diff --git a/permute/core.py b/permute/core.py
index fa7533c..f9f21b8 100644
--- a/permute/core.py
+++ b/permute/core.py
@@ -6,7 +6,7 @@
 import numpy as np
 import warnings
 from scipy.optimize import brentq, fsolve
-from scipy.stats import ttest_ind, ttest_1samp
+from scipy.stats import ttest_ind, ttest_1samp, rankdata
 from fractions import Fraction
 
 from .utils import get_prng, potential_outcomes, permute
@@ -52,7 +52,7 @@ def corr(x, y, alternative='greater', reps=10**4, seed=None, plus1=True):
     return tst, pvalue, sims
 
 
-def spearman_corr(x, y, alternative='greater', reps=10**4, seed=None, plus1=True):
+def spearman_corr(x, y, alternative='greater', reps=10**4, seed=None, plus1=True, method='average'):
     r"""
     Simulate permutation p-value for Spearman correlation coefficient
 
@@ -72,6 +72,11 @@ def spearman_corr(x, y, alternative='greater', reps=10**4, seed=None, plus1=True
         flag for whether to add 1 to the numerator and denominator of the
         p-value based on the empirical permutation distribution. 
         Default is True.
+    method : {‘average’, ‘min’, ‘max’, ‘dense’, ‘ordinal’}
+        passed to scipy.stats.rankdata. 'average' uses midranks.
+        'ordinal' uses distinct ranks, with ties broken by order
+        of appearance in the original list.
+        Default is 'average'.
 
     Returns
     -------
@@ -79,8 +84,8 @@ def spearman_corr(x, y, alternative='greater', reps=10**4, seed=None, plus1=True
         Returns test statistic, p-value, simulated distribution
     """
     
-    xnew = np.argsort(x)+1
-    ynew = np.argsort(y)+1
+    xnew = rankdata(x, method=method)
+    ynew = rankdata(y, method=method)
     return corr(xnew, ynew, alternative=alternative, reps=reps, seed=seed)