# Two-sample t confidence interval

#### mackerbuddy

##### New Member
Am I correct??!!

Corn yields (bushels per acre) for experimental plots controlled to have 1 weed per meter of row and 3 weeds per meter of row:

1 weed/meter - 166.2, 157.3, 166.7, 161.1
3 weed/meter - 158.6, 176.4, 153.1, 156.0

Explain carefully why a two-sample t confidence interval for the difference in mean yields may not be accurate.

The two-sample t confidence interval for the difference in mean yields may not be accurate because the sample size is less than 15 and it doesn't appear to have a Normal distribution.

Is that right?

#### JohnM

##### TS Contributor
If I'm reading this correctly (and I may be wrong here), the underlying distribution of average weeds per meter may not follow a t-distribution (or normal distribution either)....

hint: what distribution is typically used to model counts of occurrences....

another similar example would be a car manufacturing facility wants to estimate the number of scratches per door panel on a new paint finishing process....