diff --git a/tests/test_ttest.py b/tests/test_ttest.py
index fd5e78d..cf2e60b 100644
--- a/tests/test_ttest.py
+++ b/tests/test_ttest.py
@@ -40,6 +40,7 @@ def test_ttest_ind_ol6_1():
 	assert result.delta.ci_upper == approx(0.808, abs=0.01)
 	
 	expected_summary = '''t(18) = 4.24; p < 0.001*
+μ(Fresh) (SD) = 10.37 (0.32), μ(Stored) (SD) = 9.83 (0.24)
 Δμ (95% CI) = 0.54 (0.27–0.81), Fresh > Stored'''
 	
 	assert result.summary() == expected_summary
diff --git a/yli/sig_tests.py b/yli/sig_tests.py
index 2a63819..92bc5bc 100644
--- a/yli/sig_tests.py
+++ b/yli/sig_tests.py
@@ -35,13 +35,25 @@ class TTestResult:
 	See :func:`yli.ttest_ind`.
 	"""
 	
-	def __init__(self, statistic, dof, pvalue, delta, delta_direction):
+	def __init__(self, statistic, dof, pvalue, group1, group2, mu1, mu2, sd1, sd2, delta, delta_direction):
 		#: *t* statistic (*float*)
 		self.statistic = statistic
 		#: Degrees of freedom of the *t* distribution (*int*)
 		self.dof = dof
 		#: *p* value for the *t* statistic (*float*)
 		self.pvalue = pvalue
+		#: Name of the first group (*str*)
+		self.group1 = group1
+		#: Name of the second group (*str*)
+		self.group2 = group2
+		#: Mean of the first group (*float*)
+		self.mu1 = mu1
+		#: Mean of the second group (*float*)
+		self.mu2 = mu2
+		#: Standard deviation of the first group (*float*)
+		self.sd1 = sd1
+		#: Standard deviation of the second group (*float*)
+		self.sd2 = sd2
 		#: Absolute value of the mean difference (:class:`yli.utils.Estimate`)
 		self.delta = delta
 		#: Description of the direction of the effect (*str*)
@@ -53,7 +65,7 @@ class TTestResult:
 		return super().__repr__()
 	
 	def _repr_html_(self):
-		return '<i>t</i>({:.0f}) = {:.2f}; <i>p</i> {}<br>Δ<i>μ</i> ({:g}% CI) = {}, {}'.format(self.dof, self.statistic, fmt_p(self.pvalue, html=True), (1-config.alpha)*100, self.delta.summary(), self.delta_direction)
+		return '<i>t</i>({:.0f}) = {:.2f}; <i>p</i> {}<br><i>μ</i><sub>{}</sub> (SD) = {:.2f} ({:.2f}), <i>μ</i><sub>{}</sub> (SD) = {:.2f} ({:.2f})<br>Δ<i>μ</i> ({:g}% CI) = {}, {}'.format(self.dof, self.statistic, fmt_p(self.pvalue, html=True), self.group1, self.mu1, self.sd1, self.group2, self.mu2, self.sd2, (1-config.alpha)*100, self.delta.summary(), self.delta_direction)
 	
 	def summary(self):
 		"""
@@ -62,7 +74,7 @@ class TTestResult:
 		:rtype: str
 		"""
 		
-		return 't({:.0f}) = {:.2f}; p {}\nΔμ ({:g}% CI) = {}, {}'.format(self.dof, self.statistic, fmt_p(self.pvalue, html=False), (1-config.alpha)*100, self.delta.summary(), self.delta_direction)
+		return 't({:.0f}) = {:.2f}; p {}\nμ({}) (SD) = {:.2f} ({:.2f}), μ({}) (SD) = {:.2f} ({:.2f})\nΔμ ({:g}% CI) = {}, {}'.format(self.dof, self.statistic, fmt_p(self.pvalue, html=False), self.group1, self.mu1, self.sd1, self.group2, self.mu2, self.sd2, (1-config.alpha)*100, self.delta.summary(), self.delta_direction)
 
 def ttest_ind(df, dep, ind, *, nan_policy='warn'):
 	"""
@@ -106,8 +118,8 @@ def ttest_ind(df, dep, ind, *, nan_policy='warn'):
 	
 	# Do t test
 	# Use statsmodels rather than SciPy because this provides the mean difference automatically
-	d1 = sm.stats.DescrStatsW(data1)
-	d2 = sm.stats.DescrStatsW(data2)
+	d1 = sm.stats.DescrStatsW(data1, ddof=1)
+	d2 = sm.stats.DescrStatsW(data2, ddof=1)
 	
 	cm = sm.stats.CompareMeans(d1, d2)
 	statistic, pvalue, dof = cm.ttest_ind()
@@ -115,11 +127,18 @@ def ttest_ind(df, dep, ind, *, nan_policy='warn'):
 	delta = d1.mean - d2.mean
 	ci0, ci1 = cm.tconfint_diff(config.alpha)
 	
-	# t test is symmetric so take absolute values
+	# t test is symmetric so take absolute value
+	if d2.mean > d1.mean:
+		delta, ci0, ci1 = -delta, -ci1, -ci0
+		d1, d2 = d2, d1
+		group1, group2 = group2, group1
+		# Now group1 > group2
+	
 	return TTestResult(
 		statistic=abs(statistic), dof=dof, pvalue=pvalue,
-		delta=abs(Estimate(delta, ci0, ci1)),
-		delta_direction=('{0} > {1}' if d1.mean > d2.mean else '{1} > {0}').format(group1, group2))
+		group1=group1, group2=group2, mu1=d1.mean, mu2=d2.mean, sd1=d1.std, sd2=d2.std,
+		delta=Estimate(delta, ci0, ci1),
+		delta_direction='{} > {}'.format(group1, group2))
 
 # -------------
 # One-way ANOVA