Solved
Two blocks (randomized) with identical survey - Combine results
Hi Community,
I created a survey with 20 questions in one block and added another block with the same 20 questions, however, added some additional information to one the questions and then randomized the two blocks.
Now, I received almost 300 responses in total, however, whenever I download the data to SPSS I have 150 responses in one block and 150 responses in the other one and instead of 20 variables I have a total of 40 in the output file. It is really difficult to analyze the data as the data is not combined but rather split up into 2 variables. Is there any way to combine the data so Question 1 of block 1 and Question 1 of block 2 are combined as one variable.
Thanks in advance for the help.
Best answer by PeeyushBansal
This is simple since you are showing one block out of two to the respondents, so you have two sets of data for 150 records in one set of variable and for other 150 in another set of variable right.
So in SPSS just create one extra set of variable and move data from these two set of variables in third set of variable. And then delete these existing two set of variables and also create one extra variable say brand which store 1 or 2. 1 for block1 and 2 for block2.
Using this variable now you can analyse your result.
View original
+20I am not sure why you are surprised by this result. This is how you have set up your survey. 20 questions in one block and another 20 question in the second block (even if these 20 questions are similar it will be considered as unique).
The manual solution would be to create 20 embedded data in survey flow such as Q1, Q2, Q3... Q20 than in Data & Analysis, edit response and provide Question 1 of block 1 and Q1 of block 2 response in embedded data "Q1".
If you want to use the same survey again then in survey flow (at the end of it) you will have to assign the response of Question1 of block 1 and Question 1 of block 2 to embedded data "Q1"
Thanks for your help.
You are right, I am not really surprised by the result, I guess I jumped into publishing the survey too quickly.
Regarding the analysis, by following your approach, can I then still filter the results for the one question to which I added additional data? For my analysis, it is essential to do a T-test for these two groups. So I want to show demographics data for all 300 responses, but for one question of the survey, I would like to compare the means of 150 to the other 150.
+20Sure, for this you should add 2 more embedded data such as "block1" and "block2".
If the respondent has answered block 1 question then embedded data "block1" should be "true" likewise for block2.
Then you would be able to filter block-wise data.
+39This is simple since you are showing one block out of two to the respondents, so you have two sets of data for 150 records in one set of variable and for other 150 in another set of variable right.
So in SPSS just create one extra set of variable and move data from these two set of variables in third set of variable. And then delete these existing two set of variables and also create one extra variable say brand which store 1 or 2. 1 for block1 and 2 for block2.
Using this variable now you can analyse your result.
Leave a Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.