I´m using sequence analysis. I have a question about how to deal with missing data within the observation window. The starting point of the analysis is when respondents leave secondary school (t0). I want to examine respondents' life course over a time-span of 36 month after leaving school. The dataset contains longitudinal information of repondents toward their educational histories. I arranged the data in the ‘states-sequence’ (STS) format. So in each month the dataset provides information on respondents' status (for example "employed" or "training"). For 58 % of the sample the data provides information over the whole observation window. So for this group I can tell in every single month what they are doing. The sequences of the rest of the sample are shorter. Thus, the length of the sequences is not the same for all respondents. How do I handle sequences of respondents that end before month 36. What would be the way the missing values should be handled in TraMineR?
-
$\begingroup$ There are different ways to handle missing values in TraMineR. See Sec. 6.5 of the User Guide. The choice between e.g. dropping or treating as an additional state will depend on the kind of analysis you plan to do. $\endgroup$– GilbertApr 4, 2019 at 15:34
1 Answer
(Sorry, not enough reputation to add a comment.)
Are you sure sequence analysis in required for your sort of problem? I'd suggest using time series analysis for interval variables (in case you're monitoring characteristics that can be projected on a numeric scale). This will help you to deal with missingness in an 'interval' way and apply various sorts of imputation methods like those mentioned here: Imputation methods for time series data .
If you deal specifically with ordinal data (say, switching from state to state monitored periodically), the recommended way is to convert your data to patterns (sequences of states) and model transitions between patterns, with methods like Discrete Autoregressive mentioned here: Analysis of Ordinal Value Time Series .
