested the Polk County Sheriffs Office to perform a manual check of all prints on file from all felony arrests in Polk County. The postconviction court stated that “counsel’s simple request to have the unknown prints compared to any prints on file with the same, local law enforcement agency that had arrested Mr. Schofield could have revealed a match with Mr. Scott.” In his brief, Scho-field notes that at the time of trial he had no other specific suspect or similar murder case on which to base any request for a manual check of prints. We agree with Schofield that due dibgence cannot demand speculation by trial counsel as a prerequisite to the consideration of newly discovered evidence. See Cherry v. State, 959 So.2d 702, 708 (Fla.2007) (rejecting circuit court’s finding that defense could have discovered evidence earlier by due dibgence). We therefore reject the post-conviction court’s conclusion that Schofield failed to exercise due diligence.
*94Finally, we conclude that the postconviction court applied the wrong legal standard in determining that the new evidence would probably not produce an acquittal on retrial. Because the postconviction court failed to conduct an evidentiary hearing, it did not thoroughly analyze the newly discovered evidence. Taking all the facts put forth by Schofield in his motion as true leads us to conclude that the newly discovered evidence is of such a nature that it would probably produce an acquittal on retrial. The postconviction court’s attached trial excerpts have not convinced us otherwise.
Schofield argues on appeal that this court should accept all the facts put forth in his motion as true and conduct a de novo cumulative analysis to determine if his evidence meets the standard set out in Jones II, 709 So.2d at 521-26. He cites Haywood v. State, 961 So.2d 995 (Fla. 4th DCA 2007), among other cases, as authority for this court to conduct the cumulative analysis in the first instance. In Haywood, a sexual battery case, the newly discovered evidence was a DNA analysis of the semen found on the victim’s underwear. Haywood sought a new trial based on the fact that the DNA was that of the victim’s boyfriend. The postconviction court did not conduct an evidentiary hearing because the newly discovered facts were undisputed. The appeals court reversed the summary denial of Haywood’s rule 3.850 motion and remanded for a new trial after concluding that the evidence would probably produce an acquittal on retrial.
In the instant case, we decline to conduct the cumulative analysis of the facts de novo. Although the postconviction court accepted all of the alleged new evidence as undisputed, the court did not engage in evaluating the weight of both the newly discovered evidence and the evidence that was introduced at trial. The supreme court has noted that ordinarily an eviden-tiary hearing is required for the court to properly evaluate, in accordance with Jones II, whether newly discovered evidence is of such a nature that it would probably produce an acquittal on retrial. See McLin, 827 So.2d at 956. Because part of the new evidence in this case involves proffered testimony from witnesses as to the background of Scott, on remand the court must conduct an evidentiary hearing and make findings on the weight of the evidence, including “whether the evidence would have been admissible at trial, the purpose for which the evidence would have been admitted, the materiality and relevance of and any inconsistencies in the evidence.” Cherry, 959 So.2d at 705.
Grounds Two and Three: Failure to Preserve Evidence
Grounds two and three of the post-conviction motion allege facts relating to six hairs that were recovered from the victim’s car and on or near her body that the State has since lost or destroyed. It is alleged that six days after the body was found, on March 5, 1987, Detective Weeks, the lead homicide investigator, submitted certain physical evidence to FDLE for testing. The first submission included three hairs plucked from the passenger and cargo compartments of the victim’s car. On March 80, 1987, Detective Weeks submitted additional hairs for testing, including one recovered from the victim’s left foot, and requested that the hairs be tested and compared with head and pubic hair samples of the victim and with any human hair found in the first submission. In November 1987, the FDLE Microanaly-sis Section Supervisor returned the six items submitted, among other exhibits, with a letter stating that examination of the items was “no longer necessary.” Records from the Property and Evidence *95Unit of the sheriffs office indicate that some exhibits were checked out by a State Attorney investigator on March 3, 1989, and were not returned. The investigator, who died in 1996, habitually failed to return evidence that he logged out. In 1996, the assistant state attorney who prosecuted Schofield authorized the destruction of all property related to the Schofield case except for eight exhibits that he designated as “possibly needed in [the] future for evidence.” Deputy Sirera, who worked in the Property and Evidence unit, testified that the six hairs at issue, if returned from FDLE as “sweepings” or “debris,” were most likely destroyed.
The postconviction court disposed of grounds two and three by adopting the State’s response to the court’s order to show cause. In that response, the State first noted that it “does not contest that the hairs Defendant complains of were destroyed in 1995 on authorization of the lead prosecutor in this case.” The State contended that Schofield’s ability to bring a claim for postconviction evidence testing was procedurally barred in 1995. It contended that Schofield could have moved for the testing in 1989 prior to trial and in 1993 during the postconviction proceedings. It asserted that in 1995 the assistant state attorney and the Polk County Sheriff had no duty to preserve the evidence after Schofield had exhausted his appeals.
In Schofield’s reply to the State’s response, he argued that no procedural default or time bar applies to his constitutional claims. He asserted that the State’s duty to preserve evidence has nothing to do with exhaustion of appellate remedies, but with the State’s knowledge of the value or potential value of the physical evidence both before and at the time the State lost or destroyed the evidence. In arguing that the circumstances here lead to a conclusion of “bad faith,” Schofield contended that the State believed that the untested hairs would have formed a basis for exonerating Schofield well before the time that the assistant state attorney authorized the destruction of the hairs in 1996. He pointed to the fact that the State originally asked for all evidence to be tested by FDLE, but when some testing results were exculpatory, in August and September 1987, the State cancelled the testing of the hairs and other items. FDLE memorialized the cancellation in a November 1987 letter, and after November 1987 the State never reauthorized testing of the hairs even though Detective Weeks acknowledged that the State sought any possible evidence, especially scientific testing, that could link Schofield to the murder. Schofield asserted that the police or prosecutors feared that testing of the hairs would further exonerate or exculpate him. He asserted that the assistant state attorney knew in 1996 what he knew in 1987, that the six hairs existed and remained untested, but he did not exempt the hairs from destruction as “possibly needed in [the] future for evidence.”
On appeal, Schofield argues that because the State’s response did not dispute or challenge the facts contained in his claims two and three, he is entitled to relief on the merits on the record before this court. He reasserts the facts supporting his claim that the State believed that additional forensic testing could be exculpatory.
In Arizona v. Youngblood, 488 U.S. 51, 58, 109 S.Ct. 333, 102 L.Ed.2d 281 (1988), the Court addressed “the failure of the State to preserve evidentiary material of which no more can be said than that it could have been subjected to tests, the results of which might have exonerated the defendant.” Id. at 57, 109 S.Ct. 333. The Court held that “unless a criminal defendant can show bad faith on the part of the police, failure to preserve potentially use*96ful evidence does not constitute a denial of due process of law.” Id. at 58, 109 S.Ct. 333.
We conclude that the denial of this claim must be reversed for an evidentiary hearing because the State’s response does not refute Schofield’s allegations about the State’s knowledge of the potential value of the physical evidence. Taken in the light most favorable to Schofield, the circumstances alleged suggest a possibility of bad faith that is not refuted by the record. Schofield should be given the opportunity to prove his allegations of bad faith, and the postconviction court should make findings based on the evidence presented.
In ground three, Schofield asks that his Youngblood claim be evaluated under independent state constitutional grounds. He argues that almost a third of the states have rejected the Youngblood bad faith test, and he cites cases from a number of states that have found that the state’s loss or destruction of critical evidence violates a defendant’s state constitutional rights even without a showing of bad faith. He urges this court to adopt the reasoning of these other states and find that the failure of the State to preserve the six hairs made Schofield’s conviction fundamentally unfair as a matter of due process under the Florida Constitution. We reject these arguments because Schofield has not provided any persuasive basis for us to interpret the Florida constitution as providing a broader due process protection than that recognized in Youngblood. Therefore, we draw no distinction in the analysis governing grounds two and three.
Reversed and remanded for further proceedings.
KELLY and WALLACE, JJ., Concur.