Psy 271 Research And Design Methods

Front 1

What is measurement?

Back 1

Measurement is the use of numbers to represent the quantity of abstract concepts and variables.

Front 2

Why is measurement a critical part of social science?

Back 2

-Because it is through measurement that we’re able to make direct, systematic observations of abstract variables.

Front 3

Can you have a reliable measure that is not valid?

Back 3

Yes, exp. Bathroom scale that’s calibrated wrong, keep giving wrong measurement

Front 4

Can you have a valid measure that is not reliable?

Back 4

No there not completely independent concepts

Front 5

What is meant by the reliability of a measure? How is the concept of reliability related to the concept of “error”?

Back 5

The reliability of a measure refers to: the extent to which there is error in our measurement of a variable to the extent that scores reflect error (influenced by things other than the true score), the measure is unreliable. More error = less reliable, less error = more reliable.

Front 6

Describe test-retest assessments of reliability.

Back 6

Examination of the consistency of the measure over time, when the amount of variable present (the true score) has NOT changed. Example – measuring height

Front 7

What is meant by the sensitivity of a measure? Is a 5-point Likert scale more or less sensitive than a dichotomous response question? Explain.

Back 7

Sensitivity addresses the question "does our measure reflect small differences between subjects in the actual amount of the variable that they experience?"
You can provide more response options to increase a sensitivity of a measure. So a 5-point likert scale would be more sensitive given the most options (1-5) This provides finer distinctions between people in the amount if variable present.

Front 8

What is meant by the validity of a measure? Discuss three ways of assessing the
validity of a measure.

Back 8

Validity is the question of whether the measure is measuring the variable we think it’s measuring
Face Validity: simply refers to the plausibility of the measure. Does it make intuitive sense? Does it look like a measure of variable X? For example, a set of arithmetic problems has high face validity as a measure of math ability.
Convergent validity: you collect very different measures (different kinds of measures) of a variable and then you determine if the two measures are correlated.
Construct Validity. To establish construct validity of a measure, you determine if your measure of the variable of interest (i.e., prejudice) is correlated to a measure of another variable that is theoretically (conceptually) related (i.e., education)

Front 9

Know the characteristics of each level of measurement and be able to recognize examples of each.

Back 9

Nominal –Classifies variables into categories (identity)
Ordinal— Classifies variables into ranked categories (properties of identity and magnitude.) (runners 1st, 2nd ...) , students rank in class)
Interval— Classifies variables into ranked categories with standard equal distance between adjacent points on scale. (Have the property of identity, magnitude and equal intervals.)There is equal spacing between response categories on this measure. Exp. Temperature scale
Limitation is there is no true zero point (natural origin of scale is the complete absents of the variable)

Ratio-- Classifies variables into ranked categories of equal distance with a true 0 point where variables cease to exist

Front 10

Advantages of rating scales over dichotomous response scales.

Back 10

Because rating scales provide more response options, they are more sensitive measures so they can make finer distinctions between people in the amount of variable present. Dichotomous response questions are strictly ordinal level measures. The rating scale has more response options so it can be treated as an interval level measure. This way we can extract more information from our measure using statistics. The disadvantage of a rating scale is that because there are more response options, they can be more difficult questions for the respondent.

Front 11

describe one way you would assess the reliability of a measure?

Back 11

You can infer the reliability of a measure. One way to do that is to examine the consistency or stability of the measure through a "test-retest method.”

Front 12

The properties of nominal level measurement

Back 12

identity property

Front 13

Interval level measurement

Back 13

identity magnitude and standerd equal distance between adjacent points on a scale (equal intervals)
Exp: Temperature or many psychological test

Front 14

You can increase the sensitivity of a measure by adding response categories.
How does increasing the number of response categories affect the reliability of a measure? Explain

Back 14

By making the measure more sensitive you are able to reflect small differences between subjects in the amount of variables experienced.

Front 15

What is meant by the validity of a measure?

Back 15

the question of whether the measure is measuring the variable we think it's measuring

Front 16

Describe (do not simply list) three ways of assessing the validity of a measure..

Back 16

There are three ways you can assess the validity of a measure. First you can use Face validity, which simply refers to the plausibility of a measure by asking if it makes intuitive sense or if it looks like variable X.
Then you can either use construct or convergent validity.
To establish Convergent Validity you must do two things, first you collect very different measures of a variable. Then you determine if the two measures are correlated.
To establish construct validity you must determine if your measure of the variable of interest is correlated to a measure of another variable that is theoretically related.

Front 17

describe one way you would assess the reliability of a measure?

Back 17

to examine the consistency or stability of the measure through a "test-retest method.”

Front 18

What is meant by the reliability of a measure?

Back 18

The reliability of a measure refers to: the extent to which there is error in our measurement of a variable

Front 19

How is the concept of reliability related to the concept of “error”?

Back 19

To the extent that scores reflect error (influenced by things other than the true score), the measure is unreliable. More error = less reliable, less error = more reliable.

Front 20

Likert, Likert-type, and semantic differential and all measures of underlying psychological variables are..

Back 20

ordinal level because there is no constant unit of measurement (such as inches or seconds)

Front 21

What are limitations of interval level measures?

Back 21

There is no true zero point

Front 22

A true zero point represents..

Back 22

The complete absents of the variable

Front 23

Semantic differential---

Back 23

Scales are constructed such that the end points of a scale are anchored by adjectives that are bi-polar opposites. Such as...
Not at all--1 2 3 4 5 6 7--Very Much

Front 24

Likert Scale--

Back 24

They consist of pro or con attitude statements and they consist of 5 or 7 response options that gives the scale a mid-point representing a neutral attitude.

Front 25

How would you increase the reliability of your measure?

Back 25

take MULTIPLE MEASURES OF THE SAME UNDERLYING VARIABLE. – More than 1 question assessing attitudes toward WCU. This works because error is random. Take the average of the responses of multiple questions on a self-report measure.