diff --git a/sbi/analysis/sbc.py b/sbi/analysis/sbc.py
index 14a38daba..4264e06ca 100644
--- a/sbi/analysis/sbc.py
+++ b/sbi/analysis/sbc.py
@@ -169,7 +169,9 @@ def sbc_on_batch(
     dap_samples = torch.zeros_like(thetas)
     ranks = torch.zeros((thetas.shape[0], len(reduce_fns)))
 
-    for idx, (tho, xo) in enumerate(zip(thetas, xs)):
+    for idx in range(thetas.shape[0]):
+        # unsqueeze for potential higher-dimensional data.
+        xo = xs[idx].unsqueeze(0)
         # VI posterior needs to be trained on the current xo.
         if isinstance(posterior, VIPosterior):
             posterior.set_default_x(xo)
@@ -184,7 +186,9 @@ def sbc_on_batch(
         # rank for each posterior dimension as in Talts et al. section 4.1.
         for i, reduce_fn in enumerate(reduce_fns):
             ranks[idx, i] = (
-                (reduce_fn(ths, xo) < reduce_fn(tho.unsqueeze(0), xo)).sum().item()
+                (reduce_fn(ths, xo) < reduce_fn(thetas[idx].unsqueeze(0), xo))
+                .sum()
+                .item()
             )
 
     return ranks, dap_samples