Im doing my thesis, i have conducted a survey then i have prepared graphs and percentages for the results in the analysis phase. i need to do now higher statistical analysis, i dont know which statistical method is suitable for my data. my survey is multiple choices for students and instructors that are enrolled in online universities. as example,
when you send a message to someone in the university, how long it takes to get back the answer?
a. Same day b. Next day c. 2 to 3 days d. more than 3 days
does your university use social media for Education purposes?
a. Yes b. No and I don’t care c. I wish it will be used d. I hope it will never
What is the best way for you to communicate with your instructor or class-mates?
a. University Email b. Social Media c. Mobile d. b or c
Hello Nora Hadjazi.
The analisis of survey data is normally by the use of descritive statistics i.e., you can present the results using n and % for each question.You also can present them using mean and standard deviation.
If you want to know if there are some statistically significant tendency of your data more to a. option or more to d. option (mainly in questions where the options represent a degree, i.e. for example a. very bad, b. bad, c. neutral, d. good and e. very good) you can apply the One Sample t test by defining your cutoff point of c. neutral in this example. The test will give you a p value. If the p value was significant and the mean was 3.6 (in a likert scale from a. to d., i.e., in numeric: from 1 to 4), the responses will be tendent more to response d (in numeric: 4).
If you want to compare your responses with a categorical variable/outcome variable, if you present your descritive data as n (%) you can use the qui-square test. If you present your responses as mean and standard deviation, you can use the independent two samples t test.
I hope this has helped.
You could consider each of the three questions you outlined is nominal data (also known as categorical) and subject them to a multiple correspondence analysis. This would give you a visual map where you would zone the map into quadrants and members of a quadrant would be considered to have similarity/correspondence
Dear Nora,
I agree with Ms. Azevedo. This survey seems to be descriptive. Maybe you could try to choose one question to be the regressand, and then apply multinomial regression to see if there is any correlation structure in those questions.
Thank you very much for your replies. my main issue that im not able to find suitable analysis methods rather than descriptive analysis, i have found the percentage of each question and connect it to the research i have done then i have also find the standard dev, mode, median and so on. But my prof is asking me for high level statistics but am not able to apply any high level statistics on my current data.
one more question, using chi square test, how do we calculate the expected value ?!
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis