Integrated Farm Household Survey 2003

Sampling Frame

The IFHS utilized different sampling frames at the barangay and household levels. At the barangay level, the list of agricultural barangays covered in the 1999 Barangay Screening Survey (BSS) served as the sampling frame while at the household level, the listing of households generated from the 2000 Census of Population and Housing (CPH) of the National Statistics Office (NSO) was used as basis for drawing the samples.

The 2000 CPH listing was utilized as sampling frame for the IFHS despite the limitation that households were not classified into farming and non-farming categories for two major considerations. First, the 2000 CPH provided the most updated lists of households by barangay. Second, budgetary constraints precluded the conduct of household screening in the selected sample barangays for the survey.

Sampling Design

The domain of the survey was the province. A two-stage stratified sampling design was adopted with the barangay as primary sampling unit and the farming household as secondary sampling unit. The number of farming households was used as the stratification variable. Primary and secondary sampling units were both drawn using simple random sampling.

Sample Size Determination

In getting the number of barangays as representative of the domain (province) level, the total number of agricultural barangays in the province reported in the 1999 Barangay Screening Survey (BSS) was used in proportionately allocating the target sample size of around 600 barangays to the Integrated Farm Household Survey (IFHS) provinces. Due to budgetary consideration, the total number of barangays included for small and large agricultural sampling of households with at least one member engaged in agricultural activity. provinces was set at six (6) and nine (9) barangays, respectively, depending on the computed total sample size for the province, that is,

               n'  =  6 if n < 6, and
               n'  =  9 otherwise.

Ten (10) sample households were allocated for each sample barangay. This procedure resulted in total sample size of 592 barangays and 5,920 households for the entire country.

Sample Selection

Primary Sampling Units

A general feature of the design was the division of the primary sampling units into strata of approximately equal sizes relative to the number of farming households reported in the 1999 BSS. The division of the barangays within the province and the drawing of sample was done as follows:

The barangays were arrayed in descending order based on the total number of farming households. These barangays were then divided into three (3) strata such that the cumulative total number of farming households of all the barangays in any one stratum was approximately of the same magnitude as the rest of the individual strata. Thus,

Stratum 1 barangays constitute all "large barangays",
Stratum 2 barangays constitute all "medium barangays", and
Stratum 3 barangays constitute all "small barangays"

with respect to total number of farming households.

Equal sample sizes were allocated and drawn from the three strata, resulting in two (2) and three (3) sample barangays, respectively, per stratum depending on the sample size for the province. Selection of sample barangays wss done at the BAS Central Office using simple random sampling. The generated lists of sample barangays were then submitted to NSO for the drawing of sample households and for the
photocopying of corresponding barangay maps.

Secondary Sampling Units

Drawing of sample households was made at the NSO field offices using simple random sampling of households with at least one member engaged in agricultural activity. The generated lists of samples were sent back to BAS Central Office for control and distribution to concerned Provincial Operations Centers (POCs).

Replacement of Samples

As in any survey, there were cases wherein samples need to be substituted or replaced. Following were the guidelines in replacing sample barangays and/or households.

Sample Barangays - Only two general reasons were considered valid for substituting barangays:

1. Transportation costs were way above the allocated budget for operations; or

2. Unfavorable peace and order situation in the area.

The list of replacement barangays served as the only source of substitute barangays. It was emphasized that a replacement barangay should be taken only from the list of replacement barangays in the same stratum.

Sample Households - Only the reasons enumerated below are considered valid for replacing households.

1.Household was not a qualified IFHS sample:

a. For regions except NCR: Candidate household was not a farming household;
b. For NCR: Candidate household was not into agricultural activities, or into agricultural activities but produce was not intended to generate income for the household;
c. Conditions (a) and (b) were satisfied but there was no agricultural operation during the reference period (July 2002 to June 2003);

2. Household was a qualified IFHS sample but any of the following situations arose during visit:

   a. No qualified respondent was available for interview during the entire survey period;
   b. Qualified respondent refused to be interviewed;
   c. Interview was terminated;

It was emphasized that reasons for substituting sample households should be validated first by the field supervisor before replacement is allowed. Replacement households should be taken only from the list of replacements for the barangay.