Page 117 - Special Topic Session (STS) - Volume 2

STS463 Noraliza M.A. et al.
Bank and Organisation for Economic Co-operation and Development (OECD),
among others, mostly depend on the LFS as the major data sources.
In spite of its many plus points, the LFS is not without faults and limitations.
As far as reliability and quality of the estimates goes, LFS being a household
survey is subjected to sampling errors and non-sampling errors. The sampling
errors occurrence is especially true when the estimates are disaggregated for
small groups or areas which are under-represented in the sample (DOSM,
2019; ILO, 2017a). Although the sampling error can be reduced by increasing
the number of observations sampled, it is not the most financially smart
solution in the long run. The non-sampling error in LFS might prevail due to
misleading comprehension of definitions and concepts either by enumerators
or respondents; or defective methods of data collection. Unlike the sampling
error, this error may rise with the increase in sample size. Banda (2003)
emphasised that this type of error can be more detrimental for large-scale
household surveys in the absence of proper control mechanism.
Another issue to consider is the use of proxy respondents i.e. one
household member providing the required information on all the members of
his or her household. Since 75 per cent of the sampled households in the
national LFS currently uses PAPI where enumerators visited households to
obtained information, more often than not, households are not fully occupied
due to members being at work, school or other places. According to the ILO
(2017a), this may also hamper the precision of the response.
Due to its reputation as the most cost-effective and frequent data
collection activity, the LFS is often ridden on for testing new data collection
instrument in addition to the regular supplements of Migration Survey and
Salaries & Wages Survey. At times, these added loads might compromise the
quality of responses for labour-related fields in the questionnaire.
Furthermore, this also may add to respondents’ burden and eventually cause
the response rate to decline.
Considering the sample design which does not take into account economic
activity and occupation of household members, certain information is
obtained indirectly and is perceived as the by-product of the LFS. The obvious
instances would be estimates of employment across economic sectors or
occupation categories. Although both might produce statistically reliable
estimates at major groups, disaggregation at detail subsectors or occupations
may not be able to offer nationally representative estimates. This is also
customary for other variables in the LFS such as educational attainment,
highest certificate obtained and field of studies. The sampling base on which
such estimates would depend would be too small, and the degree of variability
correspondingly high (European Communities, 2003).
Being a regular survey with multiple demographic and socioeconomic
variables, LFS is sometimes the subject of misused. The short term difference

106 | I S I W S C 2 0 1 9