topic Re: How to analyse based on age in Discussions

How to analyse based on age

ninasor — Sun, 03 May 2020 10:50:50 GMT

I’m trying to analyze some data. I want to to see if there’s any similarities in the answers based on age. What’s the best way to do this? The age is stated by the exact number, so numbers from 20-67 is in the data.

Re: How to analyse based on age

P_Bartell — Sun, 03 May 2020 11:30:05 GMT

I have three rules for successful data analysis:

1. plot the data.

2. Plot The Data.

3. PLOT THE DATA.

JMP's Graph Builder is usually a good place to start.

But we're gonna need some more details such as:

Are the response variable(s) numeric, ordinal, categorical, or free form text? Are you interested in building a model of some type? What is the practical question you are trying to answer? Did you conduct some kind of designed experiment? Is the time order in which the responses were collected important to you? For starters.

Re: How to analyse based on age

ninasor — Sun, 03 May 2020 12:00:36 GMT

I transmitted data from a survey I made using Qualtrics. When I asked for age, the person answering had to type in his/her age, example: 32. So all the data collected has already been automatically plotted into JMP. The time order is not important. My hypothesis is that the respondents age has a correlation with their attitudes (based on questions I have asked). I want to test this, but I'm experiencing some difficulties based on the fact the way I collected their age. I am not that experienced with JMP, but I was hoping to perform a k-means cluster analysis, t-test and ANOVA.

Re: How to analyse based on age

P_Bartell — Sun, 03 May 2020 12:59:04 GMT

If you are not familiar with JMP my suggestion is to complete the JMP New User Welcome Kit before you start your analysis. It can be found here:

https://www.jmp.com/en_us/events/getting-started-with-jmp/new-user-welcome-kit.html

Re: How to analyse based on age

ninasor — Sun, 03 May 2020 13:04:30 GMT

Thanks, I have already completed it.

Re: How to analyse based on age

Georg — Sun, 03 May 2020 18:50:34 GMT

I suppose, you want to know whether the answer you collected is dependent on the age of the participant.

In what form do you have the answer?

If it is e.g. a continouous variable you could simply plot the answer over the age, and see if there is a relation.

If it is yes or no, you could e.g. compare the age-distribution for yes and no answers (Fit y= by x).

You could Group the age, in two, three or more Groups ...

I very much like to see the Distribution of all cols to get a first Impression.

There are many many possibilities to do this in jmp. See the example table with a few scripts.

There is one col for age and two answers generated randomly (no dependence on age), and two biased answers. With fit y by x you can e.g. see the result of a t-test, showing you if the average age of no is different from average age of yes.

You can play around with the formulas and tests, and see what happens.

BR., Georg