Question

Graphing

Given a histogram with a line for a normal curve superimposed on it, interpretation of the distribution of scores may be ambiguous. If the bars on the histogram are above the line for the normal distribution superimposed on the histogram, then the variable distribution is leptokurtic ("peaked"). If they are below the line, then it is platykurtic ("flat"). However, if the bars cross the line on only one side, then it is skewed. If they cross more on the right side, then it is positively skewed (a few extreme high scores). If they cross on the left side, then it is negatively skewed (by a few extreme low scores). Such a visual evaluation is very imprecise, but more precise statistics for skewness and kurtosis will be presented in a later unit.

Questions:

Why can you not produce a histogram for a variable like religious affiliation?

Also, why can you not produce a line graph, bar chart or pie char for a variable like household income?

Answer 1

If the variable is categorical (e.g sex, religious affiliation,color of a person etc) we can not produce the histogram of the data. e.g Sex categorical variable will take only two kind of observation i,e Male or Female. One can count no of males and no of females and can form the frequency distribution or bar chart or pie chart for such kind of the data and can not form histogram for the said kind of the data because for histogram the data should be continuous or interval (e.g age, height, weight etc). The Variable like (age, height, or weight) can have any value from 0 to infinity.

Bar chart or Pie chart are obtained for categorical data while as line graph is obtained to find trend in the data and is used for continuous data. The variable like household income is an interval data one can not obtain obtain the said graphs for household income.