THE ANALYSIS OF MEMORY BIAS IN AGRICULTURAL LABOUR DATA COLLECTION: A CASE STUDY OF SMALL FARMERS IN NIGERIA
分析了尼日利亚129个农户每周一次的劳动数据，发现回忆偏差导致报告的农业劳动时间可能高估实际值近38%，对使用回忆法收集数据的学者有警示意义。
An analysis is undertaken of daily labour data collected once‐per‐week for a sample of 129 farm households in Benue State, Nigeria during 1979/80. The purpose of the analysis is to test for the presence of memory bias in these data for which the recall period is up to seven days. The analysis indicates significant between‐day differences in hours of agricultural work. This is partly due to the low work input for Sunday but it is also noted that the mean number of hours of agricultural work reported for the day prior to data collection is significantly lower than the means for days further removed from the time of data collection. This finding is taken to be confirmation that a significant level of memory bias exists in the labour data and it is estimated that reported hours of agricultural labour may overstate actual hours of such work by almost 38 per cent.