2
$\begingroup$

I'm French and I'm reading an English book. I don't understand the term "when missingness is by design" — what does "by design" mean?

$\endgroup$
0

2 Answers 2

2
$\begingroup$

The main idea of planned missing data designs is that you deliberately decide not to gather some data because you assume that you will be able to recover it from the other data you gathered.

The typical scenario would be that you have some test, or questionnaire but it is too long to be administered to all subjects. To shorten the questionnaire you delete some items from it in such fashion that each group of participants shares some amount of items with other groups. You can find example of such design on these slides:

enter image description here

Group 1 responded only items 1 and 2 from the Openness scale, Group 2 items 1 and 3, and Group 3 items 2 and 3. Knowing the pairwise relations between all the items we can easily "recover" the missing information. Moreover, we have data on different combinations of answers to different questions while we were able to gather the information with shortened questionnaire forms.

The above example is simplified comparing to much more complicated ones that may appear (e.g. in equating studies). Example of such study conducted by Polish Educational Research Institute can be found below (sorry but it is available only in Polish):

enter image description here

As you can see, multiple parts of test forms were shared by multiple groups. The final equating procedure made it possible to recover the missing data and learn what could be the results of students if they answered the different form. All of this enabled the researchers to learn about relative difficulty of different test forms.

You can find more information on this topic in papers by Pokropek (2011), Little and Rhemtulla (2013) or book about missing data by Graham (2012).


Pokropek, A. (2011). Missing by design: Planned missing-data designs in social science. ASK. Research & Methods, 20, 81-105.

Little, T. D., & Rhemtulla, M. (2013). Planned missing data designs for developmental researchers. Child Development Perspectives, 7(4), 199-204.

Graham, J. W. (2012). Missing data: Analysis and design. Springer Science & Business Media.

$\endgroup$
2
$\begingroup$

It means that the scientists decided deliberately not to collect that information. So for example if you do a two-phase study and in phase one give people a simple screening test and then follow up all those with high scores but only half of the those with low scores then in phase two the scores on those people (the other half of the low scorers) will be missing by design.

$\endgroup$
1
  • $\begingroup$ Thank you very much for your quick answer. I can now understand the meaning! $\endgroup$
    – Manu
    Commented Mar 15, 2016 at 10:41

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.