Term
|
Definition
CFA: Theory/Model Testing
EFA: Data exploration |
|
|
Term
| Exploratory Factor Analysis |
|
Definition
-Reducing individual items to make sense of things
-Data reduction
-Used for test construction (determining what goes on in each sub scale) |
|
|
Term
| Confirmatory Factor Analysis |
|
Definition
-Controlling what goes where!
-Test of construct validity
-Confirm items on particular sub scales
-Test measurement model as a part of validation research |
|
|
Term
| which tests whether data fit a model? EFA or CFA? |
|
Definition
| Confirmatory! this confirms what you decided.. |
|
|
Term
| which is theory driven? CFA or EFA? |
|
Definition
| Confirmatory! it confirms your theory! |
|
|
Term
| which tests whether groups differ? |
|
Definition
| Confirmatory! (ethnicity or gender differences in model fit) |
|
|
Term
|
Definition
1-Specify your model (what goes where of "Which items load on which factor")
-Internalizing/Externalizing Problems
2-CHoose computer program (AMOS, Mplus, LISREL, EQS)
3-set up the commands
4-Interpret the OUTPUT
5-Re-Specify and re-run it (If necc) |
|
|
Term
| In our model...item scores will not fully reflect the construct because....? |
|
Definition
| ...because ERROR contributes to item scores! |
|
|
Term
| Why do we look at Overall Model fit: Chi Square in CFA? Is Sig good or bad? |
|
Definition
-check for discrepancy between our our model and the data
-Bad! means poor fit or usually "not good" |
|
|
Term
| What does a Non-Significant Chi-Square mean in CFA? |
|
Definition
| your model is a good fit! |
|
|
Term
what are these?
RMSEA <.05
SRMR < .10
NFI > .90
CFI > .90
GFI > .90 |
|
Definition
|
|
Term
EFA (or PCA)
Rule of Thumb for sample size per item? |
|
Definition
| 5-25 participants per item... generally aim for a min of 100 |
|
|
Term
EFA:
you will get more _____ in small samples. |
|
Definition
|
|
Term
| is Varimax Orthoginal or Oblique? |
|
Definition
|
|
Term
| Downside or Challenges to CFA? |
|
Definition
| Its a perfect model in theory, but lets get real...there is always ERROR! |
|
|
Term
EFA (PCA- Principle Components)
Factor Extraction Assumptions |
|
Definition
-(PCA:Assumes 100% of variance can be explained
-with high loadings, similar Solutions
-with lower loadings, Most factor p. Axis |
|
|
Term
|
Definition
| superficial categories based on physical characteristics (Skin and eye color, hair texture) |
|
|
Term
|
Definition
membership in a SOCIALLY defined group (Not biologically)
-indicates national origin (Mexican American) |
|
|
Term
Problems with
"Ethnicity" |
|
Definition
| doesnt account for within-grp variations, or variations in culture whining one geographical region |
|
|
Term
|
Definition
| Broader Term! Includes: traditions, customs, language and other patterns specific to a group of ppl |
|
|
Term
| "Culture" goes beyond Race and Ethnicity to include: |
|
Definition
| Religion, geographic location (Urban vs Rural), sexual orientation, age, handicapped condition, and SES |
|
|
Term
| Okazaki and Sue (Article) argue... |
|
Definition
..."Specify and DIRECTLY measure the underlying psychological variables associated with culture that are hypothesized to PRODUCE GROUP DIFFERENCES" such as:
-Acculturation
-Individualism vs Collectivism
-Racial or Cultural Identity |
|
|
Term
| Cultural Methodological Issues: |
|
Definition
-Self-Identification (how broad to make category?)
-Rationale for hypothesized grp differences
-Selecting Participants / Recruitment (where you are located may affect your sample)
-SES (Inflexible work, transportation issues, childcare may be a problem for lower ses) |
|
|
Term
| methodologies issues: what the cultural issue with EBT? |
|
Definition
Evidence based Tx are supported by research which doesnt take into account DIVERSITY issues!
-Result: May be disproportionally represented within certain groups. |
|
|
Term
| Face Validity Cultural Bias? |
|
Definition
| Does it LOOK bias? Is it off-putting to a specific cultural group? |
|
|
Term
| Test Bias: Content Validity Bias? |
|
Definition
| Is your MEANING of your construct the same across cultures? across groups? |
|
|
Term
| Okazaki and Sue (Article): Conceptual Equivalence? |
|
Definition
| Same and Content Validity bias; A construct may not MEAN the same thing across groups. Depression may have a different meanings across cultures. |
|
|
Term
| Name approaches to building in Culturally relevant content into a measure.... |
|
Definition
-KNOW your construct
-Study your target population
-Appraise the population used in developing the measure!? (standardized on who?)
-Use Cultural Informants
-->Informally and Formally |
|
|
Term
| Possible Bias Content Validity Problems in an existing measure.... |
|
Definition
-irrelevant content
-Inadequate content
-different grps find the test MORE or LESS uncomfortable to answer specific questions |
|
|
Term
| MORE approaches to building in Culturally Relevant Content? (3) |
|
Definition
1) Create content relevant to only ONE group
2)Seek content common across groups
3) Develop content relevant to Different Groups |
|
|
Term
Statistics for Bias!
Slope Bias |
|
Definition
| Do Scores Predict Equally well? (Different validity r for members of diff groups?) |
|
|
Term
Statistics for Bias!
Confirmatory Factor Analysis
(Differential Validity) |
|
Definition
| do factors load the same across diff groups? (is the conceptual model meaning the same in different groups)... EXAMPLE: a measurement model has been created BUT if hasnt been tested in a particular group. Does the model Fit? Does it Generalize to this other group?? |
|
|
Term
Statistics for Bias!
Intercept Bias
(Differential Prediction) |
|
Definition
| scores systematically over-predict or under-predict criterion for one group (EX: same score on GRE is telling us different outcomes for foreign born students compared to US born students. Same Score should mean same outcome) |
|
|
Term
|
Definition
| "Systematic error in which SCORES behave differently for one group versus another" |
|
|
Term
|
Definition
| measure systematically produces distorted picture of a characteristic for a person |
|
|
Term
|
Definition
| decision made from measure is systematically biased... decision made from the measure. |
|
|
Term
| Does a MEAN DIFFERENCE = Bias? |
|
Definition
| Not necessarily... groups may REALLY DIFFER. |
|
|
Term
| a Steeper slope (slope bias) means? |
|
Definition
| higher r between criterion and predictor |
|
|
Term
| if flat line in slope bias? |
|
Definition
| its bad. means test isnt predicting a good r between criterion and predictor |
|
|
Term
| what does Urbina call Intercept Bias? |
|
Definition
|
|
Term
| Issues when selecting measures for use with diverse populations |
|
Definition
Test may require adaptation:
-language
-change in mode of administration
-Different Content
DATA may not show equivalence across groups
LACK Of data for target group |
|
|
Term
| Possible solutions (for using measures in diverse populations) |
|
Definition
-Use a measure with satisfactory psychometrics for majority culture
-ESTABLISH psychometric properties of a measure within diverse samples
-Build tests of reliability and validity into study/existing measure (This is time consuming and may not work)
-BUILD A NEW MEASURE?
-Use existing measure but adapt (add/remove content) |
|
|
Term
| Issue with Self Report and Bias? |
|
Definition
| Do people respond to items in the way intended by the test developers? |
|
|
Term
Self-Report: (Cognitive Processing)
Cognitive Validity? |
|
Definition
| the Degree of Congruence between intended and actual cognitive processes involved when responding to a self report item. |
|
|
Term
Self Report:(Cognitive Processing)
Cognitive Pretesting? |
|
Definition
| Semi-structured interview method for ascertaining info about what respondents are thinking / how they are responding to self report questions |
|
|
Term
Self Report:
Recall and Memory issues? |
|
Definition
-people don't remember individual instances...they have global impressions
-people Over estimate BAD bx and underestimate FREQUENT ones
-Specific bias (Current head space)
-Psychopathology (Ex: depressed informants = memory for negative events) |
|
|
Term
|
Definition
| tendency to present oneself in an OVERLY favorable way |
|
|
Term
|
Definition
| intentionally making oneself look worse (medically "sick" or psychologically unhealthy) than one actually is |
|
|
Term
| Ways to prevent Social Desirability |
|
Definition
-create items with equally desirable alternatives
-minimize anxiety
-Anonymous responding
-reduce items which pull for SD |
|
|
Term
| Statistical Solutions for Social Desirability |
|
Definition
-Measure SD then control for it
-Filter high scorers OUT of sample
-Dont use SELF-REPORT! |
|
|
Term
|
Definition
Ratings based on overall impressions
-the good one
-the loud one
-the smart one |
|
|
Term
|
Definition
| when people think two things are always related. (His academic level is low, therefore he must have ADHD) |
|
|
Term
|
Definition
| rating those we know (and like) more leniently |
|
|
Term
|
Definition
| giving special consideration to those whom we feel are similar to ourselves |
|
|
Term
|
Definition
| different constructs assessed by the same method likely have variance in common (correlations may be inflated by the same the mothod due to SMV) |
|
|
Term
| Things that contribute to scores (3) |
|
Definition
1) Trait/construct (true score)
2) Random error
3) Method of assessment |
|
|
Term
| In Method Variance Systematic or unsystematic? |
|
Definition
|
|
Term
| MTMM looks at which 3 things? |
|
Definition
Convergent --> <-- Validity
Discriminant <-- --> Validity
Shared Method Variance |
|
|
Term
MTMM
Convergent validity should be ________ than Discriminant validity? |
|
Definition
|
|
Term
|
Definition
1) Convergent Validity
2) Discriminant Validity HTHM (Dotted Triangles)
3) Discriminant Validity HTMM (Solid Triangles)
4)Discriminant Validity SMV (ONE METHOD AT A TIME)
(3 mini steps) |
|
|
Term
| Sources of Poor Validity r's in MTMM |
|
Definition
-Poor reliability
-Restricted range
-Poor content validity (diff operationalizations) |
|
|
Term
Convergent Validity should be bigger than ALL discriminant validity r's...
TRUE -or- FALSE |
|
Definition
|
|
Term
| 3 Mini steps in step 4 of MTMM |
|
Definition
Step 4) Discriminant Validity SMV
1) HTMM r's =SMALL (<.35) and NONsig!
2)Compare Convergent Validity r's w/ HTMM (Convergent should be bigger)
3) HTMM -vs- HTHM |
|
|
Term
|
Definition
| Degree of congruence between INTENDED AND ACTUAL COGNITIVE PROCESSES involved when responding to a self-report item. How they responded versus how you intended them to respond |
|
|
Term
|
Definition
| Whether the psychological CONSTRUCT under investigation holds the SAME MEANING in two or more cultural groups. May be dependent on the context in which the assessment takes place |
|
|
Term
|
Definition
| 1. Confusion 2. Carelessness (college students just rushing through measure because they are forced to) 3. Yea-saying or nay-saying. UNSYSTEMATIC ERROR IS ALWAYS A PROBLEM! |
|
|
Term
|
Definition
| Error related to the CONTENT OF THE MEASURE. 1. Social desirability response bias 2. is source of bias an aspect of the construct being assessed, part of psychopathology! |
|
|
Term
| Okazaki and Sue (Article) |
|
Definition
-Assumptions should be made clear
-fuller discussions
-report more stats and methods! (for meta analysis)
-Consider using multiple measures
-Consult cultural professionals |
|
|
Term
| What is the issue of Cognitive processing? |
|
Definition
How are items being interpreted??
Core issue: do people respond to self-report items in the way intended by the test developers? |
|
|
Term
|
Definition
| Yea-saying and Nay-saying |
|
|
Term
| Podsakoff article refers to the need for |
|
Definition
| social approval and acceptance and the belief that it can be attained by means of culturally acceptable and appropriate behaviors |
|
|
