COMPARISON OF PSYCHOMETRIC PROPERTIES OF WEST AFRICAN EXAMINATIONS COUNCIL AND NATIONAL EXAMINATIONS COUNCIL TEST ITEMS UNDER ITEM RESPONSE THEORY

ABSTRACT

The purpose of this study was to investigate the psychometric properties (reliability, validity, difficulty, discrimination, guessing and differential item function) of the items of the Biology examinations conducted by the National Examination Council (NECO) and the West African Examination Council (WAEC) using the Latent Trait Theory (IRT).  This was against the backdrop of persistent public outcry that NECO questions were too cheap to pass.  14 Research questions and

7 hypotheses were formulated, tested, and analyzed.   The sample was made up of 1800 senior secondary year three students from 36 secondary schools in the urban and rural areas of Benue State.  The multistage stratified sampling technique was used.   The NECO and WAEC Biology examination questions from 2000–2002 were the instruments for data collection.   Maximum Likelihood estimation technique (using BILOG MG computer programme) was used to analyze the data in order to answer the research questions, according to IRT procedures.  The t-test was used to analyse the data in order to test the hypotheses.  It was found that the Biology examination items from the two examination bodies were equally reliable and valid.  Biology items in the NECO – conducted examination were more difficult than those of WAEC of same years.   Therefore, NECO questions were really not cheap to pass.   WAEC items were more prone to guessing than those of NECO, and boys performed better in

the WAEC items than in the NECO items.  It was recommended that IRT  procedures  should  be  adopted  by  all  examination  bodies  in Nigeria so that our measurement problems could be put to rest.

CHAPTER ONE

INTRODUCTION

Background to the Study

In Nigeria, today, there are three examination bodies charged with the responsibility of examining and awarding ordinary level certificates at the end of  senior  secondary  school,  namely  West  African  Examinations  Council (WAEC), the National Examinations Council (NECO) and National Business and Technical Education Board (NABTEB).

As reported by WAEC (2002), the origin of WAEC dates back to 1949 when the late Dr. G.G. Jeffery was invited by British Secretary of State for the colonies to visit West Africa to study and advise on a  proposal that there should be instituted a West African School Examination Council. In 1950, Dr. Jeffery submitted a report strongly supporting the proposal for a West African Examination   Council   and   making   detailed   recommendations   on   the composition  and  duties  of  the  council.  The  report  was  adopted  without reservation by the  four West African governments (Nigerian, Ghana, Sierra Leone  and  the  Gambia)  and  an ordinance  establishing  the  Council  as  a corporate body was drafted by the West African Inter-Territorial Secretariat in consultation with the governments.

In 1953, the Nigerian Government  made available a large block of

offices at the Technical Institute, Yaba which became the seat of the Deputy Registrar of the Council.   The member countries of WAEC  include Nigeria, Ghana, Sierra Leone and The Gambia.

The council conducts different categories of examinations:

i.       National  Examinations  -These  are  restricted  to  the  specific member countries for which they are developed and reflect their local policies, needs and aspirations.

ii.       International Examinations – these are developed for candidates

in all the member countries.   The West African Senior Schools Certificate Examination (WASSCE) is one of such examinations. It was introduced as part of the educational reform programmes in member countries.

The  Federal  Government  of  Nigeria  in  April,  1999  established  the National Examinations Council (NECO) with its headquarters in Minna, Niger State.   The aim of establishing  this council is for the nation to  have an independent/national examination outfit parallel but of the same standard as WAEC.   The National Examinations Council conducts  national  examinations such as:

i.       examinations into Unity Schools (federal government secondary colleges);

ii.       examinations into schools for the gifted children;

iii.      ordinary level school certificate examinations (facts about NECO,

2001).

As part of measures to uplift Technical and Business Education in the country,  the  Federal  Government  in  June  1992  established  the  National Business and Technical Examinations Board (NABTEB) with its headquarters in Benin City.   The board (NABTEB) examines similar  subjects handled by

WAEC and NECO and they award the National Technical Certificate (NTC) and

National Business Certificate (NBC) (NABTEB, 1997).

Test is one of the devices for measurement  in schools. In terms  of given  objectives,  testing  shows  the  achievement  of  students  (Ogunniyi,

1984). It consists of a set of uniform questions or tasks to which a student is to respond independently and the result of which can be treated in such a way as to provide a quantitative comparison of the performance in different students  (Nworgu, 1992). Testing  is a fundamental  part of  the teaching- learning process used not only as a basis for ranking students at the end of the  teaching-learning   process  but  to  guide   teaching,  and  aid  in  the development of curriculum, as well as in the assessment of needs, learning difficulties, level of mastery and  differences among students. Based on the level of performance  criterion,  there are three types of test which are the General  Mental  Ability  Tests,  Separate  Ability  Test  and Achievement  Test (Nkemakolam, 2003).

Achievement  tests  measure  the  present  proficiency,  mastery  and

understanding of general and specific areas of knowledge (Kerlinger, 1973). According  to  Nkemakolam  (2003),  achievement  tests  are   designed  to measure the outcome or level of accomplishment in a specified programme of instruction in a subject area or occupation, which a student had undertaken in the recent  past.   Achievement  tests may be  classified  into Teacher-made (classroom)  Tests  and  Standardized  Tests.  Teacher-made  tests  are  tests constructed and administered by the  classroom teacher for the purpose of measuring the achievement of pupils.  Standardized tests are more carefully

and accurately designed to cover many classes of a specified type in areas that are common to these classes.  It is normally constructed by a team of teachers in the field with specialists in test development.  On  the basis of response, achievement  test is classified into objective tests  and subjective (essay) test.

Essay  test  allows  students  to  express  themselves  freely  in  their answers  to  particular  questions.  In an objective  test,  however,  students’ responses are restricted to a number of symbols, words, phrases or simple sentences, one of which is considered to be the best answer out of several plausible alternatives (Ogunniyi, 1984). One of the strengths of objective test over the essay test is its amenability to quantitative item analysis, the result of which could be used to improve the item itself, specifically, and classroom learning, in general (Nenty, 1992).   Other advantages of the objective tests include  adequate  sampling  of the  content  to  be  assessed,  objectivity  of scoring, ease  and objectivity of item analysis and greater reliability (Aiken,

1987; Bennet & Ward, 1993).

Despite   being   convenient   to   use   and   having   some   desirable psychometric  properties, objective test items have been  criticized for their inability  to  assess  beyond  the  level  of  rote  memorization  (Ndalichako  & Rogers, 1997). Bennet & Ward (1993) argued that objective tests encourage teaching and learning of isolated facts and rote procedures at the expense of conceptual understanding and the development of problem-solving skills. This is right to a very large  extent. However, Maguuire, Hattie & Haig (1994) indicated that objectivee items may also measure higher levels of thinking.

The NECO and WAEC Biology examinations are made up of three parts, namely:  Paper  1  (practical),  Paper  2  (objectives)  and  Paper  3  (easy questions).  For  this  study,  only  the  objective  questions  were  used  and analyzed. This is because  psychometric  properties  of a test  are easier to obtain using objective test items.

The practical relevance of these tests is largely dependent on  their levels of reliability, validity, difficulty and discrimination.  All these add up to the psychometric properties of a test.    The development of achievement, ability,  aptitude,  interest  and  personality  tests  is  generally  a  multi-step process that can follow one of two distinct measurement frameworks. These are usually called the classical Test Theory (CTT) and Item Response Theory (IRT) measurement strategies (Macdonald & Paunanen, 2002).

The CTT is a measurement  framework used in the development of

achievement, aptitude, interest, ability, and personality tests.   Measurement in Education  and  Psychology  has, for a long  time,  based  its theory  and practice on the Classical Test Theory (CTT). Under the CTT framework, item analysis largely consists of calculating difficulty and discrimination indices for each item.    According to Thorndike (1977),  the  classical test theory views the score X that a student receives on a test as the sum of two observable components (True score and Error score). In Psychometrics, item response theory (IRT) is a body of theory describing the application of mathematical models to data from questionnaires and tests as a basis for measuring things such as abilities and attitudes.

IRT models are mathematical functions that specify the probability of a discrete outcome, such as a correct response to an item, in terms of person and item parameters, Person parameters may, for example,  represent the ability of a student or the strength of a person’s attitude. Item parameters include difficulty (location), discrimination (slope), and pseudoguessing (lower asymptote).  Items  may  be  questions  that   have   incorrect  and  correct responses, or statements on questionnaires that allow respondents to indicate level of agreement.

Among other things, as a body of theory, IRT provides a basis  for evaluating how well assessments work, in terms of being able to assess the individual’s ability and how well individual questions on assessments work. In education,  psychometricians  apply IRT in order to  achieve  tasks such  as developing and refining examination  items,  maintaining banks of items for exams, and equating for the difficulties of successive versions of exams (for example, to allow comparisons between results over time).

Item Response Theory (IRT) deals with the characteristics of the items

in terms of difficulty, discrimination and response pattern of examinees. Its major interest is to determine what a particular examinee might do  when confronted  with  a  test  item.  Baker  (2001)  explained  that  a  reasonable assumption is that each examinee responding to a test item possesses some amount of underlying ability. Thus, one can consider each examinee to have a numerical value, a score that places him or her  somewhere on the ability scale.  Responses  to  items  are  modeled  as  a  function  of  a  person’s performance level of the trait being measured and the characteristic of the

items completed. In other words, once an examinee’s ability level has been established, it is possible to determine the probability of a correct response to an item the examinee has never taken assuming that certain item parameters have already been determined (Lord, 1980; Wilcox 1987; Harris, 1989).

Nenty (1992) further explained that when an examinee encounters a test item, the intention is that he will respond to it with the best of his ability, that is, on the basis of how much of what the item measures he possesses. He added that during this person-by-item encounter, the person puts in all that he knows, his ability, in order to overcome the item. The ability of the examinee  in  what  is being  measured  and  the  difficulty  of the  item  are supposed to be the only determining factor in the person-by-item encounter. The difficulty of an item, according to Nenty (1992), could be viewed as the highest  resistance  it  could  withstand  before  it is overcome  or answered correctly. If an examinee’s ability is higher than that necessary to overcome the item, he gets the item right, but if the ability demanded by the item is higher than that possessed by the examinee, then he cannot overcome the item.  With this relationship,  examinee’s  ability  and  item difficulty  can  be measured  on  the  same  scale  with  the  same  unit  such  as  means  or percentages.    The  examinee’s  ability  is the  amount  of what  the item is measuring which is possessed by the examinee, while item difficulty is the amount of what the item is measuring just necessary to overcome the item. In a testing situation, the examinee should be able to answer all the items which demand ability less than that which he  possesses and should fail to

answer correctly all the items that demand ability higher than that which he possesses.

The West African Examinations Council and the National Examination

Council base the analysis of the psychometric  properties  of items on  the classical test theory frame work.  According to Korashy (1995); Smith (1996); as well as Ndalichako & Rogers (1997), the CTT is no longer valid for ensuring objectivity in measurement because it is seen as being too  restrictive and inadequate   in  scope  for  the  measurement   require  in   evaluating   the effectiveness of achievement test.

The CTT has some limitations that make its use in establishing the

psychometric  properties of tests questionable.   The major limitation is  the variant nature of the indices used to describe the item parameter (Douglas,

1990; Gruiton & Ironson, 1983).  This implies that item parameters (e.g item difficulty) vary from sample to sample of the same population of testees.  The proportion of examinees in a sample who get an item correct changes from a sample whose mean ability is high to one whose mean ability is low Nkopne (2001).  Weiss & Davison (1981) further explained that the same individual tested  in  two  different   samples  may  obtain   two  different  errors  of measurement and estimate of true score.  For this limitation, Douglas (1990) says that item parameters or statistics which remain invariant or stable from one sample of examinee to another have been viewed as a desideratum for any psychometric measurement.

All estimation procedures used to establish reliability co-efficient under

the CTT frame work are sample-based  in nature.   This contributes  to  its

weakness.  The reliability estimates are specifically a function of a particular set of items and sample of examinees  on which the data  were collected Nkopne (2001).  It means that in CTT the contribution of each item to the test reliability and validity depends upon what other items are in the test. The CTT provides an overall reliability index for all items on the entire test. To this Gruiton & Ironson (1983) say that it is unreasonable to assume that scores throughout an entire test will have  the same degree of precision of measurement or one overall index of dependability.

Scores obtained by an examinee in a test under the CTT are expressed as the total raw score the examinee obtained on the content areas concerned. It does not consider the patterns of response nor what a particular examinee might do when confronted with difficult test items.  It does not take care of individual differences among examinees.   These  limitations have made the CTT framework unfit for proper establishment of the psychometric properties of tests (examinations).  This calls for a  better framework which the Item Response Theory (IRT) offers. To avoid the limitations of the CIT examining bodies should adopt IRT in establishing the psychometric properties of tests. The IRT is invariant in nature.

Item  parameters  are  not  described  based  on  samples.    The  item parameters (e.g. item difficulty) do not vary from sample to sample as with the CTT.  Gulliksen, (1969) and Lord & Novick (1988) identified invariant item parameters  as  an  outstanding  measurement  issue  and  suggested  that whosoever elucidated it would have made a significant contribution to item analysis theory.  The invariant nature of IRT models makes the estimate of

the test item parameters and person’s ability independent of which subgroups the person belongs to or on the selection of the specific set of items provided the data fit the model (Wainer, Morgan & Gustafsson, 1980).

Another  major advantage  of IRT is that  it provides  a measure  of

precision of ability estimate at each ability level.  Thus instead of providing a single  standard  error  of  measurement  that  applies  to  all  examinees,  it provides a separate estimates of error for each  examinee and each item. Based  on  this  a  precise  and  confident  statement  can  be  made  than  a conventional one overall standard error of measurement for the total test as in CTT (Guiton & Ironson, 1983).

Some of the major advantage of IRT are stated here:

1.      In IRT a person’s score is based on a probability model for each item.   This makes the score item-based.   This is unlike the CTT  where a person’s score is obtained by summing across all items in the  scale which makes the score test-based.

2.      IRT models  consider  the pattern  of response  by examinees.

According to Baker (1977), the knowledge of the patterns of response  has implication for counseling both for students and teachers.

3.      The objectivity of IRT gives it an edge over the CTT because in IRT comparison of ability of examinees does not depend on which particular group of examinees used to validate the item parameters or on a specific set of items (Tinsley & Dawis, 1975; Douglas 1980).

The CTT model is simpler and easier to use than the IRT models.

Although the IRT models are more complex and more challenging to  learn

because they require some technical expertise to apply and requires the use of  specialized  soft  wares;  their  practical  advantages  make  their  use  a worthwhile venture.

The IRT models are three (3): the 1 – parameter model, 2 – parameter

model and 3 – parameter model.

The 1 – parameter model, also, known as the Rasch Model ascribes only the difficulty parameter of an item as the trait level required to correctly answer the question. Items that fit the model most differ only in difficulty.

The 2 – parameter model deals with the discrimination parameter of an item in addition to the items difficulty parameter.

The 3 – parameter model gives the probability of an individual  with ability responding correctly to an item  with a difficulty index, a discrimination index and a guessing index of.  The model assumes that the three parameters (difficulty, discrimination and guessing) are necessary  for an estimate of a valid relationship between the probability of a correct response of an item and the trait level (ability) of an individual. If the  public is made to know the advantages of IRT, there will be no doubts about the quality of the items in the examinations.

All examinations conducted in Nigeria have been based on the  CTT framework  for years. Examination  bodies have  tended  to rely on  CTT in testing their candidates. The public perception about the examination bodies varies from time to time. NECO has been worst hit by negative perceptions over the quality of its examinations.

A comparison of results from three secondary schools based on  the results of various subjects is presented in Appendix 1.  A general observation is that candidates passed more in all subjects, with the exception of just a few, in examinations conducted by NECO than those by WAEC. For example, in 2000, the percentage passes in NECO examinations were less in only three subjects in one school, one subject in another school and three subjects in the third school, as shown in Appendix 1. Put differently, the failure rate was higher in examinations (including Biology) conducted by WAEC than those by NECO  (details  in  Appendix  1).  Something  must  be  wrong  with  the  CTT framework being adopted at present in the country, going by public concerns.

In terms of gender, boys (Government College) performed better than girls  (Government  Girls College)  in 10 out of 12  subjects  in  the  WAEC- conducted  examination.     This  situation  is  similar  in  the  rural  schools (Appendix  2).    In  terms  of  school  location,  candidates  of  Government Secondary  School,  Ikpayongo  (a rural-based  school),  performed  better  in most subjects  than their counterparts  in urban  schools  (Government  Day Secondary School, Makurdi an urban centre), as shown in Appendix 2.  Both schools are mixed (boys and girls).

Statement of the Problem

Some people (students and parents) believe that if a student wants to pass and make his ordinary level examinations at one sitting, he should write the   NECO   examinations   because   candidates   who   write   the   NECO examinations always passed most of their papers, for the reason that NECO examinations are cheap to pass (Newswatch, 2004, P.9).   The  public is of

opinion that WAEC examinations are much more difficult to pass than  the NECO examinations.   Alarm had been raised about “the high rate of failure in all public examinations”, particularly that of WAEC (The Guardian, 2002).  An evidence of this is the record that as far back as  June, 1994, out of the

524,294 candidates who sat for English Language, only 74.157 or 14.10 per cent had the effective grades of 1 to 6.  A whopping 290,237 or 53.30 per cent had F9, and an additional 159,900 or 30.50 per cent made P7 or P8 in a subject  where  a  credit  pass  is  required  for  admission  into  any  tertiary institution (The Guardian, 2000).  That same year, 518,188 candidates sat for Mathematics and only 83,192 had the effective grades of 1 to 6; 219,818 had the F9 grade, while 215,108 had P7 or P8.  This situation of high failure rate was  attributed  to  the fact that  the  reading  culture  among  students  had drastically reduced over the years.

Similarly, in the 2000 examinations for Biology, the percentage failure was higher with WAEC examination than that of NECO in the schools sampled (see App. 1). In terms of gender the failure rate was higher among girls than the boys in the schools sampled. In terms of location, there was no significant difference in performance of students in urban and rural schools (App. 2).

The pioneer  Registrar/Chief  Executive  of NECO, Professor  Ojerinde, agrees that there are several examples of those who pass  NECO and fail WAEC examinations and vice versa (Newswatch, 2004, p. 9).  He agrees that some people say that “NECO is cheap to pass but that NECO papers are not cheap”.   For this reason, some tertiary institutions  tended to discriminate against NECO result by denying candidates  placements during admissions.

Similarly,  there  were  allegations  of  massive  leakages  in   examinations conducted  by  NECO.  To  this,  NECO  (2001)  suggested  that  examination malpractices  were perpetrated  with the connivance,  active support or the neglect of some unscrupulous supervisors and invigilators.   Both the WAEC and NECO examinations are based on the CTT framework in terms of analysis of their psychometric characteristics.

Obviously, the CTT has some serious disadvantages that make its use

in  developing  and  establishing  the  psychometric  properties  of  classroom achievement tests improper.  For instance, the CTT model typically provides only one overall index of reliability for all items in a test.  The CTT model itself does  not  ensure  objectivity  in  measurement  due  to  the  fact  that  one technique of measuring a particular attribute is not interchangeable with an alternative measure of the same attribute.   CTT fails to provide solutions to test score equating, test design and  testing items for bias.   Generally, the latent trait theory, unlike CTT  provides invariant item parameters, equates different forms of the same test and presents test items tailored to individual test examinees. It tests items for bias and precisely determines test item fit. Accompanied  by its  standard error of measurement.  Thus, the IRT model takes into  account  some qualitative information regarding the psychometric properties  of  each  test  item.  Could  the  above  reasons  account  for  the observed   differences   in   the   performance   of   candidates   who   write examinations  conducted by both WAEC and NECO?   Could the  latent trait theory (item response theory) offer a way out?

If  the  above  observations  are  true,  then  IRT  should  be  tried  in developing  and establishing  the psychometric  properties  of tests. In  both WAEC  and  NECO  tests,  what  are  the  relative  psychometric  valules  for candidates in urban and rural areas and between sexes (boys and girls), a check on this is, therefore, needed. What are the  sources of performance differences by candidates who write examinations conducted by WAEC and NECO?   Are the tests in one  examination more valid and reliable than the other or more reliable  and  not valid? Having used/tried  CTT through  the years, can one confidently continue to accept and depend on the model for our future examinations?  Can IRT offer acceptable and objective alternative to  testing  in Nigeria?    When  applied  to  tests  conducted  by examination bodies, would IRT give a better index of ability to students? Differences in performance of candidates continue to be observed?  The application of IRT models  to  the  said  examinations  would  further  confirm  the  levels  of standardization of the tests.   In other words, analyzing the tests using IRT framework   would   establish   the   nature   of   differences   between   the examinations in relation to specified  psychometric properties.   One way of investigating the validity of these claims is to find out how far the WAEC and NECO  examinations  have  kept  up  high  testing  standards  with  particular reference to psychometric characteristics.  Are the NECO test items of lower quality than those of the WAEC?

Purpose of the Study

The primary purpose of this study is to determine the relative strength of  the  psychometric  characteristics  of  the  test  items  used  in  Biology

examinations  conducted by WAEC and NECO using IRT.    Specifically,  the study determined the:

1.      reliability   of  the  test  items  in  the  Biology  examinations

conducted  by NECO and WAEC,  using  the standard error  of measurement.

2.      validity of the test items in the Biology examinations conducted by NECO and WAEC; using the fit statistic technique.

3.      difficulty parameter of the test items in the Biology examinations conducted by NECO and WAEC;

4.      discrimination   parameter   of   the   tests   items   in   Biology

examinations conducted by NECO and WAEC;

5.      differential    item   functioning    (DIF)   of   the   test    items administrated by WAEC and NECO; in terms of gender.

6.      differential item functioning of the test items administered by

WAEC and NECO with reference to school location;

7.      levels  of  guessing  parameter  of  the  test  items  in  Biology examinations conducted by WAEC and NECO.

Significance of the Study

The  study  will  provide  input  into  test  construction  skill  of  test developers/examining bodies in determining the existence or non-existence of item differential functioning (DIF). Detecting item differential functioning or item bias as a problem in examinations will only be reasonably possible using the latent trait model because of its sample-invariant property.

The results from this study would encourage test developers to  take pains to conduct rigorous item analysis before and after test administrations. The  latent  trait  models  ensure  the  equating  of  test  scores  using  ability estimates and, thus, are useful in controlling the problem of non-equivalent groups associated with CTT. The understanding of the practical approach and procedure for estimating examinee ability even when some items might not have been responded to (that is total scores determined from partial scores) would solve the problem of using graded scoring model.

The   study   will   be   significant   to  classroom   teachers,   guidance counselors  and  educational  institutions  by  providing  clues  to  meaningful interpretation of examinees result through person-by-item encounter (latent trait model) during examinations.

The study will, also, serve as an efficient tool for diagnostic evaluation

by the  teachers,  guidance  counselors  and  educational  institutions,  which involves  not only the identification  of errors  but the  identification  of the misconceptions that underlie such errors. It will ensure improvement in the teacher’s  teaching  behaviour  in terms  of  coverage,  emphasis  of curricula content  and  instructional  practices.   The  strengths  and  weaknesses  of candidates   in  examinations   will   be  readily   known   because   guidance counselors would be amply armed with the necessary information/data about each examinee’s performance item by item.

The results of this study would help to indicate the overall quality of the examinations conducted in Biology by WAEC and NECO. A confirmation of the reliability and validity of the examinations conducted by WAEC and NECO

would help to establish public confidence and acceptability of results  from their examinations. The public needs to be convinced that the examinations conducted by both WAEC and NECO are of relatively equal standards and that no one is of inferior quality. Thus, the public will  be enlightened  on the interpretation of students’ results from the examinations conducted by these two examining bodies. They will not just assume high quality of either of the examining   bodies.     Presently,   the   performances   of  students   in  the examinations conducted by these  two bodies are interpreted based on the sum of their total scores which is typical of CTT. The use of these scores just summed across all items to consider the performance of examinees hides the characteristics of both the examinee and the test. For objective and adequate decisions to be  taken on the performance  of students in examinations  by WAEC and  NECO, the psychometrics  of the tests need to be determined. Examining bodies need to consider the psychometric properties of tests  in taking decisions on the observable performance  of candidates in  order to improve upon test construction, administration and analysis. In addition, from this study, WAEC and NECO would have a clearer  understanding  of their performance in test construction and be appropriately guided from now on as they, hopefully, accept and adopt IRT evaluation framework.

Scope of the Study

The study covered all the secondary schools in the three  education zones of Benue State, Nigeria. The study is limited to the  May/June WAEC (2000 – 2002) Biology examinations and the August/September (2000) and (June/July, 2001 – 2002) NECO Biology  examinations.  This is because the

maiden  examination  conducted  by  NECO  was  in  2000,  and  so   many candidates  were said to have  passed the examinations  with  good grades especially those who could not pass the WAEC-conducted Biology examination in that same year. The 2000 May/June WAEC Biology examination was the second year of WAEC administering the West African Senior School Certificate Examination (WASSCE).  This study covers One, Two and Three – parameter models. As required in all IRT models, a relatively large sample size was used in order to ensure very high reliability of the findings.

Research Questions

The study was carried out through answering the following  research questions:

1.      What are the standard errors of measurement of the test items in the

Biology examinations conducted by NECO?

2.      What are the standard errors of measurement of the test items in the

Biology examinations conducted by WAEC?

3.      How valid are the test items of the Biology examinations conducted by

NECO?

4.      How valid are the test items of the Biology examinations conducted by

WAEC?

5.      What are the difficulty parameters of the test items in the  Biology examinations conducted by NECO?

6.      What are the difficulty parameters of the test items in the  Biology examinations conducted by WAEC?

7.      What are the discrimination parameters of the test items in the Biology examinations conducted by NECO?

8.      What are the discrimination parameters of the test items in the Biology

examinations conducted by WAEC?

9.      What are the guessing parameters of the test items in the  Biology examinations conducted by NECO?

10.     What are the guessing parameters of the test items in the Biology

examinations conducted by WAEC?

11.     How do the test items of the Biology examinations conducted by NECO

function with respect to sex (boys and girls)?

12.     How do the test items of the Biology examinations conducted by WAEC

function with respect to sex (boys and girls)?

13.     How do the test items of the Biology examinations conducted by NECO

function with respect to location (urban and rural)?

14.     How do the test items of the Biology examinations conducted by WAEC

function with respect to location (urban and rural)?

Hypotheses

1.      There  is  no  significant  difference  in  the  mean  standard  errors  of measurement of the items in the Biology examinations conducted by NECO and WAEC from 2000 – 2002 based on One, Two and Three – parameter models of IRT.

2.      There is no significant difference in the validity (fit statistic) of items of the Biology examinations conducted by NECO and WAEC from 2000-

2002 based on the One, Two  and Three – parameter models of IRT.

3.      There is no significant difference in the difficulty parameters of items in the NECO and WAEC conducted Biology examinations from 2000-2002 based on the One parameter model of IRT.

4.      There is no significant difference in the discrimination parameters of

items in the Biology examinations conducted by NECO and WAEC from

2000-2002 based on Two – parameter model of IRT.

5.      There is no significant difference in the guessing parameters of items of the Biology examinations conducted by NECO and WAEC from 2000-

2002 based on the Three-parameter model of IRT.

6.      The test items of the Biology examinations conducted by NECO and

WAEC do not differ in function among examinees with respect to sex.

7.      The test items of the Biology examinations conducted by NECO  and WAEC  do  not differ  in function  among  examinees  with  respect  to location.

₦5,000.00 – DOWNLOAD FULL PROJECT DOWNLOAD FULL PROJECT Added to cart

---

---

---

---

---

Related Project Topics :