You are on page 1of 3

general comments in our thesis paper entitled “imputation procedures for partial

nonresponse: the case of the family income expenditure survey

1. inconsistency in formatting (like the excess spaces, sheets, etc.)


2. inconsistency in citing reference (e.g. (cheng & sy, 1999) and (cheng, 1999))
3. excessive use of copy and paste that results into sentence redundancy
4. revision in the results and discussion section. try to analyze and interpret the
results even better not just saying the numerical results presented in each table.
5. mismatching of ideas (in particular the abstract and the conclusion & the
methodology and results and discussion)

table of contents

1. where is your discussion on the deterministic regression imputation?


2. it would have been better if the simulation method was named as “simulation of
nonresponse”.
3. where is your discussion on the comparison of the distribution of the true vs.
imputed values? the discussion should be placed after the discussion of the
comparison between the actual data vs. the imputed data for better
understanding.

abstract

“in choosing for the imputation method to be applied in a data set that contains
missing data, one should consider several factors, namely, the types and patterns of
nonresponse, the availability of the auxiliart data that provides a strong association to
the variable with missing data and the nature of the data itself. “ (first paragraph,
p.vii)

“… hot deck imputation, …. values for both variables under the second visit data
were set to nonresponse to satisfy the assumption of partial nonresponse. … (a) for
this data, the hot deck and overall mean imputation methods were not appropriate for
handling missing data; (b) stochastic regression imputation provides smaller biases
for both the estimates and the real values than its deterministic and model-free
counterpart; and (c) the imputation classes must be homogeneous to produce less-
biased estimates. “(second paragraph, p.vii)

omit the line “when predicting…”! your conclusion doesn’t state this thing!

chapter 1: the problem and its background

introduction

“… ;(d) noncoverage refers to the failure to include a sampling units in the survey of
the sampling frame. “ (last line, first paragraph, p.1)
delete the last line of the first paragraph in page 2. redundant sentence!

“… spending patterns and poverty incidence, ensuring the precision of estimates in


the survey is important. given the great impact of this survey to the country,
employing imputation techniques will help statisticians to provide a method in
handling nonresponse, which could lead to a more meaningful generalization about
our conuntry’s income distribution and spending patterns. hence, having estimates
with less bias can contribute in making our policymakers and economists provide
better solutions in improving the lives of the filipinos.” (third paragraph, p.5)

what do you mean when you said “consistent results”? consistent results of what, the
estimates or the data itself?

remove the “it” word in the last sentence of the second paragraph (p.6). word
redundancy!

“… hot deck imputation (hdi), deterministic regression imputation (dri), and


stochastic regression imputation (sri). “ (third paragraph, p.6)
“… also, the variables that will be imputed for this study would only be the total
income (totin2) and total expenditure (totex2) in the second visit of the fies 1997 data.
… and (c) …. “(p.7)

chapter 3: conceptual framework

replace the word “invalid” with “unreliable”. it makes sense actually! (first paragraph,
p.17)

replace the words “that for” with “the”. (last line, p.18)

“for an imputation method to work, the pattern of nonresponse must at least be mar.”
(first paragraph, p.21)

the definition of imputation presented in page 23 is questionable. i don’t recall


kalton(1983) stating something like statistical and mathematical techniques in his
report! i think this was taken in some website like the national statistical board in the
us. i still have to check it pa.

insert this line: the terms “donors” and “recipients” are only used in hot deck
imputation which will be discussed in the preceding sections. (first paragraph, p.25)

“problems might arise if the imputation classes are not formed with caution.” (first
line, second paragraph, p.25)

You might also like