other bodies interested in collecting adequate data on the socio-economic status of marginalised groups. This is a long-term investment that goes far beyond the validity of the results of surveys and censuses. These kinds of partnerships with local communities and NGOs are required to improve the data collection process and respective results. TOOL NO 5: SURVEY-DESIGN53, DATA COLLECTION54 & SAMPLING METHOD: CASE EXAMPLE FROM UNDP UKRAINE This tool is provided by UNDP Ukraine. The Autonomous Republic of Crimea is the region where the surveys were conducted. It is a multiethnic region of the Ukraine, comprised primarily of ethnic Russians, Ukrainians, Crimean Tatars and of numerous other minorities. Using this tool: This tool provides a useful example of the rationale, processes and outputs of data collection on ethnic groups. It complements Tool 4 by demonstrating the application of many of the principles of good practice in ethnic data collection and how to operationalize them. If current public records and statistics do not provide reliable data based on ethnic demographic divisions, statistical evidence drawing on existing data cannot be an option. Since ethnicity is a key demographic category within the Crimean region, it would be beneficial for ongoing monitoring of social inclusion and cohesion at the structural level to ensure that future public statistical data includes an ethnic profile. Priority needs to be given to the collection of comparable and concrete data on social conditions linked to economics, land/housing/ living and education issues. These are currently pressing areas of concern and at the forefront of general discourse. The data collection needs to occur once and be repeated following a period of 1-1.5 years to compare changes in the socioeconomic situation. This data will only be useful if the initial preliminary results are examined thoroughly and strategic plans and objectives are devised and rolled out to address the potential disparities and inequalities observed in the data collected. The scale of data collection would need to be across Crimea’s population, including all districts and cities. As the aim is to understand differences between various key demographic subgroups and ensure an overview of the general population of the region, the use of a stratified random sampling method55 is advised. To allow for a margin of error of 2% and a confidence rate of 95% with regard to the comparable data collected, the following steps are advised: 1. Outline the population size within each district/city (“Area unit”) stratified according to the 3 majority ethnic subgroups with percentage of population size above 0.4 % of total country population size. 2. Calculate 0.97% of total population in each ethnic subgroup of every district. This is the size sample group you would need to randomly select. 3. To ensure that the data collection adds up to 0.97%, estimate the respond rate and based on that, increase the number of each sample size. Resources drawn from include: The Gallup Organization, Gallup Poll Questionnaires, http://www.gallup.com/; and Cornel University, Empire State Poll 2006, Survey Research Institute, http://sri.cornell.edu/sri/esp.reports.cfm (accessed 9 August 2009). 53 Resources drawn from include: UN-ECOSOC, Handbook of Household Surveys, Statistics Office, New York, 1984 http://unstats.un.org/unsd/publication/SeriesF/ SeriesF_31E.pdf; UN-ECOSOC: Designing Household Survey Samples: Practical Guidelines, Statistics Division, Series F No.98, New York, 2005. 55 Resource useful for calculating sample: http://www.custominsight.com/articles/random-sample- calculator.asp (accessed 9 August 2009). 54 Chapter 9: Data Collection Tools 135