The public isn’t buying Robinhood’s excuse for shutting down trading in stocks like GameStop earlier this year.

Despite the fact that Robinhood CEO Vlad Tenev went before a congressional committee to explain that capital requirements caused his firm to stop, then throttle trades in companies widely traded by Reddit’s WallStreetBets community, a recent poll of more than 1,300 Americans from pollster Invisibly found that more than three-quarters of respondents, 77%, thought the shutdown itself was a form of market manipulation.

Just 23% of the respondents thought the shutdown was intended to protect people or the markets.

“Robinhood’s brand certainly took a major hit after deciding to halt trading on GameStop and other stocks in order to protect hedge funds from continuing to lose money on their short sales,” wrote Invisibly in discussing its results. “An overwhelming number of Americans—even those who have no connection to the stock market at all—found that Robinhood was guilty of market manipulation through their actions. It’s hard to believe that there could be a bigger loser in this whole event than the hedge funds who lost tens of millions on the deal, but Robinhood may have done serious long-lasting damage to their brand.”

Still, when asked whether they thought Robinhood controlled trading in GameStop and other stocks to protect hedge funds, most of the respondents, 55%, said no.

Additionally, the respondents expressed affinity for the WallStreetBets short squeeze campaign, which inflicted billions of dollars of losses on some hedge funds. Two-thirds of poll respondents, 67%, had some level of positive sentiment about the GameStop frenzy, with 39% expressing that it was exciting and good for everyday investors, and 28% thinking it was good to shake up the traditional way that financial markets operate.

Current investors were even more positive on the short squeeze than non-investors; 41% of current investors viewed the frenzy positively, while only 36% of non-investors did.

Invisibly’s poll was fielded from February 2 to February 5 among 1,352 people using its Realtime Research tool.