Chapter 3: Assessments, Effect Size, and Ethics Flashcards
What are personality assessments for?
An attempt to measure traits (determine motives, intentions, goals, strategies, & subjective representation)
What are the categories of personality tests?
(Omnibus or specific/single trait)
- S-Data or B-Data
- Projective tests or objective tests
What are projective tests? (+ two types)
B-Data, inkblot, drawing, story
- problem: not certain meaning, clinicians might be fooling themselves or use as an ice breaker
- must have standard scoring
- Rorschach: widely used, best when scored by two techniques, some validity
- TAT: implicit motives (not show how pursue), help with predictions
What are objective tests?
Questions not really objective, there is still interpretation required, but ambiguity is necessary for there to be benefit.
- a lot of test questions are needed to increase reliability
What are the three types of creating objective tests? Best one?
Rational (logic) - many still use
Factor - few use
Empirical (reality speak for self, certain kinds people answer similarly) - few use
- best test makers use a combination, intentional questions that confirm similarities, and are able to make predictions
What are rational object tests?
Questions rationally relate to what the test measures (S-Data)
- question must mean same to participant as did to researcher
- participant must be capable of self-assessment
- participant must be honest
- test questions must be valid
What are factor object tests?
Bunch of items get grouped together statistically then psychologist must determine how the factor they have in common
- factor label is highly debatable
- limitation: quality of test depends on quality of items
What are empirical object tests?
Allows reality to speak for itself
- get items, get participants, test, compare answers, cross-validate,
- believes similar kinds of people will have similar answer
- not concern with identity of actual items –> hard to fake answers, only as good as criteria/what cross-validated against, no way to determine if valid at another time or group of people and serious problems with the law/public relations
How are assessments evaluated?
- significance testing and effect size
What are the aspects of significance testing?
Significance means that the results are not likely due to chance
- usually use NHST with 5% rule
- p-value: probability of finding difference if it does exist which addresses type I error (determining effect when it is not there)
- NHST not very good to determine if according to chance or not, replication is a better test
What are the aspects of effect size?
It is an indication of the strength of a relationship
- most commonly determined by the correlation coefficient
- Calculation: x and y columns use formula, can convert between this and statistics for experimental research
- Interpretation: ranges with “quite strong” sorts of phrases or square value to get a percent that is “explained” –> thought that correlation does not do much
- Binomial Effect Size Display (BESD) begin assumption of “no effect” so 50/50 so 50 add correlation coefficient to this by turning it into a two-digit number than dividing by two and adding it to the 50
What are the ethical issues of the purpose of personality tests?
Is it for the benefit of the tester or the benefit of someone else, like employer
- employment screen may be problem as it may not be accurate
- however, judgements will happen anyway so this may be better than other forms of decision making
What are the ethical issues of the use of personality research?
- Question: wouldn’t you want to know
- Issue: behavior control research, who decides who gets to control the research
What are the ethical issues of truth and deception?
- Truth: without truthfulness the attempt of research is worthless
- Deception: gave informed consent, does no harm, cannot be investigated otherwise
- leads participants to not trust
What are theories?
1) An attempt to explain a phenomena
2) with tests to predict new information
* The question is more important than the answers
What make good tests?
- How well they explain or predict
- How well they fit within ones worldview
- Is it stimulating
- Establish relationship among variables - which cause which, or additional variable cause it?
How do we know what we know
- Rationalism (logic)
- Empiricism (evidence)
- Faith: saved by faith, calling on His name, it is the evidenced of what we hope for cand cannot see, believe that God exists and rewards
** We have the answers the rest of the world is seeking
What does the scientific approach cause and what should we consider?
- Healthy skepticism or unhealthy cynicism (we must be skeptical, humble, and curious)
- May discourage reliance upon authority and intuition as ways of knowing
- Compatible with the theistic worldview?
What is cross-validation in personality tests?
Ask about the same trait in different ways to validate the trait, limiting chance answers (MMPI)
What are ethical principles in personality testing?
Confidentiality, conflict of role/interest, informed consent, discrimination (do not do this by personality tests), abuse, user qualification (is the administrator qualified, does the one taking the test meet the qualifications), privacy
What are types of research?
- descriptive (qualitative: transforms data into manageable units) / correlational
- experimental (manipulation, may be deceptive, IV, DV, causational)/inferential
- looking for relationships
What are the aspects of the Null Hypothesis Statistical Test?
likeliness of a result
- Hypothesis
1) Null: no relationship
2) Relationship between variables exists
- Significance: relationship due to chance
- Reject/fail to reject the null
- Errors:
- Type I: reject Ho that was true (most say this is worse, life it depends)
- Type II: accept Ho that was false
What is the effect size?
Size or extent of the result
- correlation coefficient: Pearson’s r -1.0 to + 1.0
- squaring effect size/BESD
What are the potential relationships between variables?
Positive linear relationship
Negative linear relationship
Curvilinear relationship
No relationship
