2

I am currently analyzing longitudinal data that was repeatedly measured (four times) but 1000 participants were lost to follow up after the first survey.

Shall I include or exclude these participants in the longitudinal analysis that have only participated once?

Glen_b
  • 257,508
  • 32
  • 553
  • 939
Eshetie Dres
  • 21
  • 1
  • 1
    Related question [here](https://stats.stackexchange.com/questions/90677/should-the-outcome-variable-be-measured-at-least-twice-for-a-longitudinal-study) but I am not sure the answers quite respond to what you seek – Glen_b May 23 '19 at 01:00

1 Answers1

0

If you have a number of 'participants' engaging is generating a set of data over time, this is called "panel data". (See Cheng Hsiao's (sp.?) excellent book on this topic, here: https://www.amazon.com/Analysis-Panel-Data-Econometric-Monographs/dp/B000KCI5P0 )

In general, with panel data, if you have one data point for N of M subjects, they are useless. Panel data analysis is about taking advantage of both the cross-sectional and time-series components of a data set, and one entry sort of violates the linkage of value.

eSurfsnake
  • 1,004
  • 5
  • 13