Trying to use statsmodels in Python for a 2 way ANOVA and getting a PatsyError (Tried using pingouin as well, also posted about that issue in another post). Here is my code where param is a variable I define earlier and data_rostral is my data frame:
for param in params:
print(param)
model = ols('param ~ C(ShamBlast) + C(RostralCaudal)+C(ShamBlast):C(RostralCaudal)', data=data_rostral).fit()
anova_table = sm.stats.anova_lm(model, typ=2)
anova_table
Here is the error I am getting:
PatsyError: Number of rows mismatch between data argument and param (23 versus 1)
param ~ C(ShamBlast) + C(RostralCaudal)+C(ShamBlast):C(RostralCaudal)
Tried being less fancy and using totalcells instead of param, which is one of the dependent variables defined by param, thinking I would calculate each ANOVA separately, but then got a ValueError.
60 raise ValueError("wrong shape for coefs")
61 if self.coefs.shape[0] == 0:
---> 62 raise ValueError("must have at least one row in constraint matrix")
63 if self.coefs.shape[0] != self.constants.shape[0]:
64 raise ValueError("shape mismatch between coefs and constants")
ValueError: must have at least one row in constraint matrix
TIA, I know there are other posts similar to this, but I couldn't figure out how those suggestions applied to my code.
