Semester 2- Module 2: questionnaires and scales Flashcards
What are the types of measures?
behavioural measurement, self-report measures, physiological measures.
Explain behavioural measures.
the measure of observable behaviours, such as where a participant sits. Data can be difficult to obtain and interpretation is not always clear.
Explain self-report measures.
participants report their own feelings / behaviours / attitudes towards something. Easy to gather but relation to behaviour not always clear.
Explain physiological measures.
Measure of physiological changes. These changes are associated with psychological processes. examples include brain scans, and heart rate. Can be expensive to obtain and issues with interpretation may exist.
Define Questionnaire.
Different name for survey. usually consists of measurements for multiple constructs.
Define scale.
A set of items designed to measure a single psychological construct.
Define item.
an individual question.
What is psychometrics?
The subfield of psychology concerned with measurement.
What are latent variables?
Constructs which can only be observed indirectly, such as self-esteem.
Explain the classical test theory.
Item score = true score + error. (X = T + E). The true score reflects the latent vairable, error is assumed to be random.
Where does error come from?
Poorly worded questions, fluctuating state of person being tested, test items that are narrow, difference in skill/knowledge of testers or observers.
Why do most scales have multiple items?
Because you are more likely to measure the true score. As items are included the influence of errors (E) becomes smaller, E approaches 0 because it is random, and what will be left is the true score (T).
What is Cronbach’s α?
A measure of internal reliability, it indexes (describes) how well multiple items from the same scale correlate among each other.
What do Cronbach’s α scores mean?
α > .60 acceptable
α > .70 good
α > .80 excellent
however, Cronbach’s α can be too high. a score of .99 means items are redundant. You can measure the construct with fewer items without losing information.
what are the three biases in data that can arise from a single person answering all the items, and most psychological measurement being based on self-report?
Acquiescence bias, Consistency bias, and social desirability bias.
