Having trouble finding out how to estimate a criterion-related validity for my data. Also need to estimate Incremental Validity. What is the difference, and how would I enter it into SPSS?
Hi Ms. O'Toole! I'd like to ask you about the significance of correlation. If the output shows that the correlation is not statistically significant (p>0.05), can we say that there is no correlation between variable 1 and variable 2?
Hello, I would like to ask and I need the answer asap if possible please. I am trying to validate a 7 point likert questionnaire against a 6 point likert questionnaire. How do you do this if they have different scales? Thank you.
Does the sampling need to be equal for various factors? For example, does the number of women and men need to b equal? Please help me. I am really confused with what spling shall I use when doing factor analysis and taking some sources of bias as consideration (gender, course etc.). Thx
Aika Lu If you ask 5 different statisticians that question you'll get 5 different answers. Personally, I go for Factor Analysis is a fairly robust stat and with all the statistical options these days you can typically deal with unequal sample sizes statistically. Granted you don't want 5 females and 95 males or vice versa but 50/50 isn't really necessary for most stats these days. However, if this is for a dissertation or thesis the rule of law is "Whatever your chair says!"
Hi , I test this model in the pearson correlations I got for example positive .757, .544, .473 and in sig all they are 0.000 what does this mean ? is this good or bad ? I compare here only the general constructs which have another constructs
Hani- I am assuming you got a positive value of .757 for the first correlation and the sig value listed under it was .000. I would make sure to comment on 4 things when describing this result in a write up: 1. the value (strong correlation), 2 direction (positive), 3. statistical significance, 4. interpretation. So something like X and Y had a statistically significant, strong, positive correlation (r(put your n size here) = .757, p < .001). As X increases Y tends to increase as well. It's hard to give more than that without knowing your variables... When I am learning a new stat I always look through 5-6 journal articles that used it to see how it's typically written up so I know what to include and get some ideas on wording...
Siobhan O'Toole well after getting output i talk like about variable 1 had strong value% correlation with "variable 2" , i did in all of them like this , in the end i describe which correlatino the highest and which one the lowest , Example: for trust factor it has moderate 49.1% correlation with "benefits perceived". Where i can put n size ? do i need to put p < .001 ? and why ?
Thank you Ms O'Toole!
A new scale has concurrent validity with an existing scale of .80. What does this mean? Is this good? Bad?
Having trouble finding out how to estimate a criterion-related validity for my data. Also need to estimate Incremental Validity. What is the difference, and how would I enter it into SPSS?
Hi Ms. O'Toole!
I'd like to ask you about the significance of correlation. If the output shows that the correlation is not statistically significant (p>0.05), can we say that there is no correlation between variable 1 and variable 2?
Hello, I would like to ask and I need the answer asap if possible please. I am trying to validate a 7 point likert questionnaire against a 6 point likert questionnaire. How do you do this if they have different scales? Thank you.
Does the sampling need to be equal for various factors? For example, does the number of women and men need to b equal? Please help me. I am really confused with what spling shall I use when doing factor analysis and taking some sources of bias as consideration (gender, course etc.). Thx
Aika Lu If you ask 5 different statisticians that question you'll get 5 different answers. Personally, I go for Factor Analysis is a fairly robust stat and with all the statistical options these days you can typically deal with unequal sample sizes statistically. Granted you don't want 5 females and 95 males or vice versa but 50/50 isn't really necessary for most stats these days. However, if this is for a dissertation or thesis the rule of law is "Whatever your chair says!"
Hahaha... thank you! :)
Hi , I test this model in the pearson correlations I got for example positive .757, .544, .473 and in sig all they are 0.000 what does this mean ? is this good or bad ? I compare here only the general constructs which have another constructs
Hani- I am assuming you got a positive value of .757 for the first correlation and the sig value listed under it was .000. I would make sure to comment on 4 things when describing this result in a write up: 1. the value (strong correlation), 2 direction (positive), 3. statistical significance, 4. interpretation. So something like X and Y had a statistically significant, strong, positive correlation (r(put your n size here) = .757, p < .001). As X increases Y tends to increase as well.
It's hard to give more than that without knowing your variables... When I am learning a new stat I always look through 5-6 journal articles that used it to see how it's typically written up so I know what to include and get some ideas on wording...
Siobhan O'Toole well after getting output i talk like about variable 1 had strong value% correlation with "variable 2" , i did in all of them like this , in the end i describe which correlatino the highest and which one the lowest , Example: for trust factor it has moderate 49.1% correlation with "benefits perceived". Where i can put n size ? do i need to put p < .001 ? and why ?