How to hypothesis test in Python

stats.py

import scipy.stats as stats
import numpy as np
import scipy.stats.distributions as dist
import statsmodels.api as sm


# calc by hand
va = prop_sub * (1 - prop_sub) # variance

se = np.sqrt(va * (1 / total_high + 1/total_mod)) # grouped standard error

# get t value and p value
test_stat = (phat1 - phat2) / se
p_value = 2 * dist.norm.cdf(-np.abs(test_stat))

print(test_stat, p_value)

# Using SciPy and only stats (no records)
# https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.ttest_ind_from_stats.html

va1 = phat1 * (1 - phat1)
va2 = phat2 * (1 - phat2)
print(va1, va2)

stats.ttest_ind_from_stats(
  mean1 = phat1,
  std1 = np.sqrt(va1),
  nobs1 = n1,
  mean2 = phat2,
  std2 = np.sqrt(va2),
  nobs2 = n2,
)