Representative sampling | 2014 Support Materials MS

For the collection of the common longer-term result indicators, there are two waves of reporting, and distinct samples with non-overlapping participants should be drawn.

• The first wave (to be reported in the AIR due by 30 June 2019) should cover participants who left operations up to mid-2018. For this wave, sampling might not be possible until end-2018.
• The second wave (for indicators to be included in the final report in 2025) should cover participants who left operations between mid-2018 and 31st December 2023. For this wave, collection might not be possible until mid-2024.

For the two common output indicators "homeless or affected by housing exclusion” and “from rural areas" data shall also be collected based on a representative sample of participants (although collecting and reporting on the full population is certainly considered as representative sampling). These indicators are to be reported only once in the AIR to be submitted in 2017 and the full population to draw the representative sample for the data collection of such indicators must include all participants benefitting from ESF support entering operations up to end-2016. The sampling might not be possible until end 2016.

Note that the YEI longer-term result indicators are to be reported annually and thus representative samples shall be drawn each year. Samples for YEI longer-term result indicators should include participants that left operations 6 months before the reported year ends (i.e. for the AIR to be submitted in 2017, the sample of participants should include participants leaving ESF operations between mid-2015 and mid-2016).

See section 3.1.3 and the summary table on the reporting schedule in section 3.5 of the EC guidance on monitoring and evaluation.

Representative samples are randomly selected at the level of investment priority to reflect the socio-economic characteristics (variables) of the participants as captured by the common output indicators covering personal non-sensitive data: gender, employment status, age, educational attainment and household situation. Representative samples have to be established per investment priority, category of region (not for YEI), and gender.

No. The indicators have to be reported on the basis of samples that are representative of all participants benefitting from ESF support up to end-2016. Selective reporting is not acceptable.

Yes. Collection and reporting on the basis of a representative sample is a “shall” provision for common output indicators on homelessness and rural areas as well as for longer-term result indicators (Annex I of the ESF Regulation). However, collecting and reporting on the full population is certainly considered as representative sampling.

No. Beneficiaries or intermediate bodies may collect the data and might assist in drawing the sample. However, since it is important to consider factors with precise statistical meaning when a sample is drawn (size, representativeness, quality, etc.), it is strongly advised to consult an appropriate expert (internal or external), who can design and document this process accordingly. 

The indicators for which data can be collected only for a representative sample are:

• Output indicator "homeless or affected by housing exclusion"
• Output indicator "from rural areas"
• All longer-term result indicators listed in Annex I and Annex II of the ESF Regulation (Regulation (EU) No 1304/2013)

Refer to section 3.5 of the EC Guidance document on Monitoring and Evaluation (Reporting) for further details on reporting of these indicators.

As mentioned in section 3.4.2 of the EC Guidance document on Monitoring and Evaluation, ensuring representativeness also for the regional dimension of the output indicators is considered as a good practice, but it is not a requirement. Regional representativeness could be achieved by establishing the sample one NUTS level lower than the level of the programme area (i.e. category of region). For instance, in the case of an Operational Programme at NUTS 2 level the sample would have the same distribution by NUTS 3 region as observed amongst all participants (see also section 4.4.1 of Annex D). 

Yes. Samples for longer-term result indicators (and indicators on rural and homelessness for the AIR to be submitted in 2017) have to be fully representative of the total population of participants at the level of the investment priority, and regarding employment status, age group, educational attainment and household situation.  There is no exception to this. All individuals counted as participants must be included in the population from which samples are drawn.