 |
1 | |
Reliability of Measures
When psychologists measure behavior, the measurement strategy or device is considered if it does not fluctuate from one reading to another. |
|
|
2 | |
For example, a reliable measurement of extroversion, a personality characteristic that is relatively consistent over time, would produce the score each time it is administered to the same person. |
|
|
3 | |
To understand reliability, one must first consider that there are components in any measure. |
|
|
4 | |
The first component is the , which is the real score on the variable. |
|
|
5 | |
The second component is or the variability in the scores each time the same person is measured. |
|
|
6 | |
Reliability is assessed using correlation . |
|
|
7 | |
The most common correlation coefficient used for assessing reliability is called the correlation coefficient. |
|
|
8 | |
This coefficient ranges from 0.00 to and 0.00 to -1.00. |
|
|
9 | |
The closer a correlation coefficient is to (either positive or negative), indicates a stronger relationship. |
|
|
10 | |
In addition to the value of the coefficient, the sign (positive or negative) indicates the type of . |
|
|
11 | |
If the sign is positive (+), then the relationship is , which indicates that scores on both of the variables covary in the same direction. |
|
|
12 | |
That is, as scores on one variable increase the scores on the other variable also or the scores on both variables decrease. |
|
|
13 | |
When the sign is negative (-), then the relationship is and the scores on one variable increases as the scores on the other variable decreases. In other words, the two variables covary in opposition directions. |
|
|
14 | |
Reliability of the measure is assessed by obtaining two scores on the measure from many individuals. When the measure is reliable, the two scores should be and the Pearson correlation coefficient should be a high positive relationship. |
|
|
15 | |
One method used to assess reliability is called the method and the same behavior of the same individual is measured at two points in time. |
|
|
16 | |
In addition to this method, researchers have devised methods to assess reliability without separate assessments by using a measurement called consistency reliability. |
|
|
17 | |
Internal reliability can be assessed by correlating an individual's total score on one half of the test with their total score on the other half of the test, called reliability. |
|
|
18 | |
Internal consistency can also be assessed by calculating the correlation of each item score with every other item score on the test, called alpha or by correlating each item score with the total score based on all items known as the reliability. |
|
|
19 | |
Sometimes behavior is not measured by a test measurement but rather by raters making ratings or judgments of observed behaviors. In this case, a observation of one rater might be unreliable. |
|
|
20 | |
The solution is to use at least raters to observe the same behavior and then computing the reliability. |
|
|
21 | |
Construct Validity of Measures
Besides the reliability of measures, the validity of measures is also important. For example, refers to the adequacy of the operational definition of variables. Does the measure actually measure the construct itself? |
|
|
22 | |
To know whether a measure is valid, there are several indicators of construct validity. One indicated is called , which involves the judgment of whether or not the measure appears "on the face of it" to actually measure the variable. |
|
|
23 | |
However, although, the measure may look like it is measuring the construct, there is no evidence to support this conclusion. This is why, researchers conduct research to assess the construct validity of the measure. |
|
|
24 | |
Construct validity research examines the relationship between scores on the measure and some criterion, called validity. |
|
|
25 | |
A researcher makes a , based on the theory underlying the construct, of how the measure will relate to another variable called the criterion. This can be done using four types of criterion-related research approaches. Each of these four approaches differs in the type of criterion that is used. |
|
|
26 | |
They are called , concurrent validity, convergent validity, and discriminant validity. |
|
|
27 | |
Reactivity of Measures
Besides measurement reliability and validity issues, there is a potential problem when measuring behavior called . |
|
|
28 | |
A measure is if participants are aware their behavior is being measured and their changes. In other words, the participant's behavior is not natural under the circumstances because they know their being observed. |
|
|
29 | |
However, there are ways to minimize reactivity, such as allowing time for individuals to become used to the presence of the observer, using equipment or measures. |
|
|
30 | |
Variables and Measurement Scales
As mentioned earlier, every variable must be operationally defined. The values or levels of the variable may be or categorically different depending on the measurement scale used. |
|
|
31 | |
There are four kinds of measurement scales are , ordinal, interval, and ratio. |
|