Exam #2 Flashcards
Standardization
2
Self Deception
A form of socially desirable responding. An unconscious distortion. The person is distorting results, but not aware of it.
Impression management
A form of Socially Desirable Responding. Conscious distortion of results. They are aware that they are being deceptive.
What are the 5 sources of error in an interview?
1) Successful interview depends on the interviewee providing honest information- Sometimes clients are not immediately forthcoming about the various facets of their life
2) lack of insight- clients do not always have insight and they paint an inaccurate picture of themselves- this is why it is sometimes useful to interview a significant other
3) Interviewer error: bias. We all have biases and it is important to come to an understanding of these biases
4) interviewer error: judgement- don’t rush to judgement based on a single piece of information. ie client hearing voices does not necessarily mean schizophrenia
5) Interview error: interpretation and inferences. - if report writing it is best to stick to behavioural data(You believe X because client said or did X)- Speculating about the meaning of behavioural data is dangerous (both to the client and you) so be conservative about your statements about what you report about a client
Social Facilitation
People impact eachothers moods and behaviours. People tend to act like the models around them. Therefore the therapist must model relaxed confidence, in order to create conditions of openness and warmth.The client will be impacted by the therapist’s behaviours and moods.
False reassurance
Attempts to comfort or support but the support is false. Ie Someone was just diagnosed with HIV+. telling them “Don’t worry, everything is going to be all right” is a false reassurance.
What are the 5 basic steps in test construction
`1) Test Conceptualization
2) Item generation
3) Item Tryout
4) Item Analysis
5) Test Revision
1) Test Conceptualization (4 parts)
Test construction
1) Specify the objective of the test
What is the goal of testing?
Ie Create a measure of personality
The purpose of the test ought to be guided by a theory that links the test and the construct
2) Define the attribute
- What is the construct that you are measuring?
Operational definition- provide an objective and measurable definition of the construct
3) Identify the population
- Which is the group on which you will develop the test?
4) Review of the literature
- What is the previous research on the construct?
Step 2) Item generation
Test construction
Generate a pool of items to measure the trait by systematically identifying and specifying the items in the content domain
-Items should relate to the central theory and be expressed in clear and precise language
Rules for Item Writing
test construction
1) Define what you want to measure
2) Generate item pool
3) Use short items
4) Use appropriate reading level
5) Avoid double-barreled items
6) Reverse code items
Response Biases
1) Social Desirability
- Self Deception- An unconscious distortion. The person is distorting results, but not aware of it.
- Impression management- Conscious distortion of results. They are aware that they are being deceptive.
Acquiescence- Yay Sayers, agree with everything.
Random Responding*
Extreme Responding
Negative Impression*
Step 3) Item Tryout
Test construction
Develop a method of data collection, ie a Questionnaire
-Administer the test items to a sample that is identical to population for which the test will be developed
Step 4) Item analyses
Quantitative Evaluation
- The purpose is to obtain more information on each item in order to determine the retention, deletion, or revision of items.
Step 5) Test Revision
Assessing strengths and weaknesses of items.
Modifications on the basis of the analyses.
What items should be retained or revised?
Review purpose of the test to determine any modifications.
Continue the cycle of tryout, analyses, and revisions
Reliable
Consistent measurement
Valid
Measuring what it is supposed to measure
A good test must be what 2 things
Reliable and Valid
4 Forms of Reliability
1) Test Retest-person takes test then takes exact same test again after a period of time.
2) Alternative Forms- You have two different versions of a test. Tests have different items but they are tapping the same construct
3) Internal Consistency-degree to which items hang together. Different forms of it – chronbach’s alpha (very popular in psychology). Also reported with KR-20 and split half. You might have a test that doesn’t have good test retest but it might be internally consistent. Internal consistency is very important for validity.
4) Inter Rater- The 3 that we just have discussed make use of the correlation coefficient. The interrater is known as a Kappa statistic. E.g., you have 3 psychologists rating a client, to what degree do they decide the same diagnosis. Looking at agreement across judges.
Test Retest, Alternative Forms, and Internal Consistency report on the correlation coefficient. Inter-rater reports the Interclass coefficient) ICC
5 Forms of Validity
1) Face validity:
2) Content
3) Criterion
4) Construct
5) Factorial
1) Face validity – non-statistical and non-expert approach to assessing validity. You make a surface judgment about a test. Say if you were given a social psychology exam instead of a psychometrics one that you were expecting, you would give it low face validity.
2) Content validity – non-statistical approach established by experts. Experts determine if the test is measuring the domains its supposed to measure. How can you establish content related validity from your last test? Ask another psychometrics prof to look at your last exam.
3) Criterion Validity – Looking at association of bivariate – behavior (criterion – will be placed on the y axis) and test (predictor – placed on the x axis). Correlating your test with actual behavior. Two different forms.
4) Construct Validity – correlating your test with a well established gold standard test.
5) Factorial Validity – Trying to test validity of some sort of theoretical structure. ie 3 factor structure (psychoticism, neuroticism, and extroversion). So you administer a test to lots of people, and see if in fact the outcomes boil down to the 3 theoretical factors.
5 Forms of Validity
1) Face validity:
2) Content
3) Criterion
4) Construct
5) Factorial
1) Face validity – non-statistical and non-expert approach to assessing validity. You make a surface judgment about a test. Say if you were given a social psychology exam instead of a psychometrics one that you were expecting, you would give it low face validity.
2) Content validity – non-statistical approach established by experts. Experts determine if the test is measuring the domains its supposed to measure. How can you establish content related validity from your last test? Ask another psychometrics prof to look at your last exam.
3) Criterion Validity – Looking at association of bivariate – behavior (criterion – will be placed on the y axis) and test (predictor – placed on the x axis). Correlating your test with actual behavior. Two different forms.
4) Construct Validity – correlating your test with a well established gold standard test.
5) Factorial Validity – Trying to test validity of some sort of theoretical structure. ie 3 factor structure (psychoticism, neuroticism, and extroversion). So you administer a test to lots of people, and see if in fact the outcomes boil down to the 3 theoretical factors.
Relationship between Reliability and error
There is an inverse relationship between reliability and error
Reliability = 1/Error
