# Power and T1

#### AlexanderC

##### New Member
People say that low power due to low sample size makes it more likely to miss a true effect (i.e. Type 2 error). But low sample size results in higher variability. This means e.g. with a t-test, one of our groups could be larger simply due to this variability, but nevertheless be significant. So doesn't low power sometimes inflate Type 1 error rate as well?

#### hlsmith

##### Less is more. Stay pure. Stay poor.
You have a target population and you sample from it. The sample may have more variability and may be slightly non-representative. Also, most things are based on using a standard error (standard deviation of population estimate). It's formula has sample size in the denominator - thus sample size will impact confidence intervals and formal hypothesis testing, right?

#### AlexanderC

##### New Member
You have a target population and you sample from it. The sample may have more variability and may be slightly non-representative. Also, most things are based on using a standard error (standard deviation of population estimate). It's formula has sample size in the denominator - thus sample size will impact confidence intervals and formal hypothesis testing, right?
Ahh of course! A bigger standard error in denominator requires a larger difference in the numerator for a significant test result, meaning less chance of a Type 1. I see it now, thank you hlsmith (and fed2)