Item Analysis

Question 1

Mindfulness

Answer 1

Is a construct, a latent variable. It can be trait, state, practice, or skills. Mindfulness needs to be inferred and cannot be measured directly.

Question 2

15 statements, and the respondent ticks the appropriate answer on a Likert scale

Answer 2

Mindfulness Attention & Awareness Scale (MAAS)

Question 3

30 statements, and the respondent ticks the appropriate answer on a Likert scale

Answer 3

Freiburg Mindfulness Inventory (FMI)

Question 4

37 statements with a Likert scale. Scores are calculated for five different subscales

Answer 4

Five Facet Mindfulness Questionnaire (FFMQ)

Question 5

Generation of New Items

Answer 5

Top-down approach
>. Interviewing experts
>. Useful when the concepts is difficult to define and relies in insight and experience of experts (e.g., mindfulness)

Bottom-up approach
>. Interviewing/focus groups with relevant stakeholders (e.g., patients with certain conditions)
>. Works best for measuring subjective aspects such as attitudes, experiences etc.
>. Time consuming

Question 6

Piloting

Answer 6

A common way to test the appropriateness and comprehensibility of a measure is through cognitive interviewing
>.Participants complete the measure and think out loud
>.Often qualitative data are collected but also some quantitative

Question 7

Observed score = True score + Error
Emphasis in test development on correlations:
 >. Item – total score correlations 
 >. Cronbach’s α
 >. Factor analysis

Answer 7

Classical Test Theory - Factor Analysis

Question 8

Analysis of variance (ANOVA) to estimate variance components
>. Powerful method for examining test reliability

Answer 8

Generalisability Theory – ANOVA

Question 9

Alvin Feinstein, American physician, 1980s
Face validity is as important for clinical measures as statistical criteria.
>. Apgar scale, Glasgow Coma Scale Impact factor method for selecting items

Answer 9

Clinimetrics – face validity

Question 10

Each item must be shown to be measuring the same latent trait or variable
A precise statistical model defines the relationship between each item and the trait or construct of interest

Answer 10

Item Response Theory – Rasch analysis

Question 11

Specifies mathematical relationship between how much of a latent trait a person has and the probability of responding to an item
This may mean passing or failing on dichotomous (true/false) item or scoring higher on a Likert scale (1 , 2, 3, 4) on a personality measure

Answer 11

Item Characteristic Curve (ICC)

Question 12

Pros and Cons of G Theory

Answer 12

Pros:
>. Major advance on CTT
>. Powerful approach to reliability – e.g. we can look at single items not simply total score reliability
>. Tells us how generalizable our results from our sample are to the wider population

Cons:
X Requires good statistical background or friendly statistician
X Software more challenging
X Best kept secret in psychometrics

Question 13

They are deterministic Human responses are rarely so orderly
Rasch analysis aims to produce a stochastic or probabilistic Guttman scale

Answer 13

The problem with Guttman scales

Question 14

Rasch scaling

Answer 14

Rasch analysis aims to produce a stochastic or probabilistic Guttman scale
Guttman scale = unidimensional and hierarchical

Question 15

Pros and Cons of Rasch

Answer 15

Pros:
>. Robust method for developing unidimensional measures
>. Interval measurement
>. Requirement for certain journals e.g. JRM

Cons:
X Not widely understood in clinical world
X Not all constructs unidimensional
X Items with important clinical content may not fit the model