Читать книгу Interpreting and Using Statistics in Psychological Research - Andrew N. Christopher - Страница 73

1 We use bar graphs to display frequency distributions of nominal data, whereas we use histograms to display frequency distributions of ordinal and scale data.

2 Recall that nominal data are categorical data. That is, a datapoint is in one and only one category. Bar graphs, as we said in the previous question, are meant to represent frequencies for nominal data. Because such data are discrete (i.e., can fall into only one category), they are separate from all other categories. Therefore, the bars do not touch each other.For question 4, consider the following two histograms on the next page that contain the distribution of scores on a class exam.

3 The histogram in Figure 3.18a shows that many students tended to score toward the high end of the distribution on the test, whereas the histogram in Figure 3.18b shows more students toward the low end of the distribution on the test. So before knowing your own score, the top distribution would make you feel better about your chances of scoring well on the test.For question 5, consider the line graph in Figure 3.19.

4 This line graph shows that life satisfaction tends to decline in people’s late teenage years, continuing through their 20s and 30s, but then it starts to increase once people reach their 40s and increases as people continue to age from that point in their lives.

5 This is a nonlinear relationship because the relationship between age and life satisfaction is different across different age levels. That is, it tends to decline in our teenage years until our 40s (a negative linear relationship), but then it tends to increase starting in our 40s as we age from that point (a positive linear relationship).

6 A bar graph displays categories. Scale data do not contain categories, so there is no practical way to make a bar graph with scale data.

7 Nominal data contain categories. But scatterplots are useful only when the two variables being plotted are both scale data. There is no practical way to make a scatterplot with nominal data because categories cannot be plotted on a scatterplot.

Figure 3.18a

Figure 3.18b

Figure 3.19