Page 50 - Contributed Paper Session (CPS) - Volume 1
P. 50

CPS863 Abubakar S. A. et al.
                  corn production estimates at the provincial level. Specifically, this paper aims
                  to  perform  simulation  to  determine  the  best  sample  size  (barangays  and
                  households), the best sampling design, the best estimation procedure and
                  model for estimation of production of unsurveyed quarters.

                  2.  Methodology
                  2.1 Phase 1: Determination of Initial Number of Sample Barangays
                     The  LH  algorithm  is  an  iterative  procedure  that  allocate  samples  into
                  strata5. Different combinations of number of strata (3, 5 and 10 strata) and
                  coefficient of variations (1%, 5% and 8%) were explored to compute for the
                  initial number of sample barangays.

                  2.2 Phase 2: Determination of Best Sampling Design
                  2.2.1 Probability Proportional to Size-Systematic (PPS-Systematic)
                     Corn farm area was used as the measure of size. Barangays were selected
                  using  PPSSYS  and  households  were  selected  using  either  SRSWOR  or
                                                                 ∑       
                                                                  =1
                  systematic. The designbased weight,  =  =        ×   ,  ≡weight for all
                                                       
                                                             
                                                                             
                                                                      
                  household  in  barangay ,   ≡ weight  for  all  household  in  barangay  by
                                             
                  quarter,  ≡ corn  farm  area  in  barangay  by  quarter,  ≡ total  number  of
                            
                  barangays per quarter,  ≡ number of sample barangays per quarter,  ≡
                                                                                         
                  number of corn farm households in barangay  by quarter, and  ≡ number
                                                                                  
                  of  sample  corn  farm  households  in  barangay  by  quarter.  The  total  corn
                                                               ̂
                  production  in  the  province  is  computed  as:  = ∑   ∑      ,  ≡ corn
                                                                               
                                                                     =1
                                                                                    
                                                                         =1
                  production  of  household    in  barangay    per  quarter.  The  variance  was
                                                     
                                            
                                                 2
                                                                                2
                                                                      2
                                  ̂
                  computed as: () = (1 − )  + ∑    (1 −    )    where:  ≡ first stage
                                                                     2
                                               1     =1              1
                  sample  variance  (provincial  sample  variance),   ≡ second  stage  sample
                                                                   2
                                                                   2
                  variance (barangay sample variance).

                  i.     Stratified using Iterative Method
                     This method is known as cumpf rule which is based on constructing equal
                  intervals  on  the  cumulative  of  the  square  roots  of  the  frequencies  of  the
                  stratification  variables6. Proportional  allocation was  employed and  stratum
                  boundary was determined using the iterative method.

                  Case 1: PPS selection of barangays, and SRSWOR and systematic selection of
                                                     ℎ
                                                   ∑   ℎ
                  households. Note,   =  ℎ   =  =1  ×   ℎ  ,  ℎ  ≡ weight for all household
                                      ℎ
                                                     ℎ  ℎ   ℎ
                  in  barangay   at  stratum ℎ,  ≡ corn  farm  area  in  barangay   in  stratum
                                               ℎ
                  ℎ,  ≡ total number of barangays in stratum ℎ,  =total number of sample
                                                                  ℎ
                     ℎ
                  barangays  in  stratum  ℎ,  ≡  total  number  of  corn  farm  households  in
                                            ℎ
                                                                      39 | I S I   W S C   2 0 1 9
   45   46   47   48   49   50   51   52   53   54   55