Performing a t-test in Python

To conduct a t-test in Python, all you have to do is specify the alternative hypothesis and indicate whether variances are roughly equal (homogeneous).

The ttest_ind() function within scipy.stats handles the rest. Below is the syntax:

Parameters:

a — the first sample;
b — the second sample;
equal_var — set to True if variances are approximately equal, and False if they are not;
alternative — the type of alternative hypothesis:
- 'two-sided' — indicates that the means are not equal;
- 'less' — implies that the first mean is less than the second;
- 'greater' — implies that the first mean is greater than the second.

Return values:

statistic — the value of the t statistic;
pvalue — the p-value.

We are interested in the pvalue. If it is lower than α(usually 0.05), then the t statistic is in the critical region, so we should accept the alternative hypothesis. And if pvalue is greater than α — we accept the null hypothesis that means are equal.

Here is an example of applying the t-test to our heights dataset:


              123456789101112131415
            
import pandas as pd
import scipy.stats as st

# Load the data
male = pd.read_csv('https://codefinity-content-media.s3.eu-west-1.amazonaws.com/a849660e-ddfa-4033-80a6-94a1b7772e23/Testing2.0/male.csv').squeeze()
female = pd.read_csv('https://codefinity-content-media.s3.eu-west-1.amazonaws.com/a849660e-ddfa-4033-80a6-94a1b7772e23/Testing2.0/female.csv').squeeze()

# Apply t-test
t_stat, pvalue = st.ttest_ind(male, female, equal_var=True, alternative="greater")

if pvalue > 0.05:
# Check if we should support or not the null hypothesis if pvalue > 0.05:
    print("We support the null hypothesis, the mean values are equal")
else:
    print("We reject the null hypothesis, males are taller")

Tudo estava claro?

Seção 6. Capítulo 6