Practical Issues on Using Weighted Logrank Tests with Interval-Censored Events in Clinical Trials - Interval-Censored-Time-to-Event Data: Methods and Applications - page 362

scores=\logrank2" gives the model with the scores c i in Equation (13.5)
proposed by Finkelstein (1986);
scores=\wmw" gives the proportional odds model with the scores c i in
Equation (13.7) which reduces to the Wilcoxon-Mann-Whitney test with
exactly observed events;
scores=\normal" assumes F is normal, which is equivalent to assuming F
is lognormal because h() is an unspecied monotonic transformation;
or
scores=\general" which allows general F.
Themethodargument determines how the inferences will be made. Here are
the options:
method=\exact.ce" gives the complete enumeration exact permutation p-
value (Section 13.3.4);
method=\exact.network" gives a network algorithm exact permutation p-
value for the two-sample case only (Agresti et al., 1990);
method=\exact.mc" gives a Monte Carlo approximation for the exact per-
mutation p-value; different values of b are made using themcontrol ar-
gument (default is b = 999);
method=\pclt" gives the normal approximation to the permutation p-value;
method=\scoretest" gives the score test with the ad hoc adjustment sug-
gested in Fay (1996) if needed (Section 13.3.5);
method=\wsr.HLY" gives the multiple imputation method (within-subject
resampling method) of Huang et al. (2008) (Section 13.3.6);
method=\wsr.pclt" gives the multiple imputation method where the vari-
ance is estimated by the permutation variance for each imputation (Sec-
tion 13.3.6);