Home  |  News  |  Sitemap
 

The Methodology

Sampling and screening methodology
Boosted sample for EMS Select
Fieldwork
The data and the databases
Weighting of the EMS Regular screening database
Establishment of universe size and basic sample profile
Projection to the estimated universe sizes
Ascription of the postal questionnaire


Sampling and screening methodology

The EMS objective is to identify a representative sample of main income earners in the top income (20%) of households in each country. In order to do this, we purchase random telephone sample from Survey Sampling International, a US based company specialised in sampling solutions since 1977. SSI random digit dialling (RDD) methodology provides an extremely representative random digit sample. Each exchange and working block will have a probability of selection equal to its share of active working blocks. Selected telephone numbers may be protected against reuse for a period of six months. Business numbers, when detected, are eliminated.

Following sample generation, SSI employs a process that identifies most randomly generated, unassigned telephone numbers. Vacant, unavailable, or unassigned telephone numbers, fax numbers, and modem numbers are detected and removed.

The income level to qualify for the top 20% varies by country and is determined empirically. That is to say we can observe at what income level 20% of all households qualify. In the course of the fieldwork the qualification percentage is closely monitored and income thresholds are adjusted if necessary.

Respondents are asked to include all sources of income from all income earners in the household before tax (except for Germany and Austria where net incomes are used). Annual income is calculated on the basis of a 13-months salary package (except in the UK and Denmark where it is 12 months).

Individuals are eligible if they contribute at least 40% of the total household income. Where two persons within a household qualify the EMS respondent is randomly selected. A household is defined as a group living together as family or partners. Where a number of individuals share accommodation but not as a family, each individual is treated as a separate single person 'household' and is potentially eligible for the survey.

A separate boost sample is interviewed to support more detailed analysis of the top-end segments.

Boosted sample for EMS Select

The objective of EMS Select is to identify a representative sample of the top 3% in Europe. This EMS Select sample is composed of:
  • Top managers: senior managers working at an establishment with 10+ employees, and responsible for 1 or more employees with an annual personal income of €40.000 or more (€ 30,000 in France, Italy, Portugal and Spain; € 50,000 in Norway and Switzerland; no personal income threshold for C-EMS countries).
  • Frequent flyers: people who make 6 or more international business flights a year.
  • High-income earners: people with a personal income of € 80,000 or more yearly (€ 30,000 in Czech Republic, Hungary and Poland; € 60,000 in France, Italy, Portugal and Spain; € 100,000 in Norway and Switzerland).
Two additional sample sources are used to boost the EMS Select eligible respondents:
  • For top managers we use Dun & Bradstreet lists of larger companies (10 or more employees). From these lists, a random selection of senior managers is invited for the EMS Select interview. The respondents in this sample are approached at work and - after qualification for EMS Select - invited to co-operate in the EMS Select interview either as a re-contact interview when at home, or at work. Most respondents prefer to complete the interview at work.
  • Frequent flyers and high-income earners are selected from consumer lists.
The screening interview is dedicated to select the three different target groups. When a respondent is eligible, the main interview may be conducted with the eligible individual at that stage, or at a later date. Non-response to the survey is recorded as appropriate.

Fieldwork

The annual fieldwork is conducted on a continuous basis, with no break during the summer months. The table below shows the periods during which telephone interviews were conducted for each of the four fieldwork waves (i.e. including the first and second half of each year).

2005-1 31 January 2005 - 9 July 2005 2006-1 16 January 2006 - 16 July 2006
2005-2 11 July - 23 December 2005 2006-2 17 July 2006 - 23 December 2006


Interviews are conducted each weekday and Saturday.
The telephone numbers issued for interviewing are controlled by day of week and week of the year to ensure a balanced sample over time. Up to 12 recalls for non-response are made at different times and on different days up to 4 weeks. Numbers, which are not in use, business, other non-residential or fax /computer lines are excluded from the survey universe.

The data and the databases

The survey data are weighted on a national basis and grossed to the national universe estimate. Each 6 months the data set is weighted separately. The weighting is carried out in several steps, which are described below.

Weighting of the EMS Regular screening database

In each country all random screening interviews of EMS Regular are weighted back to the distribution of households by Nielsen region, by household size and by age group found in the most recent data from the national statistical agencies.

Establishment of universe size and basic sample profile

All random interviews with EMS eligible households are analysed. Households with two main income earners are given a weight of two. These weighted data are the base for:
  • Making an estimate of the EMS universe size by country. In this calculation the resulting average number of main income earners per household, from the EMS data, is combined with the proportion of households qualifying for EMS (found in the screening stage of the EMS interview).
  • Establish the sample distribution by sex, region, household size, number of main income earners, and age groups.
Select criterion and weighting of the Select sample completes in the total database For the weighted random EMS Regular completes is established which part of the respondents falls within the criteria of the Select sample. The Select completes (including the Regular completes which qualify for EMS Select) are weighted by country to the proportion of Select found in EMS Regular, and on an equal distribution of days of week of interview. At the same time the sample distribution for a number of variables, found in step 2, are kept constant.

Projection to the estimated universe sizes

The universe size estimate is arrived at by the following calculation: multiplying the proportion of households qualifying for EMS (source: the EMS screening interview) by the average number of main income earners per EMS household (source: EMS interview).

Ascription of the postal questionnaire

Not all respondents who participate in the telephone interview completed the postal questionnaire. For those who did not, the information has been ascribed.

In both the 2005 and 2006 databases, the missing variables have been ascribed to the non-respondents of the postal questionnaire by using sophisticated statistical techniques. The resulting database EMS Ascribed contains the full set of variables for each telephone respondent. Within this the media data and demographics originate directly from the telephone interview; the marketing data from the postal questionnaire, however, originate directly from the questionnaire as filled in by 34% of the respondents, while for the others the data are ascribed based on matching respondents.

The ascription process is conducted after several external conditions are set as to on what basis donor A is to be married to a recipient B: to find the best donor for each recipient.


  Print



© 2008 Synovate. All Rights Reserved.