An analysis of the information in our possession highlighted the fact that distance teaching in the EU is extremely heterogeneous. Some countries have a great number of institutions providing distance education (about two-thirds of the institutions surveyed are in the UK and Germany), while others offer very few courses of this kind, or none at all.

This heterogeneousness meant that the various strategies for defining samples available to us had to be evaluated carefully in order to ensure that the sample was representative.

There were three possibilities. These were:

• to extract a simple random sample from the whole reference universe;
• to extract a stratified random sample according to the typology defined by our Irish partner (Distance Education International);(2)
• to extract 16 samples, one for each country in the survey

The weakness of possibilities (b) and (c), taken singly, was the fact that (b) would not have allowed us to conduct representative analyses of separate countries, while (c) not only did not consider different types of institutions but also did not guarantee a representative sample of Europe as a whole, over-representing countries with fewer distance teaching institutions.

These two possibilities were considered complementary because (b) guaranteed a representative sample of the different types of institution and (c) ensured the same for single countries.

Taking these considerations into account, it was decided to use a "mixed" strategy: extracting different samples according to possibility (b) and (c) would allow analyses of both different countries and different kinds of institutions.

The sample was thus set up as follows:

1. A stratified sample of institutions was extracted according to the "type of institution" (Sample 1). The size of this sample was identified in 15 percent of the whole survey, or about 120 institutions. In order to guarantee the representivity of the sample, for each kind of institution a minimum of 10 bodies were extracted per entry, even when this quota exceeded 15 percent of the total (see Table 1).(3)

2. Sixteen samples were then extracted, one for each country being surveyed. The size of the sample was identified in 15 percent of the institutions in each country. To guarantee the representativity of the sample, a minimum threshold of at least 10 institutions per country, even when this quota exceeded 15 percent of the total number of institutions in each country.(4)

Table 2 gives the number of institutions included in each national sample.

The samples of the four kinds of institution and of the different countries were extracted at random (without reinsertion)(5) starting with the sample lists. Preparing these lists was made extremely easy by the structure of the database of institutions compiled previously by the research group.

It is important to note that the total number of institutions extracted (277) is less than the sum of the "country" sample and the "type of institution" sample (188+131) making up the overall sample in the survey. This is because some institutions were extracted in both samples.

In order to find the remaining addresses, the embassies of the following countries were contacted:

At the end of the contact phase, by analysing the contacts that had been established, it was possible to weigh up the situation and develop a few considerations about the effective contactability of institutions providing distance teaching.

Table 4 gives the number of institutions in the country samples actually contacted, distributed according to country and compared to the number of institutions in the originally planned country sample.

It is clear that the countries with the greatest problems are Sweden and French Belgium. In both cases, despite the request for help addressed to the embassies of both countries, it was not possible to contact any institutions. Other countries which were particularly difficult to contact were Austria, where in the final contact phase only two institutions were contacted, and the UK, where only 25 percent of the institutes on the list were contacted.

It should be stressed that, despite the efforts of the research group, only 50 percent of the list was actually contacted. This highlights a very low level of visibility of institutions providing distance teaching, even within their own countries.

Table 5. Comparison between the institutions in the "type of institute" sample actually contacted and the number of institutions on the list: disaggregation by type of institution.
 

Type of institution	Number contacted	Number in original sample	Difference
DTU	7	9	-2
ULDE	13	24	-11
GTDP	3	22	-19
PDTP	36	75	-39
Tot.	59	130	-71

Key:
DTU= Distance Teaching University
ULDE= University-Level Distance Education
GDTP= Government Distance Training Provision
PDTP= Propriety Distance Training Provision

An analysis of the contacts made in the "type of institute" sample (see Table 5) shows a similar situation. In this case, again, less than 50 percent of the institutions on the original list were actually reached. It was particularly difficult to reach the institutions in the Government Distance Training Provision category. This could be due to the nature of these programmes. They were often linked to specific interventions made by government institutions and were in force at the time of the original census made by our European partners but no longer operating at the time of the contact phase.

This statement can be partially mitigated by the fact that, as we shall try to demonstrate in the following pages, the overall contacts seem to guarantee a certain representativity both at a European level (at least as far as the "type of institute" sample is concerned, which we have considered at a European level), and for each country.

Analysing the institutions contacted as a whole, considering them as an overall sample, the results are much more comforting.

Table 6 presents the distribution of 129 institutions contacted according to the type of institute.

Table 6. Distribution of 129 institutions contacted according to the type of institute; comparison with "type of institute" sample.
 

Type of institution	number contacted	number in original sample	difference
DTU	7	9	-2
ULDE	40	24	+16
GTDP	9	22	-13
PDTP	73	75	-2
Tot.	129	130	-1

Key:
DTU= Distance Teaching University
ULDE= University-Level Distance Education
GDTP= Government Distance Training Provision
PDTP= Propriety Distance Training Provision

It is clear that the proportions present in the original sample are basically respected if one excludes an over-representation of those belonging to the ULDE group and an under-representation of the PDTP category.

This is basically acceptable if one takes as given the peculiarity of PDTP institutions and takes into account the importance of the role of distance training at a university level in outlining paths and strategies for distance teaching as a whole and in evaluation issues, in particular.

Table 7 presents the distribution of the population and of the institutions contacted, disaggregated by country both in absolute and in percentage terms.

Table 8 presents the size of the single country samples actually obtained at the end of the contact phase. In this case, too, all the institutions contacted were considered aside from their belonging to single country samples or overall samples by "type of institute".

Table 8. Distribution of the population and institutes contacted by nationality and size of country samples obtained.
 

Country	population	institutions contacted	percentage of institutions contacted
Austria	13	2	15,3
Flemish Belgium	7	2	28,5
French Belgium	3	0	0
Germany	263	45	17,1
Denmark	13	4	30,7
Spain	90	17	18,8
France	61	8	13,1
Finlanda	1	1	100,0
Greece	15	8	53,3
Italy	9	9	100,0
Ireland	24	8	33,3
Holland	25	5	20,0
Portugal	5	2	40,0
Sweden	8	0	0
UK	281	18	6,4
TOTAL	819	129	15,8

The table shows how in most cases the quota of 15 percent of the original list was reached, with the glaring exceptions of the UK, France and Sweden, as well as French Belgium. Unfortunately, the threshold quota of 10 institutions per country was not reached in most cases due to the difficulties already underlined in getting the institutions' addresses.

Despite this, however, a conclusion can be reached, in our view that, apart from Sweden and French Belgium, the only case in which the sample is scarsely representative of the country's real situation is that of the UK, in which only 6.4 percent of the institutions in operation were contacted. 

Notes to the text:

1. For more details on setting up the database, see the chapter in this report by L.Acciaroli, "An archive of distance teaching institutions in EU countries."
2. For further information about these four typlogies, see the chapter in this report by L.Acciaroli, "An archive of distance teaching institutions in EU countries."
3. See Appendix 1
4. See Appendix 2
5. In a sample extracted at random without reinsertion, increasing the number of subjects extracted increases the possibility of extracting the remaining subjects on the sample list. Bailey.
6. See the chapter in this report by A.Salerni "Tools for measuring and their administration."
7. See the chapter in this report by A.Salerni "Tools for measuring and their administration."
8. During the month of August, embassies made it possible to find 40 more addresses of institutions in the sample. Considering the time-span of the research project, however, it was decided not to continue the contact phase and to proceed to collecting and analysing data.