Dear Sir Thanks for the informative video. I have a query about using regression in longitudinal data (I have survey data collected at 4 different times). Now how to analyze this data? (Say a relationship Satisfaction--> Loyalty. Now I have four Beta values at 4 times. How can I use this data to help formulate a longitudinal relationship between Satisfaction and Loyalty?) Thanks
That depends on what you want to say with the data and what kind of assumptions you make about persistent difference (unobserved heterogeneity) and dynamic effect (e.g. autocorrelation over time.) With longitudinal data, the hard part (at least to me) is to figure out the right question to ask. Are you interested in the within effect, between effect, or contextual effect? Cross-sections or dynamic effects? If you are not familiar with these concepts, these two playlists contain a lot of relevant material ASRM Unit 7: Multilevel models ua-cam.com/video/dx8lZnoY8hU/v-deo.html ASRM Unit 8: Longitudinal analysis and multilevel SEMs ua-cam.com/video/c4iq8KCcy8U/v-deo.html
@@mronkko Thank you sir! i will refer these videos for more details! meanwhile to answer what is my research question: i have a model where loyalty is predicted by satisfaction and trust. i collected data from 200 people in 4 waves in 3month interval. my hypthesis is that with time: 1.impact of satisfaction on trust decreases 2.impact of trust increases i have some interesting results and i would like to use the longitudinal nature of data with the right analysis and target top journals.
@@roshr9620 That is starting point. The problem with longitudinal data is that you can estimate different kinds of effects and need to choose the right one. Also you need to consider the possibility that unobserved heterogeneity or autocorrelation over time drives the results. This is on top of considering which variables should be in the model.
Thanks for the comment. Can you elaborate? Are you talking about the content or the presentation? I shoot my video in one go and they are not scripted, so the production value is not as high as on some other UA-cam channels that have more resources.
![ПОЛНАЯ ИСТОРИЯ ЭКЗОРЦИЗМА [Топ Сикрет]](http://i.ytimg.com/vi/qBySr6Bt0B8/mqdefault.jpg)
❤️❤️❤️👍👍👍 spr
great !
You are welcome
Dear Sir Thanks for the informative video.
I have a query about using regression in longitudinal data (I have survey data collected at 4 different times).
Now how to analyze this data?
(Say a relationship Satisfaction--> Loyalty. Now I have four Beta values at 4 times. How can I use this data to help formulate a longitudinal relationship between Satisfaction and Loyalty?)
Thanks
That depends on what you want to say with the data and what kind of assumptions you make about persistent difference (unobserved heterogeneity) and dynamic effect (e.g. autocorrelation over time.) With longitudinal data, the hard part (at least to me) is to figure out the right question to ask. Are you interested in the within effect, between effect, or contextual effect? Cross-sections or dynamic effects? If you are not familiar with these concepts, these two playlists contain a lot of relevant material
ASRM Unit 7: Multilevel models
ua-cam.com/video/dx8lZnoY8hU/v-deo.html
ASRM Unit 8: Longitudinal analysis and multilevel SEMs
ua-cam.com/video/c4iq8KCcy8U/v-deo.html
@@mronkko Thank you sir!
i will refer these videos for more details!
meanwhile to answer what is my research question:
i have a model where loyalty is predicted by satisfaction and trust.
i collected data from 200 people in 4 waves in 3month interval.
my hypthesis is that with time:
1.impact of satisfaction on trust decreases
2.impact of trust increases
i have some interesting results and i would like to use the longitudinal nature of data with the right analysis and target top journals.
@@roshr9620 That is starting point. The problem with longitudinal data is that you can estimate different kinds of effects and need to choose the right one. Also you need to consider the possibility that unobserved heterogeneity or autocorrelation over time drives the results. This is on top of considering which variables should be in the model.
I don't think this is a great video as the content is not smooth at all.
Thanks for the comment. Can you elaborate? Are you talking about the content or the presentation? I shoot my video in one go and they are not scripted, so the production value is not as high as on some other UA-cam channels that have more resources.
@@mronkko But we love still the same Prof. Opened my eyes to the possibilities of longitudinal analysis.