One would expect there to be a positive correlation (r = 0.370) between "My sub (sandwich) was delicious" and "I would recommend Subsay to my friends'. This suggests that the more likely a respondent found the sandwich delicious, the more likely they are to recommend Subsay.

Run bivariate correlations with all the interval scale items. 

Looking at the 'r' values between all items and 'I would recommend Subsay to my friends'. If you had to take just ONE variable, which one you say made the least sense.

A correlation coefficient of -0.502 between the 'reason the respondent was in town' and 'needing a snack' suggests that there is an inverse association between reasons for being in town and needing a snack.

Running a correlation between the reason why a respondent "was in town' and 'Gender', would tell me about the strength of the association of gender and the reason for being in town. For example, Is one gender more likely to be in commuting than another.

Are there any subjects where there is NO significant difference between the mean scores of each of the genders? If so indicate which subject/s.

Which subjects show a significant difference between the standard deviations?

What is the t-value that we would be using to determine if there was a significant difference between the genders for Care of Magical Creatures?

Draft a bar chart of mean scores for the Hogwarts subjects. Include in this chart reference line for the 'test value' of 55%.

Which subjects are NOT showing a significant difference from the hypothesised 55%?