Page 44 - Contributed Paper Session (CPS) - Volume 1
P. 44

CPS798 Quindale E. C. et al.
                                                    1
                                                                     ̂
                                                          ̂
                     2.  Compute the mean  ̂ ()  = ∑    , where  ()  is the estimate of an
                                                           
                                                     =1
                         indicator  in  resample  ,   is  the  total  number  of  households  in  a
                         province and   is the estimate of an indicator in household .
                                       ̂
                                       
                     3.  Repeat (A) and (B) 500 times.
                     4.  Compute the bootstrap estimate of an indicator and the corresponding
                         bootstrap estimate of variance. The bootstrap estimate of an indicator,
                         ̂
                                                  ̂
                                =1
                                          =1
                            =  ∑     ̂ ()  =  ∑ 500   ̂ ()  ,    ≡  bootstrap  estimate  of  an  indicator,
                                           500
                         ̂
                          ()  ≡estimate of an indicator in resample , and  ≡ total number of
                         resamples (  =  500).  Note  that   is  just  the  average  of  the  500
                                                           ̂
                                                            
                         means  computed  in  steps  (2)  and  (3).  The  bootstrap  estimate  of
                         variance was computed as:
                                                                            ̂
                                                        ̂
                                                                     ̂
                                           ∑   ( ̂ ()  −  ) 2  ∑ 500 ( ()  −  ) 2
                                      2
                                        =  =1      =  =1       
                                                                  500 − 1
                                                                             ̂
                                  2
                         where:   2  ≡  bootstrap  estimate  of  variance,     ≡  bootstrap
                                  
                                                  ̂
                         estimate of an indicator,  ()  ≡estimate of an indicator in resample ,
                         and    ≡  total  number  of  resamples  (  =  500) .  The  annualized
                         provincial totals was computed as the weighted average of the four
                                                                                         ̂
                                           ̂
                                                        ̂
                         quarters,  that  is,  = ∑ 4    ,  = 1,2,3,4 () where:  ≡
                                           
                                                                                          
                                                 =1
                                                       
                           estimate  of    for  province  ,  ≡  estimate  of    for  quarter    in
                                                         ̂
                                                         
                                                  1
                                                  ̂ 
                         province ,  =   1  1   1   1  ≡ weight  of  quarter  in  province
                                      
                                           (  +   +   +   )
                                             ̂ 1  ̂ 2  ̂ 3  ̂ 4
                         , ̂   ≡ bootstrap standard error of  for quarter  in province . Its
                         variance,         ( ) =  ̂ ( ) +  ̂ ( ) +  ̂ ( ) +
                                                           ̂
                                                      2
                                                                                       ̂
                                                         2
                                                                                     2
                                                                       2
                                                                         ̂
                                              ̂
                                                                    2
                                                                                  2
                                                                          2
                                                                                 3
                                                            1
                                                                   2
                                                                                        3
                                                      1
                                               
                          ̂ ( ).
                                 ̂
                              2
                           2
                           4
                                 4
                  2.3 Measure of Precision
                     The precision of the annualized provincial estimates for totals and rates
                  were  evaluated  using  the  coecient  of  variation.  Estimates  with  coecient  of
                  variations which are less than or equal to 10% implies precision, and hence,
                  reliable estimates.

                  3.  Results
                  3.1 Simple Averaging Method
                     Table 1 shows the number of provinces/HUCs by LFS Indicators (Levels)
                  and CVs using simple averaging method. For employment, 91.4 percent of the
                  provinces in the country have CVs less than 10% and the remaining 8.6 percent
                  of the provinces have CVs between 10% and 20%. On the other hand, 26 cities
                  have CVs less than 10%, six (6) have CVs between 10% and 20% and one (1)
                  city with CV more than 20%. For unemployment, 7 out of 10 (or 70%) of the
                  provinces have CVs less than 10%, 23.5 percent of the provinces with CVs
                  between 10% and 20%, and the remaining five (5) provinces with CVs larger
                                                                      33 | I S I   W S C   2 0 1 9
   39   40   41   42   43   44   45   46   47   48   49