S01E04-1-2b SPSS Basics - Multiple response and semi-closed
HTML-код
- Опубликовано: 24 сен 2016
- Instructional video on how to enter three types of basic survey questions into SPSS: a semi-closed question that has an option like 'please specify____', a question which allows multiple choices, and a question that allows multiple choices and has an option like 'please specify___'.
Visit the companion website at PeterStatistics.com
extremely important brother.
Thaaaaanks!! super helpful :D
have learnt a lot, thanks
glad the video helped. Good luck with your study/research/project.
Thank you my life saver. I learn a new things.
Glad to help
How to utilise the same when you have more than 1000 responses? Kindly help.
this shows how to enter the survey itself. SPSS can handle far more than 1000 respondents. You'll need to enter them all in the Data View. If you were looking for how to analyse a multiple response question, then perhaps have a look peterstatistics.com/CrashCourse/5-ThreeVarPair/binary/MultipleBinaryPaired0.html and go over the chapters from the menu on the left. In each there are also videos on how to do those with SPSS.
Hope this is what you meant.
@@stikpet thank you so much for your help. I'm looking for the procedure where I can put yes and No as values for multiple responses. Right now, my response is in one column ( music, dance, draw, sing) and there are 400 such responses. I need to move that into spss for analysis. Request you to please share with me how to handle this bulk data into individual variables. Thanks .
If I understand correctly in your data you have responses like {music, dance, draw, sing} for one respondent and then perhaps another with {dance, draw}, and another with {music, draw, sing}, etc. Unfortunately I don't know of any easy way to convert this with SPSS. Perhaps its possible and if you do find out, please let me know.
The way I solve this, is by using Excel, but then still it is quite some work. I'd copy the one column from SPSS to Excel. Then in Excel use the text to columns option that will split the separate responses, but I then have:
row 1: music | dance | draw | sing
row 2: dance| draw | empty|empty
row 3: music| draw | sing| empty
So with sorting (the entire table!) on the various columns, then moving each to the appropriate one.
Then import all of that back into SPSS.
Sorry, don't know of any easy way of doing this :-(
Does it matter if I use eg 1=yes, 2= no?
not much, the advantage of using 0 = no, is that if you have multiple of them it is for example easy to just sum them up since 0 was no and 1 was yes. But you can also change it later when needed using a re-code, so I wouldn't worry too much about it.
How can I do it when someone gave multiple answers for "please specify"?
unfortunately I think you'll then need to create a separate variable for each, so something like alternative1, alternative2, etc. However will you be doing many analyses with these typed in answers? Often they serve just for a bit of interesting information, but not much analyses is done with it (if any). So if one person just listed a few options I wouls simply type those together.
Hope this helps a bit.
@@stikpet Alright. Thanks for your answer! :)
@@katja5121 no problem. Glad to help and good luck with your research/study/analyses.
How would you analyze a multiresponse Qs?
I'd get an impression of the results using a frequency table, then visualise the results using a bar-chart. Then perform a Cochran's Q test. If that is significant then follow up with a Dunn test and add effect size eta-squared.
For each of these steps you can look on my site at peterstatistics.com/CrashCourse/5-ThreeVarPair/binary/MultipleBinaryPaired0.html
Thank you so much! Btw, I would like to ask if you have spss tutorial in analyzing patterns (if no, skip to q4, something like that) with multi response question number
Not really sure what you are looking for. With a multiple response question, each option becomes a variable. You could cross then each option with another variable (e.g. gender, income, etc.), and look for differences between those who selected that option and those who didn't. Treating the option of the multiple response question than as a nominal variable (yes/no).
If you want to know if the results on let's say gender is different from those who skipped the multiple response question, I'd create a new variable (lets call it 'skipped') that has 'yes' if the person skipped the multiple response question and 'no' otherwise, and then perform the analysis as described on my site for two nominal variables with 'gender' and 'skipped'.
Hope the above helps.
I understand sir, thank you so much, it helped me!
Glad it helped.
On a side note, RUclips has cancelled my partner program since I don't have enough subscribers (currently around 543, should be 1000 😞), so it would really help if you subscribe (and all your friends) 😉