Table 2

 Characteristics of the six “sound” trials

AuthorsCountryStudy designSample sizeSample characteristics
GenderAge rangeSocioeconomic positionEthnicity
Allen et al, 1997w1USARandomised, controlled trialI=342; C=353Male=15%; Female=85%14-15 yearsAuthors describe sample as “high risk,” with 54% living in one parent households and 46% living in two parent householdsAuthors report “black”=67% ; “white”=19%; “Hispanic”=11%; “other”=3%
Berrueta-Clement et al, 1984w2USARandomised, controlled trialI=58; C=65Male=60%; Female=40%3-4 yearsAuthors describe sample as children from families of low socioeconomic statusAuthors report sample to be 100% “black”
Campbell et al, 2002w3USARandomised, controlled trialI=53; C=51Male=49%; Female=51%0-4 yearsAuthors state sample was “high risk” from “impoverished households”Authors describe sample as all “African-American”
Hahn et al, 1994w6USARandomised, controlled trialI=125; C=125Male=47%; Female=53%14-17 yearsAuthors describe sample as recruited from list of families receiving public assistanceAuthors report “white”=15%; “black”=76%; “Hispanic”=7%; “Asian”=1%; “other”=2%
Hawkins et al, 1999w7USAControlled trial*I (full)=156; I (late)=267; C=220Male=50%; Female=50%
6-12 yearsAuthors note that more than 56% participated in National School Lunch/Breakfast programAuthors report “white”=44%; “African-American”=26%; “Asian-American”=24%; “native American”=5%; “other”=3%.
Philliber et al, 2001w10USARandomised, controlled trialI=589; C=574Male=45%; Female=55%13-15 yearsAuthors report almost half of young people come from families with at least one unemployed adultAuthors report “African-American”=43%; “Caribbean-black”=3%; “Hispanic”=29%; “white”=5%; “Asian”=4%; “multiethnic”=14%; “other”=2%

*This study began as a randomised, controlled trial: students within eight schools were randomised into intervention or control groups, and those in the intervention group received the programme for four years. At the end of four years, children in both the intervention and the control group received the programme for two years. Additional schools matched to the original eight schools were recruited into the study. Children in these schools formed a new control group.