Is a “star” worth a thousand words?: The interplay between product‐review texts and rating valences