The state once used this measure to calculate coronavirus test positivity. Here’s what it says now

Coronavirus testing in Framingham earlier this month. Two different ways of calculating test positivity offer two different pictures of the coronavirus pandemic in Mass.Jessica Rinaldi/Globe Staff

The state has used two different ways to look at the state’s coronavirus test positivity rate. This chart shows the results from each — and illustrates why some experts say the state needs to be cautious in going forward with reopening.

The state’s closely watched daily COVID-19 data dashboard highlights a measure that simply divides the number of positive tests by the number of total tests administered.

Every day on Page 2 of its report, the state recalculates a weekly average, summing the number of positive tests for the past seven days and dividing them by total tests administered. That measure has held steady at roughly 1 percent for weeks.

But the overall number of negative tests each day has dramatically increased due to a relatively recent surge in the repeated testing of asymptomatic people, such as students on college campuses. That has played a role in driving this rate down.

A different story emerges when you remove the repeated tests from the equation and show the rate of positive tests per people tested, rather than positive tests per total tests administered. The seven-day rate of positive tests per people tested, according to data available elsewhere on the state’s website, has been climbing since late August, reaching 3.2 percent in recent days.

(The daily rate of positive tests per people tested is available to the public on Page 8 of the dashboard, but the weekly rate for that metric is not shown.)

The state switched from highlighting the positive tests per people rate to the positive tests per total tests rate in mid-August. Initially, there was little difference between the two rates but the gap has been widening — as shown on the chart.

The Globe reported Tuesday that some experts say the positive test rates per people metric — the number the state does not showcase in the dashboard — is a better reflection of the scope of the pandemic in Massachusetts today.

“Both [metrics are] somewhat informative, but the individuals one is more indicative of what’s happening now, probably,” said Marc Lipsitch, an epidemiology professor at Harvard T.H. Chan School of Public Health, and a member of Baker’s COVID-19 commission.

“From a disease transmission perspective, from a risk perspective, this calculation that they’re leading with [on the COVID-19 dashboard] is likely a substantial underestimate,” said Samuel Scarpino, a Northeastern University epidemiologist. “If you have 10,000 tests a day coming from the colleges and universities who all have a very low percent positivity, then it’s very easy to end up with a biased positivity for the state.”

In a statement, the state did not directly address questions about why it chooses to highlight the lower positivity rate, and whether the uptick in the rate of positive tests per people tested is cause for concern.

Tory Mazzola, a spokesperson for the state’s COVID-19 command center, in an e-mail, touted the state’s low rate of positive tests per tests administered, noting it had hovered at low levels for more than a month.

“The Command Center closely monitors state and community level COVID-19 data, including for the Commonwealth’s highest risk municipalities, and continues to look for trends across multiple weeks of data to ensure a snapshot is in fact a trend," Mazzola said.

Martin Finucane of the Globe staff contributed to this report.