survey questions score and aggregation?

[#1]

Hel\lo!

I had some help here for creating an inline load to map survey responses to numbers...

the next part however is throwing me off...

I have a set of survey responses, and each questions have a different number of responses..

I'm trying to now show the avg for each question (eg - question 1 has an avg of 4.54 based on the formula ((37*5)+(19*4)+(2*3)+(1*1))/(37+19+2+1)

but also I want to show the avg for all questions in each (section)

When I try to do something like AVG(AGGR ( COUNT (DISTINCT surveyID) question), I get 2.0

sample attached, hope someone can help!

[#2]

I see!

thanks for your help and answers!!

[reddy-s]

Always Welcome!

[#4]

Hello again,

OK, I realized that question #17 has different responses (e.g. the responses don't fall into SA,A, N, D, SD...they're numerical responses 1-4 on how much time spent preparing for course.

If I try to load an inline table for that with

mapping * LOAD Inline, all the questions now get 1-4 in addition to SA,A, N, D, SD...

is there a way of only showing SA,A, N, D, SD for questions 1-16, and have 17 use another set of responses?

many thanks for your help! I am starting too feel more comfortable in Qlik!

[reddy-s]

Hi Adrian,

Absolutely, by making use of the set analysis expression you can handle the way the calculations are performed. The key to this is the data model which is well built. Once that is ready , managing calculations on the front end becomes easy.

Hope this helps!

Thanks,

Sangram.

[#6]

hm OK, I'l give it a try, I have a rough idea where to begin..

something along the lines of sum({<QUESTIONS=,Code = -{17}>})/count({<QUESTIONS=,Code = -{17}>} would yield 1-16?

would I have to load another inline table to account for the other set of responses (1-4)?

apologies for being still confused!

[reddy-s]

Hi Adrain,

No problem at all. Yes you can do that for now. but inline table is a static table. I would suggest you to have all the dimension tables in the datasource instead of having numerous inline loads. (Just as a best practice)

Thanks,

Sangram

[#8]

Yes, I think having the dimensions in my table would be a better practice.

now..how would I still how all 17 questions on one bar chart? that is 1-16 with the SA,A,N,D,SD responses and #17 with the 1,2,3,4 response.

I think the quick hack would be to create 2 bar charts..one for Q1-16, and another for Q17..

many thanks!

[reddy-s]

Yes, you will be needing two different measures to achieve this.

[#10]

thanks for confirming I was on the right track!