Below is a simple scattergram of the correlation between the vote for the JVP in 1982’s Presidential election and the turnout level in the 1988 Presidential election. I restrict the sample to 19 primarily-Sinhalese districts, since the LTTE and IPKF throw everything off in Tamil areas. In 1982, the JVP was legal and putting huge effort into electoral mobilization, hoping to lead a leftist resurgence in Sri Lankan politics.
While acknowledging numerous caveats (small # of observations, possible confounders, low vote totals in 1982, etc.), what we see here is neither shocking nor unimportant: areas that voted for the JVP in 1982 were substantially less likely to turn out in 1988.The literature is unanimous in arguing that the JVP actively tried to suppress turnout in 1988 and 1989, so lower turnout, all else being equal, is likely to have something to do with JVP activities (r = -.6375). The relationship between 1982 UNP votes and 1988 turnout is the opposite (though it’s less powerful).
This suggests a political-spatial basis to the location of JVP’s ability to suppress voting, rather than a purely military, geographic, or some other functional logic of armed group activity in war. Which shouldn’t be shocking, but does helpfully direct us to get a better handle on the historical roots of JVP support, rather than focusing exclusively on wartime and tactical dynamics.