Studnet t-distribution is more accurate small samples

infocalc.py

@@ -11,7 +11,8 @@ from pathlib import PurePath##https://docs.python.org/3/library/pathlib.html#mod
 import numpy as np
 import matplotlib
 import matplotlib.pyplot as plt
-from scipy.stats import norm
+from scipy.stats import norm,t
+import scipy.stats
 import pandas as pd
 
 #Main program loop
@@ -101,17 +102,18 @@ elif args.mode == "SIM":
     mean = args.mean
     stddev = args.stddev
     samplesize = args.samplesize
+df = samplesize - 1
     
 # time to deal with the bounds
     
 # Delta Degrees of Freedom: ddof=0 for population, ddof=1 for sample std dev
 prob = 0
 if args.upperbound and args.lowerbound:
-    prob = norm.cdf(args.upperbound, mean, stddev) - norm.cdf(args.lowerbound, mean, stddev)
+    prob = t.cdf(df,args.upperbound, mean, stddev) - t.cdf(df,args.lowerbound, mean, stddev)
 elif args.upperbound:
-    prob = norm.cdf(args.upperbound, mean, stddev)
+    prob = t.cdf(df,args.upperbound, mean, stddev)
 elif args.lowerbound:
-    prob = 1 - norm.cdf(args.lowerbound, mean, stddev)
+    prob = 1 - t.cdf(df,args.lowerbound, mean, stddev)
 else:
     prob = 1# no bounds set!
     ##TODO!!!!