As Mark Twain famously said – Lies, damn lies and statistics – it is really important to understand the statistics and the modeling behind survey data to make empirical sense of it.
My brother is doing his EMBA at HaaS – Berkley and he recently became the class president. Like all good people in leadership positions, he wanted to do a survey to gather the opinions of his constituents – the students! – Of course he called me (and no – I did not charge him) – gave him a free license to QuestionPro. The only unfortunate part was that I ended up being his customer support rep – How do you tell family to “Stop wasting my time?”
Anyways – I guided him on the questions types (matrix vs. checkbox – cognitive stress etc.) but then as usual something came to my attention when we were looking at the data. One of the questions they had was:
- Academic
- Alumni
- Career
- Community service
- Networking within EW and other MBA programs
- Social – Families with kids
- Social – Singles/Couples
Here is a screenshot of the pure frequency analysis of that question:
Now this data basically tells you that Career is the most important, but Networking is also relatively important. Career got about 26% and Networking about 21% – BUT keep in mind, this was a multiple select question. So, users could choose more than one option — In such a scenario lets see what the data looks like if we did TURF Analysis (BTW – If you want to learn more about TURF Analysis – see https://www.questionpro.com/article/turf-analysis.html)
So – if you did a “Reach” analysis, as opposed to a regular frequency analysis, you get that almost 80% of the users selected Career as one of the options — This gives you a very clear picture of how you can reach the most # of users – talk about Career.
This also helps you in determining Cost/Benefit analysis. Between Networking and Career, you got almost 100% of the users covered!
For those of you interested in viewing all the other data, click on the link below:
https://www.questionpro.com//userimages/sub-1/561282/13731/index.html
