delivered the opinion of the court:
The defendant, Christine Buckley, n/k/a Christine Wolfe, was charged by indictment with the offense of involuntary manslaughter (720 ILCS 5/9 — 3(a) (West 1992)). The indictment alleged that the defendant acted recklessly and in a manner likely to cause death or great bodily harm when she gave a generic form of Dimetapp (a nonprescription decongestant) to her 12-month-old daughter, Taylor Buckley, in an amount that caused the infant’s death. A Henry County jury found the defendant guilty, and the circuit court sentenced her to a prison term of four years. No post-trial motions were filed.
The defendant raises several issues on appeal. This court, however, will primarily address the issue of prosecutorial misconduct in closing arguments. We reverse the defendant’s conviction and remand for a new trial on that basis.
The record on appeal reveals the following facts relevant to our ruling. Taylor Buckley was found dead in her crib early on the morning of March 12, 1992. Emergency medical technician Michael Size-more testified that he examined the infant at the scene and found no signs of external injuries. He did observe a small amount of dried blood on her upper lip, which he stated is often associated with Sudden Infant Death Syndrome (SIDS). In addition, he found signs that indicated to him that the infant had been dead for some time. He further testified that the defendant appeared to be in a state of total disbelief.
Dr. Louis Garcia performed an autopsy March 12, 1992. He testified by video evidence deposition that he found several factors consistent with SIDS. He noted the absence of medication in the stomach and further testified that tests on blood, fluids, and tissue samples revealed no abnormalities. Based upon these findings, Dr. Garcia listed SIDS as the cause of death. During the autopsy, several fluid and tissue samples, including blood from the heart, were taken for shipment to Baylor University for analysis in a national study on SIDS *84victims. These samples were not taken for evidentiary purposes and therefore were not subjected to chain of custody precautions associated with samples to be used as evidence, nor were any precautions taken to prevent contamination or evaporation.
On March 13, 1992, a coroner’s report was issued concluding that Taylor had died from SIDS. On March 23, 1992, Taylor’s body was exhumed and a second autopsy was performed by Dr. Mary Jumbelic, a forensic pathologist with the Peoria County coroner’s office. This second autopsy did not reveal anything unusual. Dr. Jumbelic requested, however, that Baylor University return blood and tissue samples for forensic analysis. She also took several samples for further testing during the second autopsy.
On March 31, 1992, at Dr. Jumbelic’s request, Baylor University sent the fluid and tissue samples to Dr. Eric Frow, director of the Peoria Tazewell Pathology Group, for toxicology screening. Dr. Frow testified that his tests revealed the presence in the blood samples of brompheniramine (BPA), an antihistamine found in Dimetapp. His tests, however, did not reveal the presence of phenylpropanolamine (PPA), a stimulant also found in Dimetapp. Dr. Frow testified that he sent three milliliters of blood to the National Medical Services Laboratory (NMS) in Willowgrove, Pennsylvania, for further analysis. Tissue samples were sent directly from Baylor University to NMS at the request of Dr. Jumbelic.
Dr. Robert Middleberg, a forensic toxicologist and director of NMS, testified that on April 8, 1992, he received from Dr. Frow a clear vial containing approximately two milliliters of blood. Tests performed on April 14, 1992, indicated that the blood sample contained 990 nanograms (parts per billion) per milliliter (ng/ml) of BPA and 6,300 nanograms of PPA. Dr. Middleberg testified that these concentration levels were determined partially by computer-generated results and partly by mathematical calculations done by himself and a lab technician. Dr. Middleberg also testified that tests on the tissue samples received from Baylor University detected the presence of BPA; however, it was not possible for him to determine the level of concentration.
Dr. Middleberg stated that BPA levels he found could have acted as a stimulant in a small child. He opined that although BPA acted as a depressant, in small children it could have a "paradoxical effect” in which it would act as a stimulant. BPA intoxication could, according to Dr. Middleberg, cause nervousness, shaking, cardio effects and seizures. He also stated that, in the levels he found, BPA could be expected to show some type of toxic side effects and could be a potential cause of death, in the absence of a more likely cause.
*85Dr. Middleberg could not state that BPA intoxication was the cause of Taylor’s death, as that conclusion was beyond his area of expertise. He also could not state an opinion as to how much Dimetapp Taylor had been given, nor could he state an opinion as to how much Dimetapp would need to be administered in order to produce a BPA level of 990 in the sample he tested.
Dr. Middleberg also acknowledged that after death, BPA can leach into the heart blood, thereby producing an artificially high concentration, and he acknowledged being aware of one study specifically involving BPA that showed concentrations of that chemical in the heart blood to be more than twice the level found in samples taken from other parts of the body. He did not believe, however, that the results of one study were scientifically significant. It was uncontroverted, however, that the blood sample Dr. Middleberg tested was taken by Dr. Garcia directly from Taylor’s heart.
Brian Donnelly, a toxicologist at the Federal Bureau of Investigation (FBI) laboratory in "Washington, D.C., testified that he received fluid and tissue samples from Dr. Jumbelic on June 26, 1992. Donnelly stated that he tested only one tissue sample and found BPA, but no PPA. He also acknowledged that he could not test for the level of concentration of BPA in the sample he tested.
On March 4, 1993, Dr. Jumbelic again sent fluid and tissue samples to Dr. Middleberg at NMS. Dr. Middleberg tested one-half milliliter of blood, this time finding 1,300 ng/ml of BPA and 6,300 ng/ml of PPA. Dr. Middleberg testified that there is a "built in error” of 10% that could partially explain the increase in BPA levels over the previous tests. In addition, he noted that as water evaporates from blood, an increase in the concentration level of BPA would be expected.
Based entirely upon the NMS report showing a BPA concentration of 990 ng/ml in the blood sample tested, Dr. Jumbelic opined that Taylor died from BPA intoxication. Dr. Jumbelic acknowledged that she did not perform the tests herself and relied upon the accuracy of the NMS tests. She also acknowledged that she did not know how much Dimetapp would have to be administered to produce a BPA level of 990 ng/ml, nor did she know how much Dimetapp had been given to Taylor. She recognized that no studies had been done to determine the therapeutic dosage of BPA for children, nor the rate at which children metabolize BPA, PPA or alcohol, another ingredient in Dimetapp.
Dr. Jumbelic testified that she was aware of one case reported in research literature where an adult died after taking a large amount of Drixoral, which also contains BPA. She noted that the BPA levels *86in that subject were four to five times less than the 990 ng/ml reported by Dr. Middleberg. She also testified that she was aware of one confirmed death of a child with a BPA level of 400 ng/ml. Dr. Jumbelic speculated, based upon these cases and the NMS test, that Taylor had received five times the lethal level and 50 times the therapeutic level of BPA.
Dr. John Bederka, defendant’s toxicology expert, opined that the test results from NMS were unreliable due to poor testing techniques and sample preparation, possible errors in calculations and data manipulation, lack of certain tests which should have been performed, and results which were inconsistent with tests performed by Dr. Prow on the same sample. He noted particularly the possibility that evaporation of the sample used in the NMS test could have seriously affected the reliability of the test.
Jan Buckley, the defendant’s former husband and Taylor’s father, testified he and the defendant separated immediately after Taylor’s birth and were divorced several months later. He testified that he once saw the defendant give Dimetapp to her other children in February 1990, even though the children did not appear to have a cough or cold at that time. He testified that she told him that it made them sleep. He noted that he only saw this behavior one time, and the defendant never indicated to him that she gave Dimetapp to Taylor in order to make her sleep.
Buckley also testified that on one occasion in August 1991 the defendant left Dimetapp and liquid Tylenol with him when she dropped Taylor off for visitation. He also noted that on two or three occasions in November and December 1991, the defendant brought Taylor and told him that she had put medicine in the bottle, which he assumed was Dimetapp because it had a grape smell. Buckley also acknowledged that he administered Dimetapp to Taylor when she was in his care.
Joe DeSmit, who saw the defendant three or four times per week in the months prior to Taylor’s death, testified that he saw the defendant give her older children a teaspoon of Dimetapp on more than one occasion. He stated that the defendant told him that "it helps put them to sleep and gets them out of [her] hair and calms them down.” He acknowledged, however, that on those occasions when he saw the defendant administer Dimetapp to the children, they appeared "to have the sniffles.”
Linda Buckley, defendant’s ex-sister-in-law and Taylor’s aunt, testified that she had warned the defendant in June 1991 that it was dangerous to give an infant over-the-counter medication without first consulting a doctor, and that if too much medicine was given the *87child could stop breathing and die. Ms. Buckley also testified that the defendant told her she did not realize the dangers and promised not to do it any more.
The defendant testified that Taylor had been sick with the flu two days prior to her death. On the evening before Taylor died, the defendant heard the child wake up in a "fuss.” When she checked on Taylor, the defendant noticed a runny nose, so she gave her a bottle containing a mixture of one-quarter teaspoon of Dimetapp and two ounces of water. She measured the Dimetapp with the use of a medicine dropper.
The defendant also testified that her pediatrician, Dr. Kidd, had previously told her it was appropriate to give Dimetapp to a one-year-old child, and she had learned to administer it in the baby’s bottle from an article in Parents magazine. She denied ever using medicine as a sedative, and she testified that the only conversation she had ever had with Linda Buckley concerning medication was about ear drops for Taylor’s ear infection.
After closing arguments, the jury returned a verdict of guilty to the charge of involuntary manslaughter.
The defendant maintains on appeal that she was denied a fair trial when the prosecutor, during closing argument, misstated the law regarding the requisite mental state for involuntary manslaughter by improperly equating recklessness with carelessness and suggesting to the jury that it could ignore the court’s instructions defining the proper mental state. During closing argument, the People stated as follows:
"[IJn involuntary manslaughter cases recklessness means heedless or being careless and when you’re careless you do something because you disregard what’s obvious or evident or disregard what a reasonable person wouldn’t disregard. *** The judge is going to explain a lot better than I can, but recklessness means careless. Was this defendant careless when she gave Dimetapp to her child? And look at all the evidence that we have that goes to show that she was careless.” (Emphasis added.)
During rebuttal closing argument, the People again stated:
"Recklessness is defined and there’s a definition there of conscious disregard, but in essence it comes down where somebody is careless, and when you’re dealing with a small child you can’t do that.” (Emphasis added.)
The People argue in response that the defendant has waived this argument by failing to object to these statements at trial. The People point out that in order to properly preserve a matter for review, a trial objection and a written post-trial motion specifically articulat*88ing the issue sought to be raised is required. People v. Enoch, 122 Ill. 2d 176, 186 (1988). The defendant acknowledges this fact, but urges this court to address the issue under the plain error exception to the waiver doctrine.
Plain error encompasses those errors which are obvious, which affect substantial rights of the accused, and which, if uncorrected, would be an affront to the integrity and reputation of the judicial system. People v. Campbell, 264 Ill. App. 3d 712 (1992). The plain error rule applies where evidence is closely balanced or where error is of such magnitude that the commission thereof has denied the accused a fair and impartial trial or may have significantly affected the outcome of the case. People v. Ward, 154 Ill. 2d 272 (1992).
After a careful review of the record, we believe that the evidence in this case is closely balanced. There were two contested issues at trial, i.e., whether the defendant administered a lethal dose of Dimetapp, and whether she did so in conscious disregard of the knowledge that her actions would likely cause death or great bodily harm. The evidence on both these issues is closely balanced.
The People maintain Dr. Jumbelic’s expert medical opinion established the cause of death as BPA intoxication and, based on NMS test results, that Taylor must have received a fatal dose of Dimetapp shortly before her death. This evidence is not unrefuted, however. First, we note that Dr. Garcia, who originally diagnosed SIDS as the cause of death, persisted in that diagnosis even after becoming aware of the second autopsy and the subsequent lab reports. We also note that Dr. Jumbelic’s opinion was based entirely on Dr. Middleberg’s NMS test results showing a BPA level of 990 ng/ml. The defendant, through the testimony of Dr. Bederka, called into question Dr. Middleberg’s testing methodology, the reliability of the sample tested and the calculations of the BPA and PPA concentration levels. We also find it significant that Dr. Middleberg could not state an opinion that BPA intoxication caused Taylor’s death; rather, he could only state that it might be a cause of death "absent a similar or more competent cause.”
The evidence concerning whether the defendant administered Dimetapp in a reckless manner is also closely balanced. The People presented evidence through the testimony of Jan Buckley and Joe DeSmit that could establish that the defendant had a habit of using Dimetapp as a sedative for her children. However, both witnesses acknowledged that the children seemed to be in need of the therapeutic effect of the medicine when they observed the defendant administer the medicine to her children. Likewise, the defendant denied ever using Dimetapp as a sedative for any of her children and *89testified that she gave Taylor only one-quarter teaspoon of the medication on the evening in question.
Linda Buckley’s testimony that she specifically warned the defendant of the potentially lethal effect of administering Dimetapp to an infant could establish that the defendant possessed knowledge of the likelihood of death or great bodily harm from an overdose of Dimetapp. However, the defendant denied that conversation took place, and as there are no other witnesses to that extremely important conversation, we must conclude that the evidence there was closely balanced as well.
Having found that the evidence in this case was so closely balanced as to invoke the plain-error rule to review the defendant’s claim that prosecutorial misconduct denied her a fair trial, we must now turn to the merits of her argument. In reviewing allegations of improper closing argument, the People’s argument must be examined in its entirety and the complained-of comments must be placed in proper context. People v. Morgan, 142 Ill. 2d 410, 453 (1991).
During closing argument, the People are barred from misstating the law or facts of the case, from making remarks that diminish their burden of proof, or from commenting on factual matters not based on evidence. People v. Roach, 213 Ill. App. 3d 119, 124-25 (1991). While a prosecutor has wide latitude in closing argument, he must not make comments which misstate the law in the case. People v. Gutirrez, 205 Ill. App. 3d 231 (1990), appeal denied, 136 Ill. 2d 548 (1991).
We find that the People misstated the law as to the requisite state of mind for the commission of the offense of involuntary manslaughter. In order to sustain a conviction for involuntary manslaughter, the People were required to prove that the defendant acted in a reckless manner. A person acts recklessly when he or she consciously disregards a substantial and unjustifiable risk and such disregard constitutes a gross deviation for the standard of care which a reasonable person would exercise in the situation. 720 ILCS 5/4 — 6 (West 1992).
It is well settled that an act performed accidently, carelessly, or even negligently is insufficient to prove or sustain a conviction for involuntary manslaughter. People v. Spani, 46 Ill. App. 3d 777, 780 (1977); People v. Schneider, 360 Ill. 43 (1935). When the People equated recklessness with carelessness, they clearly misstated the requisite mental state required to support an involuntary manslaughter conviction.
Having found that the People misstated the law in closing argument, our inquiry cannot end at that point. A misstatement of *90the law during closing argument does not normally constitute reversible error if the circuit court properly instructs the jury on the law, as counsel’s arguments are construed to carry less weight with the jury than do instructions from the circuit court. People v. Lawler, 142 Ill. 2d 548, 564-65 (1991). However, improper arguments that result in substantial prejudice to the defendant can constitute reversible error. People v. Tiller, 94 Ill. 2d 303 (1982). Although the circuit court in the matter sub judice properly instructed the jury on the definition of recklessness, the prejudicial effect of an improper argument cannot always be erased from the minds of the jury by an admonishment from the court. People v. Garreau, 27 Ill. 2d 388, 391 (1963).
The test to determine whether the People’s misstatement of the law constituted substantial prejudice to the defendant is whether the jury would have reached a contrary verdict had the improper statement not been made. People v. Burnside, 212 Ill. App. 3d 605, 607 (1991). Although the Burnside court relied upon objective evidence that the jury would have reached a contrary verdict to find prejudice, this court has subsequently relied upon the closely balanced nature of the evidence to conclude that a defendant was prejudiced by a prosecutor’s misstatement of the law during closing argument. People v. Gutierrez, 239 Ill. App. 3d 536 (1992); People v. Dunsworth, 233 Ill. App. 3d 258, 269-70 (1992).
In Dunsworth, this court held that the prosecutor’s misstatement of the law prejudiced the defendant:
"We find the evidence in this case was closely balanced. We cannot say with a reasonable degree of certainty that the prosecutor’s improper remarks did not contribute to [the defendant’s] guilty verdict. Also, we cannot say that the jury could not have reached a contrary verdict had the improper remarks not been made. Therefore, the judgment of the trial court is reversed, and the cause is remanded for a new trial.” 233 Ill. App. 3d at 270.
In the matter sub judice, having found the evidence to be closely balanced, we cannot say with a reasonable degree of certainty that the prosecutor’s improper equating of recklessness and carelessness in the minds of the jury did not contribute to the defendant’s guilty verdict. If the defendant was responsible for Taylor’s death, whether her responsibility was based upon carelessness or recklessness was crucially important to a finding of criminal liability, and the prosecutor’s statements could only serve to confuse that important distinction. We cannot say, therefore, that the jury could not have reached a contrary verdict had the improper equating of these two concepts not been made. Therefore, we must reverse the judgment of the trial court and remand this matter for a new trial.
*91In order to assist the trial court on remand, we will address the defendant’s argument that the trial court erred in denying her motion to exclude from evidence the samples taken by Dr. Garcia during the first autopsy. The defendant argues that the samples should have been excluded based upon an improper chain of custody. We disagree.
The trial court has broad discretion in determining the admissibility of evidence, and its discretion will not be reversed on appeal absent an abuse of discretion which prejudiced the defendant. People v. Hermann, 180 Ill. App. 3d 939, 944 (1988). To establish a sufficient chain of custody, the People need not disprove every possibility that the evidence was tampered with. Rather, they need only show that it was reasonably probable that the evidence remained unchanged in any important respect, or was not substituted. People v. Witanowski, 104 Ill. App. 3d 918, 925 (1982). Unless the defendant provides actual evidence of tampering or substitution, the People need only establish the stated probability, and any deficiencies go to the weight to be accorded the evidence and not its admissibility. People v. Shiflet, 125 Ill. App. 3d 161, 178 (1984).
In this matter, the defendant failed to introduce actual evidence that the sample had been tampered with or substituted. Rather, the defendant suggested that the sample was not properly handled in order to protect it from evaporation. The defendant merely challenges the reliability of the sample, which goes to the weight, not the admissibility of the evidence. We find, therefore, that the trial court properly admitted the challenged evidence.
Finally, we address whether there was sufficient evidence admitted at trial for the jury to conclude that the defendant was guilty beyond a reasonable doubt. We conclude that the evidence presented at trial was sufficient for a jury to decide that the defendant was guilty beyond a reasonable doubt. This does not mean that we are making a finding as to the defendant’s guilt or innocence which will be binding in a new trial, but rather our consideration of the evidence admitted at trial will protect the defendant’s constitutional right against double jeopardy. People v. Taylor, 76 Ill. 2d 289, 309-10 (1979); People v. Hammock, 121 Ill. App. 3d 874, 882-83 (1984).
For the reasons stated above, the defendant’s conviction is reversed and the case is remanded for a new trial.
Reversed and remanded.
McCUSKEY and SLATER, JJ„ concur.