The sample was made up of twenty-four families (12 with a boy and 12 with a girl). Each family had only one child between 20 and 24 months.  Both parents lived at home and both parents were between 20 and 30 years old. All the families were white. Five families lived in university housing and six in apartments. The rest lived in private homes. The sample was varied in income; some parents were still students.

Observers used an observation checklist of 46 child behaviours and 19 reactions by parents. There were five 60-minute observations completed for each family over a five-week period.  The observer used time sampling, making note of the child's behaviour every 60 seconds and then noting the parents’ response. Two observers were used to establish inter-coder reliability. The agreement between the two observers on the child’s behavior was 0.93 and for parents’ reaction 0.83.

After the observations were finished, each parent was asked to rate the 46 behaviours as more appropriate for girls, for boys or neutral.  Each parent also filled out a questionnaire on the socialization of sex roles.

The findings of the observations include:

• Boys were more likely to be left alone by their parents than girls.
• Parents gave boys more positive responses when they played with blocks than they did girls.
• Parents gave girls more negative responses when they manipulated an object than they did boys.
• Parents gave more positive responses to girls than boys for playing with dolls and more negative responses to boys.
• Parents criticized girls more when they participated in large motor activities – e.g. running and jumping.
• Parents gave more positive responses to girls than boys when they asked for help and a more negative response to boys.
• Fathers were more concerned with appropriate sex-typing than mothers and both parents found more behaviours appropriate for girls only than for boys only.

Parents reacted significantly more favorably to the child when the child was engaged in same-sex preferred behavior; children were more likely to receive negative responses to cross-sex-preferred behaviours. Parents gave girls more positive responses when they engaged in adult-oriented, dependent behavior.

On the questionnaire, parents did not see asking for help as a sex-preferred behavior; however, they were more likely to act positively to a girl than a boy asking for help. This suggests that the parents were not fully aware of the methods they use to socialize their children.

The study was naturalistic, done in the natural environment of the family, rather than in a lab.  The study has high ecological validity.

The use of two observers means that we can verify that the observations are not influenced by personal biases.  There is high inter-rater reliability, meaning that they both recorded the same data during the observations.

The study suffers from sampling bias.  The families were all linked to the university, they were all white and they were all American. In addition, 24 families is a very small sample. The ability to generalize from this sample is limited.

The researchers found a short list of behaviours that result in the parents criticizing the child.  However, the families knew that they were being watched.  This means that they have demonstrated demand characteristics.  In reality, there may be more or fewer behaviours that the parents would normally criticize when they are not being observed.