Class 3 Flashcards
What are the principles of administration
- Follow standardized procedures (important to have same stimuli for everyone)
- The examiner is non-directive
(let the client lead) - Capture the client’s performance accurately
(Write down all that is said by the client and you) - Focus on what the client sees and how they see it (i.e., solve the problems the inkblots presents)
Materials needed for the Ror.
The 10 Rorschach cards in order
At least 2 Location sheets (out of sight)
Plenty of paper
Two pens
Clipboard
OR now with the RPAS administration you can use a laptop (RPAS provides a Word template for ease on laptop)
How long do you need for the test
1 hour
Before starting, you need to ask what ….
Ask what they know about the Rorschach…
Investigate their knowledge of the Rorschach
Clarify any misunderstandings
“I want you to tell me what they look like to you”
Be worried about possible coaching or motivated distortion
Ask directly about preparation
If so, decide whether to proceed
If proceed, ask client if willing to be honest and spontaneous
Be familiar with public availability of materials
Two phases of the Ror.
- Response phase (RP)
What might this be? - Clarification phase (CP)
Carefully worded questions to resolve coding ambiguities
Things not to say or do
Avoid saying “ambiguous” or “unstructured”
Avoid saying there is no right or wrong answer (Can say “Different people see different things.”)
Avoid saying “most people…can or should say or do something”
Do not mislead the client into thinking that imagination or creativity is being tested
“Can I turn it? Should I use the whole thing?”
“It’s up to you.”
“How are they made?”
Briefly explain and move on
“Does everybody look at different ones?”
No, it’s a standard set; everyone gets the same ones.”
“What does it mean? Is there a right answer?”
“Let’s discuss that once the testing is completed.”
“Do you see it?”
Response phase: “Let’s come back to that after we go through the cards.”
Clarification phase: Yes, I see it. Or No, not yet.
Prompt
If only 1 response
> PROMPT for another, but only 1 prompt per card
> first occurence say “We would like 2 or maybe 3 to each card, so please try to give another.”
Pull
After 4 responses, PULL the card
Can you accept a card rejection
Don’t accept card rejections
“Take your time, there is no hurry. Look some more to see something else too.” Maybe add “You can do it.”
How many responses do they need? What do you do if they don’t get enough
More than 15, they are good. If 15 or fewer, go back through the cards. AND….
Do not use extra prompts
Not necessary to get a response on each card
Do not present a card on which they already gave 4 responses
Once there are 4 responses, pull the card
Add these additional responses to the protocol and in the Clarification phase, clarify card by card (integrating the supplementary responses with each card)
What happens in Phase 2 (clarification)
Used only to resolve scoring uncertainties (location, determinants)
Repeat each response verbatim
Clarify key word or phrases:
(Pretty, ugly, rotten, wild, mysterious, back there)
Focus the questions to resolve coding dilemmas.
Write down everything said as best you can.
Answer questions about the CP directly
Write down what is said and note significant gestures
Identify non-obvious and salient components on the location sheet
In CP do you always need to clarify?
Clarification questions are not always needed (you may already have enough information to score from the RP)
Location clarification is rarely needed (no need to be overly precise)
Language in CP
Be as non-directive as possible
to minimize the danger of influencing the client
Use visual language with client in the CP (“Looks”, “see” rather than “suggests…”)
Focus your attention on key words or phrases from RP and early CP (That suggest but do not confirm a determinant)
Avoid vague questions (“What makes it l.l. that)
Don’t need to see it exactly as the client does
Vary language so as not to set a problem-solving set
When confusing response during CP, say…
“I’m not sure how you see it”
What if you are unsure if the response is one or two responses?
Sometimes it is necessary to clarify whether an ambiguous verbalization is one or two responses
Wait until the CP, and repeat back the response and often the client will spontaneously clarify the question in your mind
If client reported 2 things, read the first and see if they clarify by including the second
If still not clarified, you may need to ask directly, “Was that one or two answers?”
Clarification Don’ts
Don’t ask questions that don’t have a purpose. Target one or more specific coding category
Don’t ask too many questions
Don’t keep going after a score you “know” must be there.
Don’t ask leading questions:
(Is it moving? are you using color?)
Documentation
Document well!
Document enough that another person could code the record (using commonly known shorthand)
Verbatim response, relevant gestures,
Orientation of the card (<v>)</v>
Note Prompts (PR) and Pulls (PU)
Location on the location sheet during clarification
Ask client to slow down or repeat as necessary
Put examiner comments/questions in parentheses
Document if supplemental responses were needed
Code as soon as possible!
Do you code things off the card? Ex. code the car in “a dog that got it by a car”
Code what is on the card, not something off the card
(although it doesn’t have to be seen, as long as it resides on the card)
What do you code?
Code what is articulated and sometimes gestures
(rubbing the card)
Code how client saw this cloud at this time
(not clouds in general)
Code what is on the card, not something off the card
(although it doesn’t have to be seen, as long as it resides on the card)
Code categories independent of each other
What if CP contradicts the RP
Code what was seen in the RP
Ignore CP information that contradicts the RP
Coding accurately vs seeing it the way they do
The goal is to code accurately, not necessarily to see it the way they do
The reasonably certain standard
If “reasonably certain,” code it
Main issue of coding problems are
Coding problems are most often due to clarification problems
Inter-rater reliability most well known study
The most thorough and meticulous is by Acklin, McDowell, Verschell, & Chan (2000)
Inter-rater reliability – how many studies done
Interrater agreement is a critical aspect of coding the Rorschach
Over 85 studies done on Reliability of Rorschach scoring/coding over the years
Viglione, Blume-Marcovici, Miller, Giromini, and Meyer (2012)
Two graduate students independently coded 50 Rorschach protocols administered to adults and children
Results:
Mean ICC of all 62 RPAS indictors = .88 (Std Deviation =.11; Median = .92)
Kivisalu, Lewey, Shaffer, and Canfield (2016)
everything italicized under it
evaluated interrater reliability for 50 nonclinical R–PAS protocols–at the response level of analysis.
Each protocol was coded twice, first by the original examiner and then by a blind coder
Mean ICC = .78 from the 62 codes at the response level
Pignolo et al (2017)
Standard scores and Complexity-adjusted scores in an Italian sample. Coded separately by 2 separate coders. italicized
Most scores in good to excellent range
Raw score mean = .78 (Std.D = .14)
Complexity-adjusted mean = .74 (Std.D = .14)
TEST-RETEST RELIABILITY
Pretty good, book has chart of 1 and 3 year re-test
Pretty stable over long intervals
(Some even in .90)
State variables
- Navy psychologist
- Paratroopers
- Surgery patients–elective
> 38 test-retest reliability studies done by 1985
What is test-retest not stable on
Y and m (inanimate movement and diffuse shading)
Study of test-retest reliability
Children in elementary school. Took children of the classroom to help “train” examiners
Day 2 (3-4 days later)
- Control group (same directions)
- Experimental group (told to give different responses, and they did)
FOUND:
85% same responses in the control group
14% same responses in the experimental group
Yet the correlations between the Structural Summary for each child was similar in each group!
AKA its not about you what you see, its how you see it.
Replicated in 1982 with adults (Nancy Haller) –> Similar results as with the children
Overall review of reliability
Test-Retest reliability seems strong
Inter-rater reliability seems to be good except for a few indicators
Since the Acklin et al study, a coding book has been published with detailed guidelines covering how to code, presumably increasing those variables that showed lower inter-rate reliability
Card orientation
< top facing left
v top facing down
>
top facing right
@ response is upright but the card was turned at least 90 degrees before giving the response
Reasons to Document Orientation
- to help coding FQ
- to help interpret reflections
[If card held upright but response is upside down, code the cards orientation as upright, and note location as upside down]
Prompts and pulls
Code PR or PU for every card for which you are required to prompt or pull.
If a client spontaneously hands the card back after the 4th response, do not code PU.
Location codes
W (Whole)
D (Common detail) *
Dd (unusual detail) *
*found in the FQ tables in Chapter 6, along with specific numbers (Dd99 means not in the table)
If multiple content, list location of all that apply in the Loc # column.
What if considering Dd, but almost entirely covered by a D response
Do you code Dd for near-D?
Do not code Dd for near-D if they differ only slightly from the D location. (unimportant location differences) (pages 66-67)
What location code do you use for symmetrically paired objects
“Code D for symmetrically paired objects that together reside in D location
Using white space – codes
SR: if space is the foreground (AKA SEEING WHITE SPACE)
- space is normally background but score SR if it becomes foreground
- Shows: Swimming against the current, self-assertion, resisting pressure, thinking outside the box, or oppositionalism (p. 72-73)
SI: If space is articulated detail inside or integrating an ink part as well (AKA USING WHITE SPACE WITH INK PART)
Can score both (see p. 72)
- ex. mask with white eye wholes
How do you code identifying lightly colored areas as “white”
Code SI, if the response is integrated with other parts of the blot
Do not code SR
Coding SR and SI: 4 ghosts dancing in the dark
SR: space reversal to see independent objects in the space
AND
SI: see white integrated with the ink
Content codes- what are they, do you need to have one, can you have multiple?
They describe “what” is seen in the card.
*Code all the categories that apply but only code each category once per response
*Must have content code
Sy and Vg
Sy: Synthesis
– Distinct and separate objects in a relationship
Vg: Vagueness
– objects with vague or indistinct outline or boundaries (no form demand) and no form added/injected
Synthesis thresholds
Positive threshold examples for coding Sy: “leaning on” or “leaning against,” “touching,” “looking at each other,” “standing behind (when one object obscures another),” “a bird on a branch,” “a person casting a shadow,” “a person on a rug,” “a fetus with its umbilical cord attached,” “flying above this object,” “lying next to,” “lying beside one another,” or one object “coming from” another, “trees on a pond.”
Negative threshold examples that are not coded for Sy: “a garden scene with yellow flowers, kind of like roses, daffodils, tomato horn worms,” “an underwater scene… there’s fish, rocks, plants,” “a landscape with rocks, trees and a pond,” “a sculpture of a bear and a fish.”
Pairs (2)
Two identical objects symmetrical on each side of the blot midline (not parts that come in pairs, like eyes, boots, hands of a whole person)
must be identical in all respects
if differentiated in any way, do not score pair
(Goes into summary calculations related to symmetry)
– don’t score pair when reflection is scored
Form quality (and do you code up or down)
If a response has multiple objects, code the FQ of the material objects (not an inconsequential part of the response) with the lowest FQ
FQ Tables are based on…
FQ tables are based on accumulated FQ tables (Beck et al (1961), Hertz (1970), Exner (2003), and international sample.
Based on
1) fit (derived from multiple judges from 11 countries)
2) frequency (from the multiple data sources above)
FQ and Extrapolation General
FQ extrapolations must have the same shape, form features, and spatial orientation (not just content)
FQ and Extrapolation Steps for single object
- Search the FQ tables for responses with similar shapes. (if it is Dd99). If straightforward, assign the FQ.
- Search like areas
- Search subcomponents of an object e.g., bird > wings; bull > horns
- Consider results and make a determination.
FQ and Extrapolation Steps for multiple objects
Score the lowest form quality of important components
What does FQ reflect on someone’s thinking
FQ reflects conventional and accurate ways of processing information
