Paired T test vs 2 sample t test I have been given data that includes 20 different households and what each household spends on grocery and restaurant.  I need to run a t test and I'm not sure if I should use a paired t test or a 2 sample t test.  Also what is the best way to graphically represent this data?
 A: If it's the same families you should use a paired t-test. 
You could represent the data via a sort of spaghetti plot (but it would be dry spaghetti, each line would be straight).  Make two columns, one for grocery and one for restaurants. Each family gets a dot on each line and you connect the dots in pairs. 
A: From the Cliff Notes site, we look at the Paired Difference t-test page:

This t‐test compares one set of measurements with a second set from
  the same sample. It is often used to compare “before” and “after”
  scores in experiments to determine whether significant change has
  occurred.

In your case you have two measurements on each household.  So spending from each household at the grocery store would be the first measurement and spending from each househould at the restaurant would be the second measurement.  We then wish to compare these two measurements using the paired t-test.
To best represent your data, the x-axis should represent restaurant or grocery store and the y-axis should represent spending.  You should then connect the spending in restaurant to the spending in grocery store based on household.  Here is an example, substitue in Restaurant for Before and Grocery for After.

