In my survey, I have 10 sections in the questionnaire and each of the
section will arise randomly for different respondents.
But I expect, in total sample size, each of the section will be responded
for fixed number of times.

For example, for 100 sample size, I expect each section will be
responded 10 times, but which subsection will arise to respondent will be
determined randomly.

Now, is there any option to control random number generation for fixed
number of times using ODK tool in Excel (xlsx to xml) (sequence of
appearing does not matter) ?

Please only ask a question on one list. Duplicate postings do not get
quicker replies.

···
On Sun, Nov 13, 2016 at 7:41 AM, Shahadat Hossain wrote:

Hi,

In my survey, I have 10 sections in the questionnaire and each of the
section will arise randomly for different respondents.
But I expect, in total sample size, each of the section will be responded
for fixed number of times.

For example, for 100 sample size, I expect each section will be
responded 10 times, but which subsection will arise to respondent will be
determined randomly.

Now, is there any option to control random number generation for fixed
number of times using ODK tool in Excel (xlsx to xml) (sequence of
appearing does not matter) ?

Actually, For my survey purpose, I want to generate random number between 1
to 10, and each number will arise exactly 10 times (i.e: 1 will arise 10
time, 2 will are 10 times and so on.

Yeah definitely not random. I think the closest you can get to random is
pre-assignment driven by random selection.

Here is how I would do it.

Set up a csv file with flags, one column for each section in your survey
(plus key column). One row for each sampled respondent.

Use excel or whatever to generate x (say 10) random #s between 1 and sample
size (inclusive). Give those rows chosen in the 10 random choices a flag
of 1 for that section, the rest get a flag of 0. Repeat for each section.
Note that at the end some will probably have 0 sections assigned to them.
If this is unacceptable then you really don't want random at all and you
should just go through and assign sections to sampled respondents using
whatever pattern you prefer.

Once that is set up, call those flags for a particular respondent into ODK
using pulldata() and use those flags to display the assigned sections.

Jason

···
On Sunday, November 13, 2016 at 10:41:45 AM UTC-5, Shahadat Hossain wrote:
>
> Hi,
>
> In my survey, I have 10 sections in the questionnaire and each of the
> section will arise randomly for different respondents.
> But I expect, in total sample size, each of the section will be responded
> for fixed number of times.
>
> For example, for 100 sample size, I expect each section will be
> responded 10 times, but which subsection will arise to respondent will be
> determined randomly.
>
> Now, is there any option to control random number generation for fixed
> number of times using ODK tool in Excel (xlsx to xml) (sequence of
> appearing does not matter) ?
>
> Thanks in advance
>

That isn't a random choice. As you advance through your data collection,
the set of choices is increasingly restrictive until at survey 99, you
already know what survey 100 is required to use.

You are asking to start with a 10/10/10/.../10 set of values, and randomly
re-order them for a given device.

Then, for each given device, work through that list.

The simplest solution is to create a paper with this randomized list for
each data collector and have them read and enter the value on that list
(and then scratch it off as they use it).

Doing this entirely within ODK 1.0 tools without some sort of external
artifact like this is not easy.

···
On Mon, Nov 14, 2016 at 10:31 AM, Shahadat Hossain <shahadat.ecodu@gmail.com wrote:

Actually, For my survey purpose, I want to generate random number between
1 to 10, and each number will arise exactly 10 times (i.e: 1 will arise 10
time, 2 will are 10 times and so on.