Judith A. Vessey, PhD, MBA, RN, FAAN Rachel L. DiFazio, RN, PhD(c) Tania D. Strout, PhD, RN

Increasing Meaning in Measurement: A Rasch Analysis of the

Child Adolescent Teasing Scale(CATS)

Judith A. Vessey, PhD, MBA, RN, FAANRachel L. DiFazio, RN, PhD(c)

Tania D. Strout, PhD, RN

Declaration: No conflicts of interest

Funding: National Institutes of Nursing Research, R01 NR 04838

Background Youth Teasing and Bullying are a major

public health problem ~20% of youths report being bullied or bullying

at school in a given year 160,000 on any given day Significant psychological and physical sequelae

Healthy People 2020 Objective – reduce bullying among adolescents

State legislation requiring teasing and bullying interventions

Teasing

Dynamic social interactions comprised of a set of verbal and/or non-verbal behaviors that occur among peers and that is humorous and playful on one level but may be annoying to the target child on another level.

Bullying

Repetitive persistent patterns of conduct by one or more children that deliberately inflict physical, verbal, or emotional abuse on another child and where a real or perceived power differential is in place.

Teasing & Bullying

as a Continuum

Mean-spirited bullying that is poorly

received;Objective bullying

Good-natured teasing that is poorly received;

Chronic teasing/Subjective bullying

Mean-spirited bullying that is well received;Builds resiliency to

bullying

Good-natured teasing that is well received

Need for Measurement Theory

Goal: to allow for a safe acquirement and reproducibility of measuring characteristics

Analysis of data should reflect reality rather than just the numbers

Classical Test Theory vs.Latent Trait Models CTT: has the test (not the item) as its

basis› Although the statistics generated are often

generalized to similar students taking a similar test; they only really apply to those students taking that test

Latent trait models: aim to look beyond that at the underlying traits which are producing the test performance› They are measured at item level and provide

sample-free measurement

Items Analysis

Latent Trait Models

Classical Test Theory

Item Response Theory

Rasch Models

3P

1P2P

4P

Goals of Rasch Analysis

To demonstrate the relationship between item difficulty and person ability› The latent variable is conceptualized as

existing along a continuum› Items can be hierarchically ordered along

the continuum The score provides information

regarding what it means to be at a specific place on the continuum

Results of the CATS CTT Analysis

Final instrument: 32 items and 4 domains

DOMAIN # of ITEMS

EXTRACTEDVARIANCE

CRONBACH’S

ALPHA

Personality & Behavior 13 18.4% .89

Family & Environment 8 12.4% .84

School-related 9 11.8% .85

Body Size 2 8.6% .82

Purpose

To evaluate the degree to which the CATS items have been developed in accordance with the assumptions of the Rasch measurement model

Methods Methodological study design It was hypothesized that teasing/bullying,

as measured by the CATS items, are: › unidimensional in nature › follows a hierarchical order in the way that the

items define the variable› a continuum along which the CATS items can be

ordered and people experiencing› various levels of teasing/bullying can be placed

Secondary data analysis

Sample: 666 children aged 11-15 years

Methods Each CATS subscale evaluated

independently Winsteps v. 3.69.0 for Rasch analyses Rasch Rating Scale Model:

Methods Person & Item Separation Statistics

› Hierarchical Order Analysis of Fit

› Fit to Ideal Rasch Model Principal Components Analysis of Rasch

Residuals› Dimensionality

Variable Maps› Hierarchical Order› Continuum

Results The current CATS subscales were not

uni-dimensional, did not strictly follow hierarchical order, and did not stretch along the entire continuum

DOMAIN Unidimen-sionality?

HierarchialOrdering?

CoversContinuum?

Fit to Rasch?

Personality & Behavior No No No 3 underfits

3 overfits

Family & Environment No No No 2 overfits

School-related No No No 2 overfits

Body Size Yew No No Good fit

Variable Map PERSON - MAP - ITEM <more>|<rare> 1 + | | | | | | | T| | | . | . | . |T Who I live with My jewelry/chains # |S Way I dress My money Brand of shoes I wear What my family is like Being a nerd My schoolwork My parents

0 .# +M Grades Talking Friends Act weird/diff. Not know answers How talk Trouble Gay Chicken How do school Stuff Be dork/loser Shy/quiet Weird/diff friends Sports .#### |S Smart I am Way I act Not good at sports Not being popular Music I listen to

######### S|T Way I look My body shape My weight .######## | ######### | .########### | ############# | .######## | ######## | .##### | .##### | .###### | ####### M| .## | ###### | -1 .## + .#### | | .####### | #### | | .###### S| | | .###### | | | | | .##### | -2 + T| | | | | | . | .######## | | | | -3 .############ + <less>|<frequ> EACH "#" IS 4. EACH "." IS 1 TO 3

Personality & Behavior

Family & Environment

School-related

Body size

Domain

PERSON - MAP - ITEM <more>|<rare> 1 + | | | | | | | T| | | . | . | . |T Who I live with My jewelry/chains # |S Way I dress My money Brand of shoes I wear What my family is like Being a nerd My schoolwork My parents

0 .# +M Grades Talking Friends Act weird/diff. Not know answers How talk Trouble Gay Chicken How do school Stuff Be dork/loser Shy/quiet Weird/diff friends Sports .#### |S Smart I am Way I act Not good at sports Not being popular Music I listen to

######### S|T Way I look My body shape My weight .######## | ######### | .########### | ############# | .######## | ######## | .##### | .##### | .###### | ####### M| .## | ###### | -1 .## + .#### | | .####### | #### | | .###### S| | | .###### | | | | | .##### | -2 + T| | | | | | . | .######## | | | | -3 .############ + <less>|<frequ> EACH "#" IS 4. EACH "." IS 1 TO 3

Mean-spirited bullying that is poorly

received;Objective bullying

Good-natured teasing that is poorly received;

Chronic teasing/Subjective bullying

Mean-spirited bullying that is well received;Builds resiliency to

bullying

Good-natured teasing that is well received

ConclusionThe divergent results between the CTT and Rasch analyses, while not completely surprising, underscore the need for continued refinement of an instrument’s psychometric properties to ensure that is measuring the concept of interest in the way that it was intended. Reference: Vessey, J. A., DiFazio, R. L., & Strout, T. D. (2012). Increasing meaning in measurement: A Rasch analysis of the CATS. Nursing Research, 69, 159-170.

Thank you!