12.07.2015 Views

NMICS 2010 Report - Central Bureau of Statistics

NMICS 2010 Report - Central Bureau of Statistics

NMICS 2010 Report - Central Bureau of Statistics

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>NMICS</strong> <strong>2010</strong>, Mid- and Far Western RegionsSelection <strong>of</strong> householdsLists <strong>of</strong> households were prepared by the listing teams in the field for each enumeration area. Thehouseholds were then sequentially numbered from 1 to n (the total number <strong>of</strong> households in eachenumeration area) at the CBS, where the selection <strong>of</strong> 25 households in each enumeration area wascarried out using systematic random sampling technique.Calculation <strong>of</strong> sample weightsThe <strong>NMICS</strong> <strong>2010</strong> sample is not self-weighting. Essentially, by allocating equal numbers <strong>of</strong> householdsto each <strong>of</strong> the subregions, different sampling fractions were used in each subregion since the size <strong>of</strong>the subregions varied. For this reason, sample weights were calculated and these were used in thesubsequent analyses <strong>of</strong> the survey data.The major component <strong>of</strong> the weight is the reciprocal <strong>of</strong> the sampling fraction employed in selectingthe number <strong>of</strong> sample households in that particular sampling stratum (h) and PSU (i):Whi1fhiThe term f hi , the sampling fraction for the i-th sample PSU in the h-th stratum, is the product <strong>of</strong>probabilities <strong>of</strong> selection at every stage in each sampling stratum:fhi p1 hi p2hi p3hiWhere p shi is the probability <strong>of</strong> selection <strong>of</strong> the sampling unit at stage s for the i-th sample PSU in theh-th sampling stratum.Since the estimated number <strong>of</strong> households in each enumeration area (PSU) in the sampling frameused for the first stage selection and the updated number <strong>of</strong> households in the enumeration areafrom the listing were different, individual sampling fractions for households in each sampleenumeration area (cluster) were calculated. The sampling fractions for households in eachenumeration area (cluster) therefore included the first stage probability <strong>of</strong> selection <strong>of</strong> theenumeration area in that particular sampling stratum and the second stage probability <strong>of</strong> selection<strong>of</strong> a household in the sample enumeration area (cluster).A second component in the calculation <strong>of</strong> sample weights takes into account the level <strong>of</strong> nonresponsefor the household and individual interviews. The adjustment for household non-response isequal to the inverse value <strong>of</strong>:RR h = Number <strong>of</strong> interviewed households in stratum h/ Number <strong>of</strong> occupied households listed in stratum hAfter the completion <strong>of</strong> fieldwork, response rates were calculated for each sampling stratum. Thesewere used to adjust the sample weights calculated for each cluster. Response rates in <strong>NMICS</strong> <strong>2010</strong>are shown in Table HH.1 in this report.Similarly, the adjustment for non-response at the individual level (women and children under five)for each stratum is equal to the inverse value <strong>of</strong>:RR h = Completed women’s (or under-5’s) questionnaires in stratum h / Eligible women (or under-5s) in stratum hThe non-response adjustment factors for women’s and under-5s’ questionnaires are applied to theadjusted household weights. Numbers <strong>of</strong> eligible women and under-5s were obtained from theroster <strong>of</strong> household members in the Household Questionnaire for households where interviewswere completed.The design weights for the households were calculated by multiplying the above factors for eachenumeration area. These weights were then standardized (or normalized), one purpose <strong>of</strong> which isto make the weighted sum <strong>of</strong> the interviewed sample units equal the total sample size at the199

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!