Recent literature has acknowledged that children's performance on cognitive tasks can be improved when undertaken jointly among pairs of pupils leading many researchers to claim that friendship groupings should be used more frequently in the school environment. The aim of the present research was to investigate whether children in friendship pairings perform consistently better than acquaintance pairings on educational tasks taken from the curriculum, using a matched paired design. 88, 8 to 12 year old primary school pupils undertook similar Math and English tasks across two conditions, a friendship and Acquaintance pairing. Results showed no significant difference between the scores across the two pairings. Significant differences were found between the pairings across gender, with males scoring higher in acquaintance pairings and females in friendship pairings.