The same procedure was used across cultures which means that it can be compared which gives the findings more reliability. Also, there seems to be consistency with the same three attachment types identified.
In general, the majority of children were securely attached and when other attachment types were dominant it was due to cultural differences, rather than bad parenting.
Ainsworth devised a well-controlled procedure within the lab that could be repeated and also used many observers which gives inter-rater reliability.
The laboratory is an unnatural place for the child to be and they are not used to it which is a problem with the validity of her work.
A fourth attachment type was added later and this is a weakness of her work because it suggests that it was an incomplete idea.
Attachment depends on complex issues like the responsiveness of the child (and their temperament) and these are hard to measure. Because of this they are not scientific conclusions as they are not measurable.