Interpreting and Using Statistics in Psychological Research. Andrew N. Christopher
Читать онлайн книгу.
or a scale variable. Go ahead and see whether you can identify the scale of measurement for each of our 10 variables. You can check yourself with this screenshot (or the answers to the previous Learning Check).
11 Role. We should keep the default of Input here. The other choices are beyond the scope of what we need to do in this class.
Data View
Now that we have set up our Variable View, we click on Data View (bottom left corner of the screen) to begin entering our data. Notice that across the top of the spreadsheet, each column is one of our variables. Each row will contain the data for one participant.
Figure 2.4 contains the (hypothetical) data from two people on the 10 variables we have established in our SPSS spreadsheet. Now that we are in Data View, let’s enter those data into SPSS and get some practice in this important detail.
As you begin to enter your data, it is critically important that you recall the numerical values you gave for your nominal variables. Remember, for instance, that we will give male respondents a 1, female respondents a 2, and transgendered respondents a 3 when entering data for the variable of Sex into SPSS. If you ever have any doubts about the values you provided in the Variable View, go back and look. Incorrectly entering data into SPSS renders worthless all the statistical tools we will learn in this class. SPSS is only as good as the data we enter into it.
Figure 2.4 Sample Data to Enter Into SPSS
Let’s start entering participant 1’s data into SPSS. This person reported being a female who is a sophomore and is Asian. As I just preached, we must code these nominal data correctly into SPSS or else we might as well just enter garbage into the spreadsheet. For the variable of Sex, we give participant 1 a 2. For the variable of Class_Standing, we give this participant a 2. For the variable of Ethnicity, we give her a 5.
The good news now is that scale variables tend to be easier to enter into SPSS than are nominal variables. As mentioned, we can normally just enter the value that the participant provided. So for the variable of GPA, enter 3.45. The variable of Height is a little trickier only because we cannot enter“5 feet, 5 inches” into the spreadsheet. Rather, we need to convert this height into inches (or whatever scale you want to use so that we have a single number). “5 feet, 5 inches” converts to 65 inches, so I entered 65 for this variable. Go ahead and enter the remaining data for the first participant. Then, enter the data for participant 2 (this is not an exercise in converting measurements of feet into inches, participant 2 is 74 inches tall). Here is what your Data View will look like:
The main menu runs across the top of the Data View, starting with File on the left across to Help on the right. At this point in the class, we do not have any statistical tools in our toolkit to use with these data, but know that for our purposes, we will be primarily using the Analyze tab and the Graphs tab in this class.3
Notes
1. All research involving human participants must be approved by what’s called an institutional review board (IRB) or other committee that reviews the ethics of research studies. In this example, a research study in which participants would be allowed to hit each other probably would not be approved by such a committee.
2. There are of course many ways to divide a 12-item scale into halves (e.g., even- and odd-numbered items or first six items and last six items). Typically, researchers use computer software, such as the program we will begin learning later in this chapter, to assess internal reliability. We’ll discuss more about internal reliability in Chapter 12.
3. In case you are interested in Terrell et al.’s (2008) results, they found that men who were more narcissistic gave more noise blasts (that is, acted more aggressively) than men who were less narcissistic. Women did not differ in their aggressive behavior based on their level of narcissism. In addition, women with a tendency to act aggressively were more likely to display aggression against a low-status competitor, whereas males with a tendency to act aggressively were more likely to display aggression against a high-status competitor.
Chapter Application Questions
1 How does a measurement’s reliability differ from validity?
2 What is the difference between construct validity and criterion validity?
3 What information does scale data (interval and ratio data) provide in addition to what an ordinal scale provides?
4 For each variable, identify the scale of measurement (nominal, ordinal, interval, or ratio) and whether it is discrete or continuous:
5 When your “intelligence” is defined as your score on the Wechsler Adult Intelligence Scale, you are:testing the validity of this measure.testing the reliability of this measure.operationalizing the term intelligence.determining the correlation between two variables.
6 Which of the following BEST represents a nominal scale of measurement?scores on a personality measurethe number of inches of rain in Detroit on September 15, 2015letter grades in schoola person’s membership in one particular Greek organization
7 Which of the following is NOT a continuous variable?number of siblingstemperatureheighttime needed to solve a problem
Answers
1. Reliability refers to a measurement’s consistent scores, whereas validity refers to how appropriate a measure is to use in a given context.
2. Construct validity concerns how well a variable is operationalized. Criterion validity concerns how well a measurement tool predicts some outcome.
3. Scale data are meaningful numeric data; with nominal and ordinal data, the numbers are identifiers. With scale data, the differences between numbers are meaningful and allow us to use a wide range of statistical tools to understand them.
4.
5. c
6. d
7. a
Questions for Class Discussion
1 What is the difference between test–retest reliability and internal reliability?
2 What information does an ordinal scale of measurement provide in addition to the information that a nominal scale provides?
3 Why would a researcher likely NOT record a participant’s religious affiliation on a scale (i.e., interval or ratio) measurement?
4 A study was conducted to determine whether physically fit persons sleep more hours than those who are not physically fit. Two groups of people were selected. One group consisted of people who work out at least three times a week. The other group consisted of people that do not work out at all. For one week, participants slept in a sleep lab and an experimenter recorded the number of hours each person slept.Is there an independent variable or a quasi-independent variable being examined in this research? What is this independent or quasi-independent variable?What is the dependent variable?Is the dependent variable discrete or continuous?What scale of measurement is the dependent variable measured on (nominal, ordinal, or scale)?
5 Graduate