I know it’s frowned upon to ask for help with assignments, but I’m at my wit’s end. I am writing a research proposal, and I can’t figure out which data analysis technique to use. The lecture concerning the subject won’t be until 4 days before the assignment is due, and while I have the lecture notes, they don’t mean much to me.
So, I really need help!
Okay, say I need to measure the proportion of people loyal to a store. I ask the 5000 buyers 3 questions related to loyalty, with the answers on a scale from 1 - 5 with people choosing “5” the most loyal. My hypothesis is “20% of Myer customers are loyal to Myer”. How can I test this?
I have made up a mock frequency distribution table, and I realise I will need to find the mean, SD, etc. But other than that, I am lost. Chi test? T-test? Z-test? How do I define what is “loyal”? People who score top 2 on the scale? Please help! Ultrafilter, I know you answered a stat question before… Anyone?
I have a feeling I’m really off track… I think most of those test relationships between variables, and I don’t need to test that at all. I can tell I’m thinking too hard on this…
First you’ll have to create a strift definition of “loyal to Meijer”. If the total possible “points” on your survey is 15, then perhaps ‘loyal’ is 12, or 10, or 13. You have to choose. Maybe one of the questions is more important than the others.
Once you have defined that condition, you can simply do a 1-proportion z-test. With your large sample of 5000, and a .05 level of significance, I calculate that 1047 of your 5000 (20.92%) will have to respond in the manner you’ve defined as “loyal,” for the data to have significance.
By the way, is this for an introductory statistics course? I’ve just finished a one-year AP high course in statistics, and the problem seemed fairly straightforward. Stats is my best subject and is integral to my career aspirations, though.
You first have to define what “loyal” means. It cannot mean that a person will always shop at Myer no matter what. This never happens. As an example, let’s say you have a person who routinely shops at Myer for groceries once a week. She needs to get some beer for a party. The carryout is 1 mile away, while Meyer is 20 miles away. Would you say she’s “un-loyal” for buying the beer at the carryout?
So you first have to very carefully (and statistically) define what “loyalty” means. (And be reasonable in your definition.) Once you do that, you design questions to determine the percentage of people who are loyal.
Thanks everyone for your replies. iwakura43, the course I’m doing is Business Research Methods. I completed a statistics course to get into this one, but it was 2 years ago, and I’ve forgotten everything. I did well too.
andymurph64, thanks for your kind offer. May I email you my data analysis section when I’ve finished so that you can offer suggestions?
I’m going to look up z tests in my text, hopefully from that I will be able to sort it out. I’ve been to a tute now, and so I know a little bit more.
It’s Myer, not Meijer. It’s an Aussie department store.
What formula are you using? In my text, the formula for z test for the proportion is z = X - np/sqrt np(1-p), where X is the number of successes, p is the hypothesized proportion of successes, and n is the sample size.
(So, I’m not sure where you’re getting the level of significance from.)
And now I’ve just realised I can’t use the Z test because I don’t know the population summary statistics. I guess this means I just calculate the proportion as Number of successes/Sample.
I knew I was thinking too hard. Thank you all for helping.
I can tell I’m thinking too hard on this…
May I email you my data analysis section when I’ve finished so that you can offer suggestions?