- parents may not behave naturally if the researcher was present
- mothers are subjective, may miss important details
- natural experiment so variables cannot be controlled
- limited sample of 60 babies from one district and social class = generalisation is difficult so also lacks external validity
- lacks temporal validity as conducted in 1960 when laws were different e.g now fathers can go on paternal leave
- socially sensitive - families may be judged if their childs attachment doesn't follow the stages