26
Rita C. Ramos, R.N. Assistant Professor University of the Philippines Open University Faculty of Management and Development Studies

Final PPT_ Item Response Theory

Embed Size (px)

Citation preview

Page 1: Final PPT_ Item Response Theory

Rita C. Ramos, R.N.Assistant Professor

University of the Philippines Open UniversityFaculty of Management and Development

Studies

Page 2: Final PPT_ Item Response Theory

Medical Surgical Nursing covers greater percentage in the total number of hours in the nursing curriculum

Statistics show s that there has been increasing percentage of failing among test takes in the Nurse Licensure Examination ( PRC 2010 )

Page 3: Final PPT_ Item Response Theory

1. To establish the validity and reliability of the instrument using CTT and IRT framework

2. To determine the dimensionality measure of items.

3. To compare the item difficulty and item discrimination of the Medical Surgical Nursing Achievement Examination using CTT and IRT

Page 4: Final PPT_ Item Response Theory

137 4th year nursing students form a private university in Manila.

Selected on the basis of the completion of all subjects in Medical Surgical nursing

Page 5: Final PPT_ Item Response Theory

Nursing Achievement Test for Medical Surgical Nursing ( NAT- MSN ) is comprised of 219 items of multiple choice

Parallel to Nursing Practice III, IV and 5 of the Nurse Licensure Examination

Page 6: Final PPT_ Item Response Theory

Test IIIClient in Pain

Peri- operative Care Alterations in Human Functioning

Test IVAlterations in Human Functioning

Client in Biologic CrisisEmergency and Disaster Nursing

Test VDisturbances in Perception and Coordination

Medical

Surgical Nursing

Page 7: Final PPT_ Item Response Theory

Test objectives for each major categories and its subsets were derived from course syllabi

11 topics derived from three major categories ( Nursing Practice 3, 4 & 5 )

Nursing subjects have both theoretical and clinical components ( Theory and Related Learning Experience )

Page 8: Final PPT_ Item Response Theory

Total hours summed up to 408 hours in the 11 topics ( Theory and Related Learning Experience )

Total item computed for each subtopic were distributed according to five domains of New Bloom’s Taxonomy .

The final structure and draft of the achievement examination was arranged according to the main three parts : Nursing Practice III ( 100 items ), Nursing Practice IV (100 items ) and Nursing Practice V ( 19 items )

Page 9: Final PPT_ Item Response Theory

Seeking of permission Test administration with Research assistant

Page 10: Final PPT_ Item Response Theory

Reliability of the achievement examination utilizing framework of Classical Test Theory using SPSS version 11.5 ( Chicago , Illinois ).

Item difficulty and discrimination were computed and analyzed according to formula

WINSTEP version 3.69 ( Linacre 2010 ) was used to assess the following :unidimensionality, hierarchical ordering of items, person reliability and separation, and item reliability and separation.

Page 11: Final PPT_ Item Response Theory
Page 12: Final PPT_ Item Response Theory

IRTPerson reliability .76

Item reliability .97

Classical Test Theory

Cronbach’s Alpha .7546

Page 13: Final PPT_ Item Response Theory

The sample yielded a person reliability of .76

This implies that items are working well together to consistently reproduce a participant’s score

Page 14: Final PPT_ Item Response Theory

The sample produced a person separation statistics of 1.78.

The strata formula was used to determine the number of distinct strata ( HP=( 4GP + 1 )/3 ).

Thus it resulted to strata equalled to 2.70. The sample can be grouped and separated

into three distinct ability groups.

Page 15: Final PPT_ Item Response Theory

The item separation was 5.71 when computed using the stated above formula ( ( HP =( 4GP +1 )/3 ) resulted 7.94.

Findings suggested that the test items can be categorized into eight subgroups.

Page 16: Final PPT_ Item Response Theory

The item reliability of the said achievement is .97.

The SZTD resulted to .00 ZSTD less than 0 indicated greater predictability.

Page 17: Final PPT_ Item Response Theory

Mean SD

CTT difficulty

IRT difficulty

.282

0.137854

.122

.1209

Page 18: Final PPT_ Item Response Theory

There were 139 items ( 63.47 % ) considered as Average; 80 items ( 36.52 % ) difficult.

Majority of reasonably good items, marginal items and poor items were all from test 3.

Page 19: Final PPT_ Item Response Theory

Resulted to the following : of the 219 items, 9 items were poor items ( 4, 10% ) , 8 items were marginal ( 3.65 % ), 27 items ( 12.37 % ) were reasonably good items, 11 items ( 5 % ) were good items and 164 items ( 74.88 %) were very good items.

Page 20: Final PPT_ Item Response Theory

The nine poor items and difficult were as follows : items # 106, 185, 92, 157, 6, 72, 117, 129 and 147.

Page 21: Final PPT_ Item Response Theory

Reflects the matching of ability with difficult of item.

The mean of items was 1 logit below the sample.

The ability of the participants was higher than the all the difficult items,

Page 22: Final PPT_ Item Response Theory

There were two potential gaps ( located 106, 185, 72 and 92 );from 92 and 157; additionally from 157 and 197 and 6 )

The gaps are not significant considering that is less than 2.00 logits

Page 23: Final PPT_ Item Response Theory

Mean SD

CTT discrimination

IRT discrimination

0.564729

1.02386

0.245878

0.310584

Page 24: Final PPT_ Item Response Theory

The utilization of two frameworks in testing intensified and strengthened the stability of the achievement examination.

CTT and IRT yielded reliable results : .7546 and .97 respectively

Page 25: Final PPT_ Item Response Theory

IRT has person reliability which is one of the limitations of CTT.

Person reliability shows the consistency across participant’s score.

Generated three distinct strata. Test items can be categorized into eight

subgroups instead of 12 subjects instead of 12 subtopics.

Page 26: Final PPT_ Item Response Theory

The results of CTT and IRT are almost indistinct.

Majority of the difficult items identified in CTT are synonymous with IRT results

IRT can predict the probability of each student to answer such item correctly or incorrectly based on the logit. Hence , it provides individual assessment instead of group.