[IA] Evaluating results from a survey?

[Da2Shae]

OK, I'm doing my IB Math Studies Internal Assessment now. What I'm doing is evaluating my survey results.

I conducted a survey of 50 boys and 50 girls (100 total) of which subjects they preferred the most. I gave them a list of 7 subjects each with the choices "I Like/Do not Like this subject"

I have the raw data, but I have a vague idea of what I want to do with it.

I wanted to use the standard deviation, chi-square into this data but im not sure how...

Any ideas on what I can do with these results? Any more processes? Also, whats the word count for the Math IA?

[Asheee]

I believe there is no max word count, but the min is 2000 words. You can draw tables of the results, ummm, maybe find the correlation of gender and subject preference?

It is probably a bit too late for this suggestion, but I think your project would be heaps easier if people had to give a numerical value to what they think of the subject, maybe from one to 10, because then you could have done mean,median and mode based on what score each subject got. If worse comes to worse, you could make up those results based on what you found in your initial survey.

I hope i've helped a little bit.

good luck!

[Da2Shae]

I believe there is no max word count, but the min is 2000 words. You can draw tables of the results, ummm, maybe find the correlation of gender and subject preference?

It is probably a bit too late for this suggestion, but I think your project would be heaps easier if people had to give a numerical value to what they think of the subject, maybe from one to 10, because then you could have done mean,median and mode based on what score each subject got. If worse comes to worse, you could make up those results based on what you found in your initial survey.

I hope i've helped a little bit.

good luck!

Lets just say that I do...

what would I do then?

[Asheee]

If you do have numerical values as opposed to "like/do not like"?

Well, you could work out the average value everyone gave for each subject, the median and the mode. You could then see if females had a preference for a particular subject based on which one was the more popular subject.

However, to do more advanced calculations like correlation and linear regression, you will need both parts of your survey to be numerical values, so you could perhaps widen female and male into age groups, or perhaps focus on one gender only with a wide age group (say from 10-30 years). That way you would be able to see a correlation such as "as the age of the females increases, the score given out of 10 for maths increases" or something like that.

Because basically, for a linear regression line on a graph, both axes will need to have numerical scales. such as Age on the x-axis going from 10-30 (or whatever you choose) and for the y-axis (Subject score) 1-10.

I hope that makes sense!