Chapter 8
Module 2For self-study: Running tests
Assignment 8.1Assignment 1
Read through the questionnaire of Ilonka’s study and open the document Ilonka.sav in SPSS. Answer the following question before you go to the first part of assignment 1: What do you think the purpose of her study was?
Calculate the (a) mean, (b) median, (c) mode, (d) range and (e) standard deviation for the variable hours tv (see section 8.6 of chapter 8 for instructions); (f) add a bar chart to visualize the results. (g) Using the definitions of these terms (see section 8.2 of the chapter) describe your findings. (h) As you may have seen, the bar chart is not easy to interpret: what is the problem, and how could you solve it?
Investigate the distribution of participants’ background variables. Focus on the following variables: birthcou, ethnic, language. Calculate (a) mean, (b) median, (c) mode, (d) range and (e) standard deviations – in short, you can use the settings of 1.1. However, this time (f) make for each of these variables a pie diagram based on percentages. (g) Describe the results. (h) What level of measurement do we have here? (see Chapter 5 to fresh up your memory), (i) describe which of the measures of central tendency make sense and which do not, using your answer under (h).
Now in Question 14 Ilonka examined participants’ preferences across a number of television genres. In an attempt to reduce the information, she wants to cluster the genres in larger categories. One could be “entertainment”: fiction, so soaps, TV-series, police/action-series, and cartoons. In this assignment we will see whether participants’ scores on “entertainment: fiction” would be a reliable variable.
Conduct a reliability analysis (see section 8.9 of chapter 8). Describe the results, and interpret them: What can you conclude from the output of this test?
Now do the same for another group of variables: informative programs, news, programs about animals and nature. What can you conclude here?
Examine the relation between gender on the one hand and subjects’ enjoyment of soaps on the other. Describe the results using the section on Boxplots in Chapter 8.
Do the same but this time for gender and the reported enjoyment of sports programs.
Ilonka’s study aimed to find out what differences there are between television watching habits and preferences among children of different cultural backgrounds.
You hopefully did the following.
In the ANALYZE menu you clicked on DESCRIPTIVE STATISTICS and then selected FREQUENCIES. You saw the Dialogue box as printed above, and enter the “Number of hours per day” under Variable(s). You always need to check whether the default settings of SPSS are in fact as you want them to be, in this case, whether they suffice for your assignment. Click on “Statistics” to make sure.
As you will have discovered, you need to check all the boxes yourself, as follows:
This will result in the following output:
| Statistics | |||
|---|---|---|---|
| Number of hours per day | |||
| N | Valid | 262 | |
| Missing | 0 | ||
| Mean | 3,443 | ||
| Median | 3,000 | ||
| Mode | 2,0 | ||
| Std. Deviation | 2,0192 | ||
| Minimum | ,0 | ||
| Maximum | 8,0 | ||
Now we know what the correct answers are to the first part of this assignment: the Mean is 3.44 (two digits seems enough). This tell you what the statistical average is (all the scores added up and divided by the total number of cases. The Median is 3.00. This is the value that is the middle. To the left and to the right of this value we find 50% of the respondents. The Mode is the most frequently occurring value on this variable. The group of participants said they watch 2 hours television per day is the largest apparently. The Standard Deviation, information that should always accompany the Mean, is 2.02 (you can round up the third digit). Look at the minimal score (0) and the maximum (8). Consider the relation between these final bits of information: these tell you that the variability is rather large.
To obtain the Bar chart you probably clicked on Charts in the Frequencies dialogue box, and selected Bar charts there. In that case SPSS produced the following figure for you:
The graph is indeed a bit confusing, as the text of the assignment already announced it would be. As you can see, there are some participants who entered 1 hour, and some that said they watch 1.5 hours. This results in a curve that is hard to interpret. What you could do to solve this is use the recode procedure. For instance, you could add scores 0, 0.5 and 1.0 and recode them into a score 1; 1.5 and 2 in a score 2; 2.5 and 3 into 3, etc. The results will look more transparent, but also consider that you are simplifying the original information
In the last part of the assignment you can follow the exact same procedure. Instead of hourstv you enter birthcou, ethnic, and language in the dialogue box for Frequencies. This produces the following output:
| Statistics | ||||
|---|---|---|---|---|
| Country of birth | Cultural background | Language spoken at home | ||
| N | Valid | 262 | 262 | 262 |
| Missing | 0 | 0 | 0 | |
| Mean | 1,16 | 2,45 | 1,46 | |
| Median | 1,00 | 2,00 | 1,00 | |
| Mode | 1 | 2 | 1 | |
| Std. Deviation | ,659 | 1,112 | ,795 | |
| Minimum | 1 | 1 | 1 | |
| Maximum | 4 | 4 | 4 | |
It is very important that you realize that SPSS will almost always do exactly what you tell it to do. However, there results do not always make any sense. You cannot, for instance, calculate a “mean country of birth” (and this also rules out the Standard Deviation, of course). Also, a Median is hardly informative here. What does the range tell you more than that Ilonka entered one country name as a 1, and another as a 4? A mode does exist on the level of nominal measurement. But in fact, none of the data are worth reporting; even the Mode you can better leave out, and simply report the frequencies themselves. For this, refer to the final part of your output. You were asked to present Pia charts for the results in percentages. You hopefully found this option under Charts.
We do not present the entire output here. Down below you see the Pie chart for Language spoken at home.
As you can see, most of the participants speak Dutch at home (the “mode = 1”in the descriptive statistics we saw). The second language, at least in the sample, is Moroccan. Very few of the respondents speak Surinam at home.