How much social-media activity actually causes a lift in TV ratings is still debatable.

Now, Nielsen is promoting a new use for TV-related Twitter data: The researcher says correlating tweet activity with on-air promotions can let networks and advertisers better estimate a show’s eventual ratings.

In a study published Monday, Nielsen analyzed 42 broadcast and cable series premieres from late August through early November 2014. For viewers 18-34, programs that were more heavily promoted achieved higher TV audiences when they premiered — no surprise there. (Nielsen didn’t ID the shows in the study.)

However, Nielsen created a model that it said is more accurate in predicting ratings. That combined promo data with Twitter TV activity over a four-week period ending two weeks prior to a show’s premiere (and adjusting for whether the show was on a broadcast or cable net).

Of course, nothing can really predict the future will 100% certitude. Nielsen’s Twitter-TV-plus-promos data set explained 65% of the variance in the premiere audience sizes, compared with 48% using promotions data alone. According to the research firm, the Twitter-enhanced model would help ad agencies and TV networks make better choices about where to place their bets for upcoming shows.

Nielsen noted that the study doesn’t prove TV-related chatter on Twitter necessarily leads to bigger ratings. The firm conducted a study in 2013 that analyzed live TV ratings and tweets for 221 broadcast primetime show episodes; that showed the volume of tweets caused statistically significant increases in live ratings among 29% of the eps, although Nielsen said how the dynamic worked wasn’t clear.

The latest study used Nielsen’s 24/7 Twitter TV Tracking data set. That’s currently available to clients for more than 650 English- and Spanish-language shows, according to Nielsen.