"We saw that the increase grew with each subsequent time period, so that really speaks to the latent branding effects of earned media exposure, [similar to] TV advertising," said Andrew Lipsman, ComScore's VP-marketing.

Last summer, Facebook released data that indicated fans of Starbucks tended to spend more in store than non-fans. However, that study was criticized as not necessarily differentiating between correlation (people who spend more at Starbucks also tend to be Facebook fans of Starbucks) and causation (did their exposure to content resulting from being a Facebook fan of Starbucks cause them to spend more in store?). Mr. Lipsman emphasized that this study focused on causation.

As part of the same study, ComScore was commissioned by Facebook to look at the effect of exposure to premium Facebook ads by a major retailer (who ComScore would not identify) on purchase behavior. After a month, it found that 1.47% of the test group had made an in-store purchase of the brand, compared to 1.27% of the control group, and that 0.61% of the test group had made an online purchase, compared to 0.39% of the control group.

Mr. Lipsman was also dismissive of a Reuters/Ipsos poll released last week that 80% of Facebook users had never bought a product or used a service as a result of ads or comments they had seen on the platform, noting that people tend not to be self-aware about advertising's influence on their behavior.

"It's not a sound research method," he said. "If you ask most people, 'Does the advertising work on you?,' they're likely to say no."