How Accurate Is Conception Date Based on Ultrasound?

A first-trimester ultrasound can estimate gestational age to within roughly three to five days of the true value, which makes it the most reliable clinical tool available for dating a pregnancy. But “gestational age” and “conception date” are not the same thing, and the gap between them introduces a layer of uncertainty that many people do not expect. Ultrasound measures how big the embryo is, not when sperm met egg, so converting a size measurement into a specific conception date involves assumptions about ovulation timing, implantation speed, and early growth rates that vary from one pregnancy to the next.

What Ultrasound Actually Measures

When a sonographer dates your pregnancy, they are measuring the embryo or fetus and comparing that measurement to a reference chart of average sizes at each gestational age. In the first trimester, the key measurement is the crown-rump length, which is the distance from the top of the embryo’s head to its bottom. The result is reported as a gestational age in weeks and days, counted from the first day of the last menstrual period. This convention adds roughly two weeks to the actual time since conception, because the standard calendar assumes ovulation happened around day 14 of the cycle. So if your ultrasound says you are 8 weeks pregnant, the embryo itself is closer to 6 weeks old.

This two-week offset is not a flaw in the system. It is a deliberate convention that dates back decades, and every reference chart, due-date calculator, and clinical guideline is built around it. To get from a gestational age estimate to a conception date, you subtract about two weeks. That subtraction is straightforward, but it assumes ovulation happened on schedule, and for many people it does not.

Why Pinpointing the Exact Day of Conception Is Inherently Difficult

Even if ultrasound could measure gestational age with perfect precision, converting that number into a single conception date would still be tricky. Conception does not necessarily happen on the day of intercourse. A landmark study tracking couples who were trying to conceive found that pregnancies resulted from intercourse that took place anywhere within a six-day window ending on the day of ovulation. The probability of conception ranged from about 10 percent when intercourse occurred five days before ovulation to 33 percent on ovulation day itself.¹ Sperm can survive in the reproductive tract for several days. Estimates put the average survival time at about 1.4 days, with a small percentage of sperm surviving more than four days.²

This means that even if you know exactly when you had intercourse, the actual moment of fertilization could have occurred a day or two later, after the egg was released and the surviving sperm reached it. Ultrasound cannot distinguish between a pregnancy conceived from intercourse five days before ovulation and one conceived on ovulation day, because both embryos are the same gestational age. The practical upshot is that ultrasound can narrow the conception window to roughly a week, but it cannot reliably identify the specific sexual encounter that led to pregnancy.

First-Trimester Accuracy and the Best Window for Dating

The gold standard for ultrasound dating is a crown-rump length measurement taken between about 7 and 13 weeks of gestation. During this window, embryos grow at a remarkably consistent rate, and individual genetic differences in size have not yet kicked in. An international study developing standardized size charts found that crown-rump length tracks gestational age in a nearly linear fashion during this period, with a standard deviation of only a few days.³

An older but well-regarded study that used pregnancies conceived through artificial insemination, where the exact insemination date was known, found that the best mathematical models predicted conceptional age from crown-rump length with the 5th-to-95th percentile range spanning about 3.7 days above and 7.3 days below the regression line.⁴ In plain terms, most predictions fell within a few days of the true conception date, but the spread was not perfectly symmetrical. Ultrasound was slightly more likely to overestimate gestational age (suggesting conception happened a bit earlier than it actually did) than to underestimate it.

What IVF Pregnancies Reveal About Accuracy

The most powerful way to test ultrasound dating is to compare it against pregnancies where the conception date is known with certainty, because an embryo was transferred on a documented day. These validation studies consistently show that first-trimester ultrasound is impressively close to the truth, but not perfect.

A large study comparing first-trimester dating scans against the known embryo-transfer date in IVF pregnancies found that ultrasound was accurate to within roughly a day and a half on average. Accuracy was best in term births and slightly lower in preterm, small-for-gestational-age, and multiple pregnancies, but still typically within two days.⁵ Another IVF-based study, however, identified a consistent tendency for ultrasound to overestimate gestational age by about three days compared with the true fertilization date.⁶ This systematic bias was present regardless of which reference chart the clinicians used, suggesting it is baked into the charts themselves rather than caused by operator error.

A third IVF study compared two first-trimester measurements and found that biparietal diameter (the width of the baby’s head) had a slightly smaller average error and tighter agreement with the known IVF gestational age than crown-rump length did.⁷ The differences were small enough that both measurements perform well in clinical practice, but the finding hints that head measurements may have a slight edge in certain situations.

Taken together, these studies suggest that in the first trimester, ultrasound typically places the conception date within a window of about three to five days of the real date. That is close enough to be clinically useful for due-date estimation, prenatal screening, and growth monitoring. It is not close enough to determine, say, which week of a vacation led to a pregnancy, let alone which specific encounter.

Why Some Embryos Measure Bigger or Smaller Than Expected

If all embryos grew at exactly the same rate, first-trimester dating would be nearly perfect. But they do not. Research using IVF pregnancies with known implantation windows has shown that the timing of implantation itself is a major source of variation. Embryos that implant early tend to measure larger at the 10-to-14-week scan, and embryos that implant late tend to measure smaller, even when they are growing at the same rate after implantation.⁸ Since ultrasound cannot tell when implantation occurred, it reads the bigger embryo as slightly older and the smaller one as slightly younger. This is the largest single source of the few-day spread in first-trimester dating.

Even among pregnancies with the same known transfer date in IVF, crown-rump length and gestational sac size vary significantly at a given gestational age.⁹ Some of this variation may trace to maternal factors. A recent study found that inadequate folic acid supplementation and low serum folate levels were associated with delayed embryonic development and reduced growth in the first trimester.¹⁰ If an embryo is slightly smaller because of a nutritional factor rather than because it is younger, the ultrasound will underestimate gestational age and push the estimated conception date later than it actually was.

Accuracy Drops Sharply After the First Trimester

The comfortable three-to-five-day accuracy window is a first-trimester privilege. As pregnancy progresses, individual genetic variation in fetal size becomes a much larger factor, and dating accuracy degrades quickly. In the second trimester, head measurements remain the best predictors of gestational age, outperforming femur length and abdominal circumference, but the margin of error widens to roughly one to two weeks.¹¹ One study found that biparietal diameter-based estimates in the second trimester varied by only two to three days from the last-menstrual-period-based age, but that agreement deteriorated substantially later in pregnancy.¹²

By the late third trimester, the picture gets much worse. At 38 to 42 weeks, all dating formulas show large systematic underestimation, with average errors in the range of one and a half to three weeks. Even the best-performing models have 95 percent confidence ranges of plus or minus three weeks or more at this stage, driven by inherent biological variability in how big different babies are near term.¹³ This is why a third-trimester scan is essentially useless for working backward to a conception date. A baby who measures “average for 37 weeks” might actually be a large 35-weeker or a small 39-weeker.

Measurement Technique Adds Its Own Uncertainty

Beyond biological variability, the physical act of taking the measurement introduces error. A study examining how consistently sonographers could reproduce crown-rump length measurements found that even when the same operator measured the same embryo twice, the results could differ by about ±2 millimeters. Between two different operators, the spread widened to roughly ±2.4 millimeters for crown-rump length.¹⁴ At early gestational ages, when the embryo is small and growing quickly, a couple of millimeters of measurement error can translate to a day or two of dating error. At later gestational ages, the same millimeter error matters less in proportional terms, but by then the biological variability has already made dating imprecise for other reasons.

Operator skill, ultrasound machine quality, and the position of the embryo all play a role. A squirming embryo that is curled up or positioned at an awkward angle may be harder to measure accurately. This is one reason clinicians typically take multiple measurements and average them, and why standardized training in measurement technique matters for getting consistent results across different clinics.

Ultrasound Versus Last Menstrual Period

For most pregnancies, the two main tools for estimating a due date are the date of the last menstrual period and a first-trimester ultrasound. These often agree within a few days, but when they disagree, which one should you trust? The evidence consistently favors ultrasound. A study comparing both methods against actual delivery dates found that in over half of cases the ultrasound-derived due date was more accurate, and the researchers concluded that once a satisfactory scan has been obtained, clinicians should rely on it rather than the last-menstrual-period estimate.¹⁵

A more recent study that had the advantage of knowing the actual ovulation date (through fertility tracking) gave a nuanced picture. When compared against ovulation-based gestational age, ultrasound tended to slightly underestimate pregnancy length, doing so by at least a week in about 4 percent of cases. The last menstrual period, meanwhile, tended to overestimate, doing so by at least a week in nearly 9 percent of cases.¹⁶ Both methods had their biases, but the last menstrual period was roughly twice as likely to be off by a clinically meaningful amount. This is largely because many people do not ovulate on day 14, and even a few days of irregular cycle length throws the LMP calculation off.

Clinical guidelines reflect this evidence. The American College of Obstetricians and Gynecologists recommends that once a first-trimester ultrasound has been performed, changes to the estimated due date should be reserved for rare circumstances.¹⁷ Canadian guidelines similarly emphasize that accurate dating reduces unnecessary post-dates labor inductions and allows better timing of prenatal screening tests.¹⁸

Dating Twin Pregnancies

A common concern is whether standard singleton dating charts work for twins. A study of IVF twin pregnancies found that crown-rump length measurements in twins straddled those of singletons. The larger twin’s gestational age was overestimated by about 2.4 days and the smaller twin’s by about 0.8 days compared with the known IVF date. These differences were within the normal accuracy limits of first-trimester ultrasound and were considered clinically insignificant.¹⁹ So singleton charts can be used for twins in the first trimester without introducing meaningful extra error, though clinicians typically use the larger twin’s measurement for dating purposes.

Why the Accuracy of the Conception Date Estimate Matters

For most people asking “how accurate is my conception date,” the question is not purely academic. There are several real-world scenarios where the answer matters. The most common is due-date planning. A few days of error in conception dating translates directly into a few days of error in the estimated due date. Since clinical decisions about induction of labor for post-dates pregnancies hinge on whether you have passed 41 or 42 weeks, an error of even a few days can be the difference between watchful waiting and being scheduled for an induction. More accurate dating reduces the number of pregnancies incorrectly labeled as post-dates, which in turn reduces unnecessary interventions.

Prenatal screening is another area where precision matters. Tests for chromosomal conditions like Down syndrome are calibrated to specific gestational ages. If the dating is off by a week, the risk calculation can be thrown off enough to generate a false alarm or, worse, miss a true positive. This is one of the strongest arguments for getting a first-trimester dating scan rather than relying on menstrual dates alone.

Then there is the question that many people are actually asking when they search this topic: paternity. If someone had intercourse with two different partners within a short window, they may hope that ultrasound dating can identify which encounter led to the pregnancy. The honest answer is that ultrasound is not precise enough for this. A three-to-five-day margin of error, combined with the up-to-six-day window during which sperm can survive and lead to conception, means that ultrasound cannot reliably distinguish between sexual encounters that were less than a week or two apart. DNA testing is the only reliable way to establish paternity.

AI-Assisted Measurement and the Future of Dating Accuracy

One promising avenue for squeezing a bit more precision out of first-trimester dating is artificial intelligence. Researchers have developed automated systems that use 3D ultrasound images to measure crown-rump length without relying on a human operator to place the measurement calipers. A recent comparative study found that a 3D AI-enhanced model achieved expert-level performance in crown-rump length measurement while being faster and more consistent than manual methods.²⁰ By reducing the operator-dependent component of measurement error, these tools could tighten the overall accuracy of dating, though they cannot eliminate the biological variability in embryo growth that accounts for the largest share of the remaining uncertainty.

Metabolic and blood-based approaches to gestational age estimation are also under investigation. The same validation study that tested first-trimester ultrasound against known IVF dates also evaluated a metabolic algorithm and found it performed comparably to ultrasound in certain subgroups.⁵ These methods could eventually serve as a complement or cross-check, particularly in settings where early ultrasound is not available.

Historical Use of Pregnancy Dating in Paternity Disputes

Before DNA testing became available, courts relied on pregnancy duration to adjudicate paternity claims, and the margins they accepted were strikingly wide. In Norway, the Forensic Medical Commission set the lower limit for a full-term pregnancy at 240 days after conception and the upper limit at 320 days, a span of more than 11 weeks.²¹ That 1914 statement was cited in court rulings all the way into the 1970s. The enormous range reflects the genuine biological uncertainty in pregnancy timing, a range that ultrasound has narrowed considerably but not eliminated. It also serves as a useful reminder that human pregnancy length is inherently variable, and no imaging technology can compress that variability down to a single date with absolute certainty.