I calculated the weighted proportion of individuals who were married (with both spouse absent and spouse present) for each month in the sample and restricted the sample to non-farmer private workers ages 20-60. Plotting a line chart of the time series revealed that there was a 3% increase in the amount of married individuals from December 1988 to January 1989. Was there a major change in the way that marital status data was collected in 1989 that could result in this type of increase?
I’ve been able to replicate the exact opposite of your observation. I’m not certain how you are specifying “non-farmer private workers”, but I’ve found a 3% decrease in married individuals between December 1988 and January 1989. This observation seems to be caused by the inclusion of the response category “separated”, which was introduced in January 1989. This seems to reduce the number of people reporting “married, spouse absent”, who instead report “separated”.
I hope this helps. Let us know if you have any additional questions.