Parents’ choice of daycare institution is a big thing for parents in Denmark. Here, parents can wish which institution they would like their child to attend, and the municipality and/or the leader of the institution will decide if they will get a spot at the desired institution. Often they put a lot of effort into it, although the information about the different institutions they have might be sketchy. Some municipalities offer some performance information, some do not. Parents can visit potential institutions and get an impression, but it takes time.
Generally, this is an example of free choice and quasi-market competition in public service provision that has been a popular solution in many democratic welfare countries over the last 30 years. A lot of research has looked at these quasi-markets to determine how they work and if they are efficient. However, few look at the citizens who have to make the choice – what information do they depend on in making that choice?
In this (Danish) article published in Politica, we use a conjoint experiment to look at what information makes citizens more likely to choose a daycare institution. The experiment presented citizens1 with 5 scenarios that contained two profiles of fictive daycare instituions in each. Each daycare profile included information on a set of attributes, and the levels of those attributes were randomized for each profile. The attributes were: Distance to institution, share of satisfied parents, share of children with other background than Danish, share of staff who are educated pedagogues, number of children per staff, public or private daycare, and total number of children.
And then we asked them to choose between the two profiles of daycare institutions 5 times (because there were 5 scenarios). You can see the results in the figure.

As you can see, all attributes except public/private daycare has significant effects on the choice of daycare institutions – and in a meaningful way too! Each level of an attribute is compared to the reference category. For example, in the distance attribute, being presented with a daycare institution that is situated 10 km away from home is compared to the reference category of being presented with one only 0.5 km from home. Not surprisingly, the probability of choosing such an institution 10 km away, is much lower (hence the placement of the dot to the left of the zero line).
It is important to note that it can be difficult to compare between attributes here, because the levels are different. We also test whether the share of educated pedagogues dampens some of the negative influences such as the number of children per pedagogue. This does not seem to be the case.
The article is open access at Politica:
- We have a sample representative of the general Danish population, so not all were parents. We do ask them if they are, and it surprisingly makes no difference to the results if we only focus on the parents. ↩︎
Comment