Poster Paper:
Recall Accuracy in Agricultural Data: Evidence from the LSMS
Saturday, November 14, 2015
Riverfront South/Central (Hyatt Regency Miami)
*Names in bold indicate Presenter
Policymakers-and-policy-researchers-make-use-of-household-survey-data-for-a-wide-array-ofpurposes,-
most-importantly-for-identifying,-diagnosing,-and-solving-problems.-While-we-mightprefer-
to-have-access-to-real<time-data,-technological-and-budget-constraints-generally-preventthe-
collection-of-this-type-of-data.-Given-these-constraints,-most-household-surveys-rely-onrecall-
data,-in-which-respondents-are-asked-about-previous-behavior.-The-length-of-this-recallrange-
variess-greatly,-from-just-a-day-to-more-than-a-year.-As-such,-there-is-a-worry-that-datacollected-
in-this-manner-may-suffer-from-recall-bias.-These-biases-include-such-things-as-simplememory-
loss-as-well-as-behavioral-tendencies-like-telescoping.-If-these-data-do-indeed-sufferfrom-
recall-bias-of-any-kind,-then-the-resulting-conclusions-we-draw-from-the-data-may-beincorrect.-
In-order-to-test-for-the-presence-of-recall-bias,-we-use-Living-Standards-Measurement-
Study-(LSMS)-data-from-three-sub<Saharan-countries:-Malawi,-Tanzania,-and-Uganda.-We-takeadvantage-
of-the-panel-structure-of-the-data-to-examine-the-effects-of-length-of-recall-onseveral-
variables-related-to-harvest,-including-cultivation-labor,-harvest-labor,-and-non<laborinputs.-
We-approach-this-question-as-a-follow<up-to-Beegle-et-al-(2012)1,-who-examine-thesame-
question-using-similar-surveys-but-without-the-use-of-panel-data.-If-the-timing-of-thesurvey-
is-randomly-allocated-across-households,-as-Beegle-et-al-argue-it-is,-then-the-resultingregression-
estimates-are-unbiased-However,-the-addition-of-household-fixed-effects-allows-us-tofocus-
on--across-plots-and-crops-within-a-household.-We-are-able-to-identify-the-month-in-whichharvest-
begins-and-ends-on-each-plot-within-the-household-and,-as-such,-we-are-able-toconstruct-
estimates-of-the-effects-of-recall-length-using-plot<level-data-and-household-fixedeffects.-
Moreover,-the-panel-nature-of-the-survey-also-allows-us-to-examine-these-effects-acrosstime.--
-
1-Beegle,-K.,-Carletto,-C.,-&-Himelein,-K.-(2012).-Reliability-of-recall-in-agricultural-data.-Journal(of(
Development(Economics,-98(1),-34<41.-
most-importantly-for-identifying,-diagnosing,-and-solving-problems.-While-we-mightprefer-
to-have-access-to-real<time-data,-technological-and-budget-constraints-generally-preventthe-
collection-of-this-type-of-data.-Given-these-constraints,-most-household-surveys-rely-onrecall-
data,-in-which-respondents-are-asked-about-previous-behavior.-The-length-of-this-recallrange-
variess-greatly,-from-just-a-day-to-more-than-a-year.-As-such,-there-is-a-worry-that-datacollected-
in-this-manner-may-suffer-from-recall-bias.-These-biases-include-such-things-as-simplememory-
loss-as-well-as-behavioral-tendencies-like-telescoping.-If-these-data-do-indeed-sufferfrom-
recall-bias-of-any-kind,-then-the-resulting-conclusions-we-draw-from-the-data-may-beincorrect.-
In-order-to-test-for-the-presence-of-recall-bias,-we-use-Living-Standards-Measurement-
Study-(LSMS)-data-from-three-sub<Saharan-countries:-Malawi,-Tanzania,-and-Uganda.-We-takeadvantage-
of-the-panel-structure-of-the-data-to-examine-the-effects-of-length-of-recall-onseveral-
variables-related-to-harvest,-including-cultivation-labor,-harvest-labor,-and-non<laborinputs.-
We-approach-this-question-as-a-follow<up-to-Beegle-et-al-(2012)1,-who-examine-thesame-
question-using-similar-surveys-but-without-the-use-of-panel-data.-If-the-timing-of-thesurvey-
is-randomly-allocated-across-households,-as-Beegle-et-al-argue-it-is,-then-the-resultingregression-
estimates-are-unbiased-However,-the-addition-of-household-fixed-effects-allows-us-tofocus-
on--across-plots-and-crops-within-a-household.-We-are-able-to-identify-the-month-in-whichharvest-
begins-and-ends-on-each-plot-within-the-household-and,-as-such,-we-are-able-toconstruct-
estimates-of-the-effects-of-recall-length-using-plot<level-data-and-household-fixedeffects.-
Moreover,-the-panel-nature-of-the-survey-also-allows-us-to-examine-these-effects-acrosstime.--
-
1-Beegle,-K.,-Carletto,-C.,-&-Himelein,-K.-(2012).-Reliability-of-recall-in-agricultural-data.-Journal(of(
Development(Economics,-98(1),-34<41.-
