I am pretty new to stat and having read some articles about applying random forest algorithm. I have a employee survey return which contains approx 200 questions say 50 questions per a theme so there are 4 themes (e.g how happy you are with your job, your line manager, your customers etc...)

I hope you guys can please give your insight in terms of

  • How to find out which questions are the key factor / influence for overall score.
  • Any other algorithm I should use?

Some examples

I am clear about what I am expected to achieve in my job I understand how my work contributes to the success of the Company My job makes good use of my skills and abilities I have enough information to do my job well I enjoy my job My line manager encourages my ideas and suggestions for change My line manager provides regular feedback on my performance My line manager recognises and acknowledges when I have done my job well I feel my line manager will listen to my concerns and provides support when necessary The people in my team co-operate to get the work done I am satisfied with the recognition I receive for doing a good job Poor performance is dealt with effectively where I work My line manager is considerate of my life outside work, when appropriate My line manager motivates and inspires me to be more effective in my job I am encouraged by my line manager to develop my knowledge and skills During my last performance appraisal, I was given constructive feedback

The above survey can be scored from 1-5 where 1 is strongly disagree and 5 is strongly agree to the statements.

At the back of the survey we also capture the respondents data e.g age, gender, location (hopefully to use this to cluster the group of respondents etc..)

So the key factor I want to use is "I Enjoy My job" and what influence employees to be happy at the workplace.

Any sort of advice / insight i would be most grateful

Thank you Peddie

  • $\begingroup$ We will need a clear idea of what question you're trying to answer. You write that you want to know which questions are the key factor, but "key factor" is entirely undefined. What do you want to know about your data? $\endgroup$ – Sycorax Jul 13 '15 at 19:29
  • $\begingroup$ @peddiepooh I have tagged this post as analysis. Please review the wiki and edit the post so that it gives a sufficient amount of detail for informative answers. $\endgroup$ – AdamO Jul 13 '15 at 20:20
  • $\begingroup$ @PeddiePooh Could you clarify about what are your input variables (predictors) and what is your output variable (what you want to predict). $\endgroup$ – Antoine Aug 15 '15 at 18:12
  • $\begingroup$ I would like to predict the survey result based on staff's length of service, payscale and location. Thank you :) $\endgroup$ – PeddiePooh Aug 15 '15 at 20:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.