Reading Self-efficacy

Get Complete Project Material File(s) Now! »

CHAPTER THREE STUDY ONE: COMPARISON MODEL-DATA FIT FROM IRT AND IDEAL  POINT IRT MODELS TO AN IDEAL POINT SCALE

Study One sought to compare the effectiveness of an ideal point model to a personality scale developed under the principles of an ideal point scale. In addition, the goodness-of-fit of a traditional polytomous dominance IRT model (GRM) to the ideal point developed scale is examined. All previous comparative fit analysis (e.g., Chernyshenko, Stark, Chan, Drasgow, & Williams, 2001; Chernyshenko, 2002) have been applied to response data obtained from dominance constructed scales, thus, no attempt has been made to examine the degree of fit provided by a dominance IRT model to an ideal point constructed personality scale. As such, the present investigation extends previous research by examining the fit provided by an ideal point IRT model and a dominance IRT model to an ideal point constructed personality scale.
Model-data fit will be compared using the category response functions generated under two IRT models representing each approach, namely the ideal point-based generalized graded unfolding model (GGUM) and the dominance-based GRM. The two parametric models were chosen for comparison based on the similar characterization of responses, namely, models estimate the same number of parameters, and allow for both discrimination and response thresholds to vary across items. In addition, the GRM and GGUM are both appropriate for polytomous resp Given the argument offered by Chernyshenko et al. (2001) that an ideal point response process may be a more appropriate approach to modeling personality data, and further findings confirming this conjecture (Chernyshenko, 2002), it is expected that the ideal point model (GGUM) will provide better fit to the ideal point response patterns than the dominance-based GRM. The research question for this study was: Which measurement model provides the best fit to the ideal point constructed Academic Self-Worth Scale?

Method
Participants

Participants consisted of 1020 secondary school students (527 females, 493 males) from four secondary schools in the Auckland area. Students ranged in age from 16-20 years, with a mean age of 16.70 years (SD = .77) covering secondary school Years 11, 12, and 13.

Scale Development
Instrument

Participants completed the 74-item Academic Self-Worth Scale (ASWS) consisting of three sub-scales measuring Defensive Expectations (30 items), Reflectivity (15 items), and Self-handicapping (29 items). Items for each sub-scale were developed and adapted using ideal point scale construction guidelines (see Procedures Section), from pre-existing scales for each construct namely the Academic Self-Handicapping Scale (Midgley, Arunkumar, & Urdan, 1996), Life Orientation Test (Scheier, Carver, & Bridges, 1994), Academic Process Questionnaire (Martin, 1998), and Defensive Pessimism Questionnaire (Norem & Cantor, 1986). Within the Self-handicapping and Defensive Pessimism sub-scales, six distinct strategy approaches (see Table 1) are represented.

Self-handicapping items

Self-handicapping items were derived and adapted from Midgley et al., (1996), Strube (1986), and Martin (1998). Active self-handicapping items reflect the active impediments that an individual will employ in order to purposefully impede upcoming measurements of their ability. Items include excuses that typify the deliberate and destructive actions adopted by students, such as, “I let myself get run-down when assignments or exams are due”, and “Instead of doing study after school, I tend to hang out with friends during the term”.
Similarly, self-presented self-handicapping items consist of content that portray the individual as having obstacles to their performance, but unlike the active self-handicapping content, these impediments either do not actually exist or are excessively exaggerated (Martin, 1998). Here individuals are not motivated to actively ruin their chances of performing well, instead they focus on presenting themselves as being disinterested or at a disadvantage, thus publicly managing any potentiality of academic failure. Examples of such image management items are, “I tell people that I put assignments and study off until the last moment more than I actually do”, and “I let people know that I am involved in a lot more activities than I really am when exams or assignments are due”.
In addition, self-presenting affective items were also incorporated to reflect the less tangible emotional states presented by individuals. Like the self-presented items, the individuals are either not actually experiencing the portrayed affective states or the extent to which they are occurring is significantly embellished. For example, “When an exam or assignment is due, I let people think that I’m more tense and uptight than I really am”, and “When I have an exam coming up, I sometimes tell others that I’m more frustrated with my prep than I really am”.
Based on the item design from Midgley et al. (active and self-presented) and Martin (self-presented affective self-handicapping), items finished with a similar trailer relating to the motivation for presenting an image regarding their upcoming performance scenario (see Table 1). The use of this trailer helps maintain the respondents’ focus on the underlying reasons for presenting their alibi for potential poor performance, a priori (Martin, 1998).

Defensive pessimism items

Defensive expectation and reflectivity items were based from the strategy prototypes devised by Norem and Cantor (1986) and the Life Orientation Test (Scheier, Carver, & Bridges, 1994), and drawn from the further adaptations made by Martin (1998).
In this scale, active defensive expectation items reflected the strategies whereby individuals adopt unrealistically low academic expectations in order to protect themselves from the negative expectation of failure (Norem & Cantor, 1986). Norem and Cantor’s original measure of defensive pessimism was conducted after scores from the Optimism-Pessimism Pre-screening Questionnaire had been scored. This approach was based on the assumption that students who were pessimistic regarding their performance on previous assessments were more likely to have defensively pessimistic orientation. Martin’s adaptation of Norem and Cantor’s prototypes does not make such an assumption, and suggests instead the use of the predicate “No matter how well I have done in the past…” as a way of integrating a screener into each of the items (see Table 1).
As with self-handicapping strategies research Martin (1998) adapted active defensive expectation scenarios into self-presented versions. Also, as with the self-handicapping counterparts, the content in the self-presented defensive expectations items reflect students presenting themselves as being more pessimistic than they actually are towards upcoming performance scenarios on their ability (e.g., “I tell everyone that I question my academic ability, more than I actually do”).

CHAPTER ONE  INTRODUCTION
Purpose of the Research
Significance of the Research
Design of the Research
CHAPTER TWO  LITERATURE REVIEW
Self-worth Strategies
Self-handicapping
Active self-handicapping
Self-presented self-handicapping
Self-presented affective self-handicapping
Defensive Pessimism
Active Defensive Expectations
Self-Presented Defensive Expectations
Reflectivity
Response Behaviors
Measurement Theory
Item Response Theory (IRT)
Graded Response Model (GRM)
Ideal Point Item Response Theory
Generalized Graded Unfolding Model (GGUM)
Applications of Ideal Point and Item Response Approaches to Personality Items
Reading Self-efficacy
Innovative Item Design – Framework
Multimedia Testing
Equivalence Testing
Efficiency Testing
CHAPTER THREE STUDY ONE: COMPARISON MODEL-DATA FIT FROM IRT AND IDEAL
POINT IRT MODELS TO AN IDEAL POINT SCALE 
Method
Participants
Scale Development
Instrument
Self-handicapping items
Defensive pessimism items
Procedure
Statistical Analysis
Cross-validation
Exploratory Factor Analysis of the Academic Self-Worth Scale
IRT Item Parameter Estimation
Model-Data Fit
Results
Exploratory Factor Analysis of the Academic Self-Worth Scale
IRT Item Parameter Estimation
Model-Data Fit
Discussion
CHAPTER FOUR STUDY TWO: A LITERATURE REVIEW OF HUMAN, TECHNOLOGICAL
AND ITEM DESIGN ISSUES IN COMPUTERIZED TESTING 
Participant Issues
Race, Ethnicity, and Gender
Cognitive Processing
Ability
Familiarity with Computers
Computer Anxiety
User Interface – Legibility
Screen Size and Resolution
Font Characteristics
Line Length
Number of Lines
Interline Spacing
White Space
User Interface – Interactive
Scrolling
Item Review
Item Presentation
Computerized Item Design Types
Item Format and Response Actions
Mouse-based Response Action
Figural Response Items
Drag-and-drop Item
Graphical Modeling Item
Drag-and-connect Item
Specifying Relationships Item
Create-a-tree Item
Capturing Frames Item
Multiple Selection Item
Analyzing Situation Item
Text-based Response Action
Essay/Short Answer Item
Generating Examples Item
Document-analysis Item
Text-editing Item
Innovative Item Response Formats
Task Constraints
Discussion
CHAPTER FIVE STUDY THREE: WEB-BASED AND PAPER-AND-PENCIL VERSIONS OF THE READER SELF-PERCEPTION SCALE: A COMPARISON OF MEASUREMENT EFFICIENCY AND PARTICIPANTS’ PERCEPTIONS
Method
Participants
Design
Measures
Reader Self-Perception Scale (RSPS) – Paper-and-Pencil
Reader Self-Perception Scale (RSPS) – Web-based
RSPS – Progress sub-scale modifications
Post-assessment feedback measure
Procedure
Paper-and-pencil administration
Web administration
Results
Data Analysis
Descriptive Statistics
Confirmatory Factor Analysis
Item Parameter Statistics
Efficiency Analysis
Item Level
Test Level
Mode Preferences
Discussion
CHAPTER SIX DISCUSSION
Findings
Implications
Future Research
Contribution
REFERENCES
GET THE COMPLETE PROJECT
MAXIMIZING INFORMATION: APPLICATIONS OF IDEAL POINT MODELING AND INNOVATIVE ITEM DESIGN TO PERSONALITY MEASUREMENT