1
$\begingroup$

We would like to ask your advise regarding the missing data review. We are focussing on missing GOS values only. We would like to clarify the distinction of pairwise and listwise deletion.

We have seen that the majority of studies report predictors on all patients enrolled, however many lose patients to follow up. They then carry out analyses of outcomes using the number of patients where outcome is available only. Is this considered pairwise deletion of data?

Conversely, there are studies in whom completeness of GOS data was an inclusion criterion. Therefore they did not even report predictors on those patients with missing GOS. We consider this listwise deletion.

We would welcome your opinion.

Improve this question
$\endgroup$
1

1 Answer 1

Reset to default
-1
$\begingroup$

Listwise Deletion - Delete entire record even if there is one missing value

Pairwise Deletion - Will remove only specific variables with missing values from the analysis and continue to analyze all other variables without missing values, variables chosen will vary from analysis to analysis based on missingness. 

Height age sex
12     10  m
14     17  f
8      5   f
10     7   .
19     23  m

In the above example for observation 4 while performing a correlation we will only perform correlation between height and age and ignore correlation between age and sex

Improve this answer
$\endgroup$
2
  • 1
    $\begingroup$ Please note that posting identical or essentially identical answers is not usually acceptable; if the same answer really answers two separate questions, instead of posting the same answer twice, answer the best version of the question and flag the other as a duplicate. $\endgroup$
    – Glen_b
    Commented May 6, 2019 at 8:08
  • $\begingroup$ @Glen_b sure, thank you will do from next time $\endgroup$
    – Rithvik M
    Commented May 7, 2019 at 16:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.