### Instructions for Choosing between Gambles Based on Advice

In this experiment you are asked to choose between gambles. Each gamble
consists of a glass container (a jar) containing red and white marbles. There
are always 100 marbles, but the percentage of red and white will differ
from gamble to gamble. One marble will be chosen randomly and blindly
from the chosen urn, and if the marble is red, you win $100, but if
the marble is white you get $0 (nothing). It is in your best interest
to choose the urn with the greater percentage of red marbles.
For urn A, you will always be given the exact percentage of red marbles.
For urn B, you will receive advice from someone who wants to help you
who has estimated the percentage of red marbles in urn B. There are two
sources, a friend and an expert:

The *friend* has no training in estimating percentages, but is
trying to help you by giving you an estimate of the percentage of Red in Urn B.

The *expert* has been trained in estimating percentages by eye.
The expert also makes an estimate of Urn B by eye.

These people are both trying to help you, but they do not know the exact percentages;
they make their estimates by shaking the jar, looking through the clear glass,
and estimating the percentage by eye.

Look at the first trial below, W1. In this case, we know that urn A has exactly
30 red marbles and 70 white ones. The friend, however, thinks that there are 70% red
in urn B. If you think there are more than 30% red in the urn B, you should choose urn B.
By choosing A, you know that you have exactly 30 out of 100 chances to win. But if
the friend is right, there are more than twice as many red marbles in urn B. Some people
will choose A and others will choose B.

If you strongly prefer A, click
far to the left, and if you strongly prefer B, click far to the right.
To indicate intermediate degrees of preference, click along the row of buttons closer to the
urn from which you would prefer to choose a marble.

#### Warmup Trials Choosing between Gambles Based on Advice