Answer:
B. Phone company B because the sample was not biased.
Explanation:
While Phone company A had a higher satisfaction rate, 25 people that conducted this poll were in at least 1 commerical of the phone company's in the past 3 years. This means that their reflection of the company is biased (typically higher) and not completely based off of their performance but instead how cool it was to be in a commerical or how much they were paid. If the poll was a more random sample of just customers, it would be more believable data. On the other hand, while phone company B had a lower satisfaction rate and higher deviation, their statistics are more believable due to the random sampling of customers.