MisleadingCharts
Back to the gallery

The rooftop-solar market the chart couldn’t see

Showing the misleading chart

Be the first to star this exhibit

A strategy slide we drew for an invented electrical contractor ranks all 50 states by solar’s share of their electricity, using EIA’s 2025 figures, and sends the new rooftop crew to Nevada because New Hampshire generates just 0.017% of its electricity from solar. Every number is EIA’s, but they come from surveys of power plants of 1 megawatt and up, and a home rooftop system is a few kilowatts. Add EIA’s own estimate for systems under 1 MW and New Hampshire’s 3.3 GWh of solar becomes 411 GWh, Massachusetts climbs from 11th to 3rd, and 24% of all U.S. solar turns out to have been below the line.

01The claim

The rooftop-solar market is in the desert Southwest: Nevada generates 30.5% of its electricity from solar and New Hampshire 0.017%, so a new rooftop crew belongs in Reno, not Concord.

02The trick

The numbers are EIA’s and the axis starts at zero; the problem is which installations the numbers cover. EIA’s plant-level surveys, Forms EIA-860 and EIA-923, collect only from plants with at least 1 megawatt of combined capacity. That line defines the surveys’ scope; it isn’t a statement about solar, but it means the slide counts solar farms and very little else. A home rooftop system is a few kilowatts, hundreds of times too small to file, so a rooftop installer is ranking states with data that leaves out nearly every system it would ever install, and a missing system reads as “no solar here”. North Dakota shows the pattern in miniature: EIA simply has no utility-scale solar row for it, and the slide calls that “no solar generation to report at all”. The states hurt most are the ones where rooftops dominate. In New Hampshire, the plant surveys show 3.3 GWh of solar in 2025 and EIA’s separate estimate for systems under 1 MW adds 408 GWh, so 99% of the state’s solar is missing from the slide. In New Jersey it is 70%, in Connecticut 72%, in Massachusetts 67%. The slide’s fine print even names the threshold (“plants of 1 MW and up”), which is the usual pattern: the scope is stated once, in the basis note, and then the chart is read as if it covered everything.

03The fix

Put the tier under the line back, and label it as an estimate. EIA publishes it alongside the plant data as “estimated small-scale solar photovoltaic”: small-scale capacity comes mainly from utilities’ monthly reports on Form EIA-861M, supplemented with data on third-party-owned systems, and is converted to generation with NREL solar models, so it is a modelled series, not metered output. Because EIA’s all-fuels total leaves the tier out, the redraw adds it to each state’s bar in its own colour and to the denominator, so shares are of all generation including systems under 1 MW (EIA estimates only solar below the line; other small generators stay uncounted). Nationally the under-1 MW tier made 93.0 TWh in 2025, 24% of all U.S. solar, and 61% in New England. The ranking reshuffles: Massachusetts goes from 11th to 3rd at 23.7%, level with Hawaii (up from 8th), New Jersey from 30th to 17th, and New Hampshire from 49th to 37th. A second panel ranks states by the small-scale series alone, which is closer to what a rooftop installer is asking: California leads with 34.4 TWh, then Arizona, Texas, Florida and New York, with Massachusetts 6th and New Jersey 7th, far above their places on the slide (11th and 30th), while Nevada, the slide’s first pick, is 8th. That panel measures the installed base, so big states lead; per household would be the next cut. And read any of these shares with care: a share is solar against what the state generates, not what it uses or how many roofs it has, and New Hampshire’s denominator includes the Seabrook nuclear plant, much of whose output leaves the state. The general habit: before ranking anything from a register, find its scope sentence and ask whether the thing you care about is above the line or below it.