Monday, August 25, 2014

Hints on Statistics Inquiry

OK Review Of Where You're At
In your pairs now you should have taken some time to think about how to approach this idea of whether Holden or Ford owners *love* their cars more.

I've deliberately picked the word love as it's means different things to different people (and therefore you may have different ways of approaching your inquiry).  This is fair too as Statisticians get paid to work for clients and sometimes those clients will ask for things that aren't too specific. 

I'm expecting that you've began to give some definitions (spelling out exactly what you're doing).  There are some key things to cover - who/what/where/when.  I.e. if you data comes from Trademe Website - you need to state this.  If you're looking at price there are several to go from (current bid, reserve price, buy now price, 'or nearest offer' price).  You need to specify these - justify which one you're picking. 

I'm also expecting that you've been looking at the data - how much is there?  How many XR6's are advertised?  Can you use all of these?  If too many you will need to sample (pick a few) - how are you going to do this (spell it out in your presentation)?.  How does this compare with the Commodores?  Are all the XR6's the same (if not why not) - how are you accounting for these?  This is an important aspect of the PPDAC cycle as if your data is bias from the start it will lead to bad conclusions. 

You want to think about your overall presentation (I'm thinking most pairs will do a PowerPoint - but that is not necessary the case).  Try to balance your time - don't get sucked into spending too much on making things pretty and sacrifice your actual processing time. 

From here you should start to think about finding the measures of the middle (mean, median, mode) and the measures of spread (range, upper and lower quartiles). 



1 comment:

  1. Problem - I wonder if Trademe data (August 2014) will indicate whether Holden Commodore drivers love their cars more than Ford XR6 drivers.

    ReplyDelete