Paychological Scales And Tests Flashcards
Types of variables
Categorical: not numerical
Quantitative: numerical
Steven’s measurement theory
Hierarchy of measurements
1. Nominal
2. Ordinal
3. Interval
4. Ratio
3 properties
1. Relative order (increase or decrease)
2. Equal interval (same unit step)
3. Zero point (absence of measurement)
Nominal variables
Measurement aim: categorizing cases in terms of more than 2 categories
Multinominal variables: 3 or more
Labeled like in SPSS
Ordinal variables
Measurement aim: giving value to each category
- dichotomous quantitative variables (eg level of income)
- participants arranged in order
-don’t follow one another in equal steps
Interval variable
Measurement aim: asseigning real value/numbers to each value in the same unit step
Eg: temperature
- no zero point (0°C is not the lowest point)
Ratio variable
Measurement aim: all measurement properties apply:
- can be arranged in order
- constant unit step difference
- contains zero point
Overview for variables
- Nominal: no relative order, no equal interval, no zero point
- ordinal; relative order, no equal interval, no zero point
- interval: relative order, equal interval, no zero point
- ratio: relative order, equal interval, zero point
Psychological measures
Difficulty of finding psych examples for interval and ratio as as interval and ratio
Quasi interval scale
Appears to be interval, but its scale isn’t necessarily equal
Psychological tests and scales
Attempt to quantify psych characteristics
- psychometry: technology of test creation for quantification
- standardization
Restrictions in availability
- costs
- training or qualification needed for test
Unidimensional scales
Correlation of scale items with each other are determined by the result of a single underlying dimension
Multidimensional scale
Correlation based on result of two or more underlying dimensions
Open questions
Participants can answer freely -> qualitative data
Closed question
Predefined questions can be answered-> quantitative data
Can be frustrating to answer when the predefined answer isn’t correct
Can be misinterpreted
Thurstone scale
Central idea: each statement of scale, if a person agrees, it’s given a score to strengthen that statement
Process of scaling
- Specifying specific psych construct
- Producing questions
- Asking judges to determine how the questions are (not their attitude)
- Taking judges mean
- Constructing final scale
- Scale administered for responders
Limitations of thurstone scale
- judges can never be neutral
- LARGE number of judges
Linkert-type scale
- most frequently applied scaling
- set of statements with fixed responses (never, slightly, idk, sometimes, always)
- quantification through assigning numeric value
Issues with likert-type scale
- forcing responders to predertimed answers
- middle response option
- interpretation of score
Bipolar scale
YES OR NO
Semantic differential scale
- measuring attitudes
- two extremes and then you have to give your attitude
Visual analoge scale
Mark positions in a line between two bipolar statements
- absence of categories
Problem with bipolar scales
- response bias
- problem of interpretation of middle score of the scale
General issues and limits of testing
- response set or response bias
- loss of attention
- need mix of negative and positive questions
- responders interpretation
- social desirability-> going for a nicer answer to look good