What's the issue?
Many studies have shown that there is not always coherence in answers from parents and children when asked about the same issues. For example parents might say that they monitor closely what their children do online, while the children might say that they are not closely monitored by their parents.
To collect data from both parents (or adults, such as teachers) and children is not uncommon in studies where the focus is on children's behaviour. It is possible to combine information from parents and children in various ways but these might be seen as the main alternatives:
Parent as only informant (proxy)
Parent as main informant and child as supplementary informant
Parent as main informant and child as main informant
Child as main informant and parent as supplementary informant
Child as only informant
All these approaches have their advantages as well as shortcomings. There is a twofold advantage of comparing data from parents and children. Firstly it enables cross validation of information on children's behaviour as adults are often more precise when it comes to measuring time use (especially for younger children). Secondly the difference in answers from parents and children is an interesting concept of study in itself.
Pitfalls to avoid
There are some issues linked to this kind of data. The comparison between the groups can of course be made on the aggregate level (looking at children as a group and the parents as a group). It is however not safe to assume directly that differences or similarities on the aggregate level hold true on the individual level. If for example a study reveals that a certain proportion of children do certain things on the internet and at the same time a considerably lower proportion of parents think that their children do these particular things this does not allow us to assume that parents do not know what their children do on the internet.
An example of the difficulty in generalizing from the aggregate level to the individual level is that if a proportion of teenagers in 34 countries who have had sexual intercourse is compared to the proportion of teenagers who have been drunk at least twice a very weak relationship is found between the use of alcohol and the likelihood of having had sexual intercourse. In line with that Icelandic children hold the sixth place for likelihood of having had sexual intercourse and 22nd place for the proportion of 15 year olds who have been drunk at least twice. When the same question is analysed on the individual level, however, for children in Iceland 14% of teenagers who have never been drunk have had sexual intercourse compared to 83% of those who have been drunk 20 times or more.
Questions to consider
Ideally, data from parents and children should be linked on the individual level. This however complicates the research design and (depending on countries) calls for informed consent to be obtained from both the parents and the children which in turn is likely to lower the response rate considerably.
As a general rule the younger the children the more common it is to rely on parents or other adults as informants. This of course calls for some considerations on the validity and reliability of the information obtained. As a rule of thumb it is easier to obtain accurate information on behaviour (if they use the internet for example and for how long) but attitudes are more difficult to assess.
When children are asked to give information on their parents it is sometimes possible to cross validate their information with comparison to other studies. An example of this is parent's occupation or educational level or parent's use of the internet.
The SAFT (Safety Awareness Facts and Tools) Project. SAFT 2006 Parent and Children survey. 2004-2006. Norwegian Action Plan for Children, Youth and the Internet and the European Commission Safer Internet Action Plan: Norwegian Media Authority.
UK Children Go Online Project. UK Children Go Online: Emerging Opportunities and Dangers. London: London School of Economics and Political Science.
Example of good practice
Examples of comparison on the aggregate level are some of the studies conducted as a part of the SAFT project. In these studies, parents as a group and children as a group have been surveyed separately.
An example of a study where comparison is made on the individual level is UK Children go Online which, for example (Livingstone & Helsper, in press), examined parental regulation of children and teenagers' online activities with answers matched on the individual level.