Which Tests are appropriate?

#1
I have created a survey where participants state their agreement with my statements on a 7 point likert scale. Also 3 category questions, eg their sex, and age group.

I have formed hypotheses on each of my likert statements and need a statistical test to work out how the result agree with my hypotheses. And also another to work out what role their sex and age group play in the reults.

Unfortuantely, my knowledge of this is limited. Somehow I reached the conclusion that regression testing and chi-squared tests would do what I need, but really I have no clue as to the the usefulness of these tests.

Looking forward to your responses. :)

Free Soul
 

JohnM

TS Contributor
#2
We'll need to know more about how your hypotheses are stated before we can comment on what type of statistical test to use....
 
#3
Basically it's a survey about e-commerce websites and customer preferances. So for example one of the likert statements is:-

1 = very stongly disagree 7= very strongly agree
Time is an important issue for me. I want to shop as quickly as possible.

1 2 3 4 5 6 7

And my Hypothesis is that most participants will be agreeable with this statement eg, will pick 5,6 or 7

Also there are 5 seperate sections with 4 likert statements associated with each one.
E.G. Section 1: Percieved risk of a transaction.
My hypothesis is that most users will be agreeable with the 4 likert statements in this section. Assuming the results are in favour of my hypotheses, i can say that the results from my statistical test support my hypothesis.

I have no hypotheses relating to the role of age and sex on the agreement with the statements, so it is mainly based on obsevation of the results. I need a statistical test to draw conclusions from this set of results, eg how the results differ between people aged 10-19, 20-29,30-39,40-49 and 50+. And also how the results differ between the sexes.

I hope this is the information you need

Thanks for you time :)

Free Soul
 
Last edited:
#5
Thanks alot, thats a great help to me! Will this also work with the last bit i talked about where I have no hypothesis yet for age and sex?

Is regression testing useless for this data? What other statistical tests may be useful on this data? I know that it depends on what conclusions I am trying to draw, but to be honest I just need to do as many relevant tests as possible.

Thanks in advance. :)

Free Soul
 

JohnM

TS Contributor
#6
Yes, it will work for those as well.

Regression may work, but not so well with categorical / ordinal data.

I would just keep it simple - use the simplest method that will answer the research question(s) - don't try to throw a lot of tools at the problem.
 
#7
JohnM said:
I would just keep it simple - use the simplest method that will answer the research question(s) - don't try to throw a lot of tools at the problem.
That's the best news i've heard all day :D
Thanks so much for your help :tup:

Free Soul