Types of Adjunct Question Studies Summary

Levels of Adjunct Quetion Studies    Summary

Interaction between T&L

General Conclusion of the Research Studies

Adjunct Questions have received a great interest of educators and instructors since they are considered an effective teaching instrument, hence, an effective evaluation device (e.g., Al-Smadi 1992; Bridgeman and Lewis, 1994; Darwazeh, 1997a, 1997b; Harder, 1991; Kneip & Grossman 1979; Zeidner 1990; etc.).

Accordingly, adjunct questions, as a tool of assessing students’ levels of learning, has been tackled by many specialists, educators, and researchers . 

There have been a number of studies designed to compare between multiple-choice and essay questions and to examine their effects on students’ achievement. In this section, it would be beneficial to summarize the results of some previous studies that have been conducted on adjunct questions in terms of their types, levels, and their interaction with student’s ability.

The author of this research has reviewed these studies and classified them into three main headings: (1) type of adjunct questions, (2 ) level of adjunct questions, and (3) the interaction among question types, levels and students’ ability.

Types of Adjunct Question Studies TOP
= = = = = = = = = = = = = = == = = = = = = = = 

The question formats that are mostly used in the construction of achievement tests at both school and college levels could be classified into two main categories: (a) objective type format such as multiple-choice questions, and (b) essay type format such as short - answer questions ( Darwazeh, 1997, p. 60 ; Gronlund and Linn, 1990, pp. 166-200 ). 

Some educators said that an essay type test is a better measurement tool of learners’ achieve- ment than an objective one, because: 1) it requires the learner to produce the answer not to choose it, 2) it can measure different levels of students’ learning , 3) it is relatively easier to prepare than multiple-choice type, and 4) it is the only means that we can use to assess the high levels of learning such as application, analysis, synthesis, problem solving, and evaluation (See, Ebel & Frisble 1979, p.100; Hambleton & Murphy, 1992). Other educators assume that the multiple-choice type test is more popular because it: 1) requires the learner to discriminate among the alternatives easily and choose the right answer, 2) removes subjectivity in scoring answers, and 3) reduces measurment error and elaminates random guessing by implementing a certain formula (Bridgeman & Rock 1993; Hambleton & Murphy,1992; Hopkins & Stanely, 1981, pp.231-234; Mehran & Lehman, 1978, pp. 208-300; Kubiszyn & Borich, 1990, pp. 84 - 114; etc.). 

Bridgman and Lewis (1994); Darwazeh (1997a) have also differentiated between essay and multiple-choice test items; they stated that essay questions assess productive and organized skills that can’t be measured by multiple-choice questions. Besides, the two types differ in their coverage of the content domain; essay questions usually cover a smaller range of learned content, while multiple-choice questions cover a broader range of learned content. 

There has been a number of studies which were conducted to compare between multiple-choice and essay questions to investigate their effects on students’ achievement. For example, Roderick & Anderson (1968) compared between the effect of essay versus multiple-choice questions on students’ achievement. Students were asked to read a 3400 word passage on "Classical Conditions" and to answer the questions that follow the passage. Students were distributed into two experimental groups: one received essay post questions and the other received multiple-choice post ones. Results showed that students who received post essay questions performed better than those who received multiple-choice post ones on the post test which measured the retention learning (See, Darwazeh,1997a, p.149).

Arrasmith, Sheehan, and Applebaum ( 1984 ) supported Roderick & Anderson’s results when they compared between the selected - response strategy ( multiple-choice ) and the constructed - response strategy ( essay ) for assessing a third - grade writing task. The performance of 371 third - grade students were measured by selected - response test items ( multiple-choice) and constructed - response test ones (essay ) to assess the mastery of the written composition. Results indicated that essay items were better for mastering the writing composition than multiple-choice ones. 

Foos and Fisher (1988) found the similar results when they conducted a study to assess the value of test taking as a means of enhancing rather than monitoring learning. 105 college students read a short essay about the "American Civil War" and were then given either a postpassage questions or no quetions at all on the text material. The form of the postpassage questions was either fill-in (essay) or multiple-choice; and the knowledge examined was either directly stated in the original text (verbatim) or could be logically derived (inferential) from the text. A final test, containing the above mentioned levels of learning, was given to all students two days later. Results of the final test indicated that students who received inferential (UG) fill-in (essay) questions performed better than those who received verbatim (RI, RG) multiple-choice questions. 

Hambleton and Murphy (1992) found the same results as those found by Foos and Fisher when they investigated whether the multiple-choice question type has a positive effect on students’ achievement and if they can measure high order cognitive skills copmared with the authentic measurement (essay). The researchers collected data from several articles on essay and multiple-choice items to assess the effectieveness of the multiple-choice question over the essay one or vice versa. After comparing the data, they found that the authentic measurement items (essay questions ) have a little effect over and above that provided by the multiple-choice items on students’ achievement when assessing higher order levels.

Lukhele, Thissen, and Wainer ( 1994 ) agreed with Hambleton and Murphy’s results when they investigated the relative value of multiple-choice, and constructed response (essay) items on two achievement tests in chemistry and United States History. The researchers analyzed all of the data from the operational administration of the 1989 Advanced Placement Test in Chemistry as well as from the 1988 administration of the Advanced Placement Test in United States History. Analyses showed that the constructed response items (essay) of the tests yielded little information over and above that provided by the multiple-choice items written on a retention level of learning.

On the other hand, some researchers found that objective questions (e.g., multiple-choice) have a greater effect than essay questions on students ‘learning. For example, Duchastel & Nungester ( 1982 ) conducted a study to investigate the effect of two question types (essay versus multiple-choice) on student’s achievement .125 high school students in Grade 10 from a midle-class suburban high school were distributed randomly into three groups: one group received essay postpassage questions written on remember level; another group received multiple-choice postpassage questions written on the same level; and the last group is the control group who received a true or false questionnaire concerned with general study habits. All students were asked to study a brief history text with the post questions mentioned above. On the delayed post general retention test, consisting of 24 items selected from the initial post questions, twelve items were selected from the initial short answer (essay ) questions, and twelve from the initial multiple-choice questions. Thus, for students in each of the initial questions, half of the items was repeated in the same format as seen previously, while the other half was represented in the alternate format. The students were told that the test would require them to write down the main points they could remember, and that there would be specific questions of details as well. Results showed that students who received multiple-choice questions performed better than those who received essay questions and than those who were in the control group, on a post test that measured retention learning.

Perkins (1984) reached similar result as Duchastel & Nungester (1982) when he analyzed two common types of questions in testing English as a Foreign Language (EFL) reading comprehension. These tests were: 1) objective question types (true-false, multiple-choice), and essay ones (missing letters, and grammar paraphrase). The subjects were 19 Egyptian adults from the American University in Cairo, who were enrolled in an intermediate-level English as a Foreign Language. Subjects were distributed randomly into four groups: the first group received true-false posttest questions, the second group received multiple-choice posttest questions, the third group received missing letters posttest questions, and the fourth group received grammar paraphrased posttest questions. All subjects were asked to read a 232 word reading passage selected from their required text. Results of the posttest showed that objective items produced better results than did essay ones on a comprehension level of learning.

Bridgman and Lewis (1994) agreed with Perkins’ results when they investigated the relationship between essay and multiple-choice questions. 32 college students from public and private institutions, who entered the college in the fall of 1985, were tested in four subjects: American History, European History, English Language and Composition, and Biology by using two types of questions: essay questions and multiple-choice ones. Results indicated that multiple-choice scores of the American History and Biology examinations were superior to the essay ones, but essay scores were essentially equivalent to multiple-choice scores in European History, and English Language. 

Some researchers didn’t find significant differences between essay questions and objective ones on students’ learning. For example, Frase (1968); Williams (1963) could not find significant differences between the effect of short answer (essay) questions and the effect of multiple- choice ones on the student achievement ( See , Darwazeh , 1987, p.111, for more details).

Brigdman ( 1992 ) did not also find a significant difference when he compared between open-ended (essay ) and multiple-choice formats. In a separable data collection, 364 paid volunteers who had taken the Graduate Record Examination used a computer keyboard to enter answers to the same set of questions. Results indicated that the total scores for the multiple-choice and open-tests (essay ) demonstrated remarkably similar effects on achievement.In other words, there were no significant differences between the effect of question types (essay versus multiple-choice ) on students’ learning on retention type test. 

Bridgeman and Rock (1993 ) reinforced Brigdeman’s findings when they conducted a study to explore the relationship between multiple-choice and open ended (essay) analytical questions. The scores of 349 students of the Graduate Record Examination General Test were analyzed. Results showed that the open ended (essay) items were not measuring anything beyond what is measured by the multiple-choice version of these items which means that the two question types ( essay versus multiple-choice ) have the similar effect on the student’s achievement . 

Thissen, Wainner, and Wang ( 1994 ) agreed with Bridgeman and Rock’s results when they didn’t find a significant difference between multiple-choice and essay questions on the student achievement. They used 2000 students who took Computer Science and Chemistry tests of the College Board’s Advanced Placement Program, and divided them into two groups: one group received multiple-choice questions, and the other one received essay questions. Results showed that essay sections have the same effect on students’ achievement as the multiple-choice on solving problem (application level) test.

SummaryTOP
= = = = = = = = =
From the above studies, the researcher concluded that some researchers (e.g., Arrasmith, Sheehan & Applebaum, 1984; Hambleton and Murphy, 1992; Lukhele, Thissen, and Wainer, 1994; and Roderick & Anderson, 1968 ) found that essay type questions have more effect on students’ learning than do multiple-choice ones especially on high levels of learning (UG). Others, (e.g., Bridgman & Lewis, 1994; Duchastel & Nungester, 1982; and Perkins, 1984; etc.) found that multiple-choice type questions have more effect than essay ones especially on remember levels of learning (RI , RG). But some researchers found that there were no diferrences between the effect of essay versus multiple-choice question types on all levels of learning (e.g., Bridgman, 1992; Bridgman & Rock, 1993; Frase, 1968; Thissen, Wainner & Wang , 1994; Williams, 1963).

Levels of Adjunct Quetion Studies TOP
= = = = = = = = = = = = = == = = = = = = = =
Adjunct questions are usually designed to measure the variation of different cognitive levels. Thus, they vary from simple level such as Remember an Instance (RI) and Remember a Generality (RG) to complex level such as Use a Generality (UG) and Find a Generality (FG) according to the cognitive process the learner employs during answering the required question (Merrill ,1983 ). Low- level questions like remember - an - instance and remember - a - generality have been consistently used in school and teachers’ textbooks, but high level questions like apply and find a generality levels have seldom been used. Thus, adjunct questions which require recall of specific information or facts (RI,RG) may produce a different level of learning from adjunct questions which require students to apply (UG) or transfer the learned idea to new situations (FG). Simply, low levels of adjunct questions induce low levels of learning and high levels of adjunct questions induce high levels of learning ( Darwazeh, 1992, 1995; Martin, 1979).

There has been a number of studies which were conducted to investigate the levels of adjunct questions which teachers use in evaluating their students’ achievements. In the measurement of reading students’ achievement, Hoeppel ( 1981 ) conducted a study aiming at categorizing questions found in reading skills development books used in Maryland Community College Development Remedial Programs via the " Bloom Taxonomy of Educational Objectives Domain." A random sample of (555) questions was selected for the analysis. The questions were classified according to Bloom’s Taxonomy. The classification showed that 145 questions (26%) were for knowledge, 408 questions (74% ) were for comprehension, 2 questions ( 0.0035%) were for application and no questions were for analysis, synthesis and evaluation. In other words, results showed that (99%) of the questions were categorized within two levels: knowledge and comprehension. 

Al-Makzoomy (1986) agreed with Hoeppel’s results when he analyzed the Jordanian secondary- school teachers’ responses through a questionnaire, on the levels of questions used in teaching reading comprehension. Results showed that (68%) of the teachers usually place more emphasis on literal-type questions ( remembrance ) than on inferential-type questions (application). 

Rinser (1987 ) supported the results of the above two studies (Hoeppel,1981; Al-Makzoomy, 1986) when he studied the cognitive levels of questions demonstrated by test items that accompanied fifth grade science textbooks. Based on Bloom’s taxonomy, results showed that about (95%) of the test questions devoted to knowledge or comprehension ( RI, RG ), but (5%) were used for application ( UG ), and (0.2%) for evaluation, analysis, and synthesis questions were neglected completely (p.126).

Tollefson ( 1989 ) summarized the results of some studies which investigated the relationship between teachers’ questions and students’ learning. These results were:
1. The cognitive complexity of the students’ responses was largely determined by the cognitive complexity of the questions.
2. Recall and recognition (RI, RG) questions, which have a low level of cognitive processing, generally require simpler responses than higher level (UG) questions requiring students to express opinions or to provide interpretations (pp. 6-7).

Harder (1991) agreed with "Al-Makhzoomy’s" results when he studied the levels of classroom oral questions which were used by the teachers of Arabic language at the basic educational stage in Jordan. A sample of 60 teachers were given 45 minutes to ask their students different levels of questions. Results, which were classified according to Bloom taxonomy, showed that the percentage of questions at the level of knowledge was 47.5% of the total questions, 32.4% for the comprehension questions, 13.1% for application questions, 5.6% for analytic questions, 1% for synthetic questions, and finally, .7% for evaluation questions.

Perry, VanderStoep, and Yu (1993) conducted a study to investigate the levels of questions that were asked in first grade addition and subtraction lessons in Japan, Taiwan, and the United States. Some researches have argued that knowledge is, in part, constructed through questions and these may be used differently in U.S. than in Asian classrooms. Thus, each question of addition or subtraction in 311 observed lessons was coded as 1 of 6 levels of questions. Analyses revealed that the Asian teachers asked significantly more questions about conceptual knowledge than did U.S. teachers. In addition, Chinese teachers asked significantly more questions that were embedded in a concrete context (verbatim) than did U.S. teachers . 

Martin, et al. ( 1994) reinforced Harder’s (1991) results. Martin and her friends stated that teachers usually use questions which require factual answers and low levels of thinking (i. e., knowledge and comprehension questions) make up at least 70% of the questions, while questions that require application, analysis, synthesis, or evaluation thinking are used much less often (p.156).

Watts and Anderson (1971) investigated the effect of questions that require students to apply (UG) what they have read to new situations. 300 high school seniors answered an inserted post questions after reading a passage that consisted of a 450-word passage explaining psychological principles. Subjects were distributed randomly into three groups: one group received application questions (UG), another group received name questions ( RI, RG), and the control group was asked to read the passage only. Results showed that subjects who received questions requiring them to apply (UG) principles to new examples performed better on the posttest (which measured application questions) than did subjects who received questions requiring them to recall (remember) previous examples. 

Rickards & Vesta (1974) found similar results to Watts and Anderson’s results when they studied the effect of question levels on learning. The subjects were 80 college sophomores enrolled in an introductory educational psychology course. Each subject responded to one of the four types of questions: rote - learning - of - facts, rote - learning - of- ideas (RI, RG ), meaningful - learning , or task - irrelevant questions (UG). Results showed that students who received questions on meaningful learning (UG), which required subsumption or organization of facts under given ideas, performed better than students who received questions requiring them to recall specific facts or ideas (RI or RG ) - rote learning of facts and ideas - on a posttest containing these levels of learning .

Kneip & Grossman ( 1979 ) summarized some studies which reported that students’ achievement was significantly and positively affected when teachers use mostly high level questions. For example, Rayan ( 1973 ) compared between the effects of high and low level of questions on the social studies achievement of fifth and sixth grade students. Results indicated that questions which demand high cognitive levels (UG ), beyond the recall level, were superior to the low level (RI, RG) questions in producing not only the high level understanding but also producing high levels of achievement.

Redfield (1981)conducted a study to examine the effect of teachers’ questions on student achievement. In this study, twenty studies on teachers’ use of higher (UG) and lower (RI, RG) cognitive questions were reviewed . Higher cognitive questions require the student to manipulate information to create and supply a response; lower cognitive (RI, RG) questions call for verbatim recall or recognition of factual information. Results of the studies reviewed showed that teachers’ use of higher cognitive (UG) questions had a positive effect on student general achievement on a retention level of learning. 

Royer and Konold ( 1984 ) examined Hunkin’s (1969 ) study in which he investigated the effect of two levels of questions, knowledge (low - level) and evaluative (high - level) on students’ achievement in two groups. One group studied social materials provided with knowledge questions, the other group studied the same passage but was provided with evaluative questions. After four weeks, the whole students took an examination consisting of questions of all six levels of Bloom’s Taxonomy. Results showed that the two groups did not differ on items from the lower taxonomic levels, but they differ on evaluative items. That is, students receiving higher level evaluation during the study phase (UG) performed significantly better on high level questions in the posttest . 

Al-Nayef (1989) supported Hunkin’s results when he conducted a study to investigate the effect of adjunct question levels on reading comprehension of the eleventh graders. Al-Nayef used two levels of questions: low level questions (knowledge), and high level ones ( i.e., comprehension, application, analysis, synthesis, and evaluation). The sample of his study consisted of 144 eleventh grade scientific students in Al-Koureh District Education Directorate in Jordan for the second semester of the academic year 1988 /89. Subjects were randomly assigned to three groups: High level experimental group ( N = 82 ), low level experimental group (N =35), and control group ( N = 27 ). Results showed that students who were exposed to high level questions (comprehension, analysis, synthesis, evaluation) performed better on the same levels of questions in the post test.

On the other hand, some researchers found that low level questions have a great effect on students’ achievement. For example, Felker and Dapra (1975) have investigated the effect of different levels of adjunct questions on students’ learning. 93 introductory psychology students at the Greensburg campus of the University of Pittsburgh were distributed randomly into three groups: one group received verbatim questions ( RI , RG), another group received application level questions (UG), and the third group was the control group which received no adjunct questions during instruction but were asked to read the passage only. Results showed that subjects who received comprehension post-questions requiring them to recall (remember) the text performed significantly better than those who received application - level (UG) adjunct questions requiring them to identify new examples of learned concepts or principles on application level posttest, and better than the control group which received no adjunct questions during instruction. 

Samson et al. (1987 ) conducted an analytical study to investigate the result of Redfield & Rousseau’s ( 1981 ) study which showed that there was a large effect of higher order (UG)questions on student achievement. The main aim of Samson et al.(1987) study was to test the effects of question levels on student achievement and to examine Redfield & Rousseau’s results. They analyzed fourteen studies which were conducted to examine the effects of different adjunct question levels on learning. The results of the fourteen studies were examined and compared with the results of Redfield & Rousseau. Results of Samson’s study indicated that higher level questions (UG) have a small positive effect on learning. This result contradicts the previous results which indicate the opposite. 

Perkins et al. (1990) agreed with Samson’s analysis when they investigated the effect of question levels on English as a Second Language (ESL) reading comprehension. A sample of 150 Japanese English-as-a- second- language students at Southern Illiois University, Carbondale was given a reading comprehension test containing three levels of questions: factual (RI), generalization (RG), and inference (UG), to measure comprehension effects at different proficiency strata . Results showed that there were significant differences for the factual questions (RI), but no significant differences among the generalization (RG), and inference (UG) levels of questions.

Some researchers did not find a significant main effect with respect to the effect of adjunct question levels on students’ achievement .For example, Andre et al. (1980 ) have done seven experiments to explore the effect of application (UG) question levels and factual ones ( RI, RG ) on students academic achievement. The subjects were (120, 155, 87, 57, 135, 194, 194) male and female students attending Ames High School in Ames, Iowa; they were asked to read a short passage of three concepts and answer two levels of questions: remembrance ( RI, RG ) and application ( UG ). Results showed that five studies found no significant differences between question levels (remembrance and application). Only two studies found that remember - levels (RI , RG ) adjunct questions were superior to application (UG) ones. 

Darwazeh & Reigeluth (1982a) have also studied the level and position of adjunct questions and their effects on memory and application learning. Seventy- four male and female eighth grade students, at a suburban school in Syracuse, New York, were presented with a 450 - word passage from a social studies textbook. Nine short - answer posttest questions were written on three levels of learning (RI, RG, and UG), and were used either before or after the relevant passage. On the nine - short answer posttest questions, which were comprised of these three levels of learning ( RI, RG, and UG), they found that there were no significant main effects with respect to the question levels. 

Darwazeh (1982b) didn’t also find any significant difference among the performance of the experimental groups who received either remember - an- instance (RI), remember a generality (RG) questions and the performance of the experimental ones who received use a generality questions (UG), though the direction of the groups’ mean indicated that RI level had higher (x= .48) effect on students’ performance than RG (x= .41) and UG (x= .39).
 

Summary TOP
= = = = = = =
From the above revision of previous studies, we can conclude that some researchers found that high level questions ( i. e., use a generality) have a greater effect on students’ achievement than low level ones (e.g., Al-Nayef, 1989; Kneip & Grossman, 1979; Redfield, 1981; Rickards & Vesta, 1974; Royer & Konold, 1984; Watts & Anderson, 1971). Others found that low level questions have a greater effect on students’ achievement than high level ones (e.g., Felker & Dapra, 1975; Perkins et al.,1990; Samson et al., 1987). Few studies did not find significant differences among the different levels of questions: remember an instance, remember a generality, or use a generality on later learning (e.g., Andre et al. 1980; Darwazeh & Reigluth 1982a; Darwazeh, 1982b).

Interaction between types and levels of adjunct questions = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = TOP
Researchers and educators have postulated that using the two types of different questions ( essay versus multiple-choice ) will lead to different levels of learning outcomes, and different levels of questions (remember, comprehension, application, etc.).

The depth of cognitive processing was influenced by the type of questions on the one hand and the level of questions on the other hand .
Some educators claim that multiple-choice questions usually induce low levels of learning such as RI and RG, whereas, the essay questions usually induce high levels of learning such as UG and FG (Hancock, 1994). 

Anderson & Biddle (1975) expected that essay questions could improve students’ achievement if they are written on high levels of learning (use a generality and/or find a generality), whereas, multiple-choice questions can improve students’ achievement if they are written on low levels of learning i.e., remember an instance and/or remember a generality levels. 

Darwazeh (1987, 1997) stated that essay questions have greater effect on students’ achieve- ment than multiple-choice ones especially when they are written on high levels such as evaluation, synthesis, solving problems, etc. (p. 79).

Al-Smadi (1992) indicated that the use of question type ( essay or multiple-choice) of low - cognitive levels ( RI, RG ) does not help students learn on high levels of learning because these levels of questions generally elicit short responses and require no deep thinking .In other words, they don’t require students to comprehend, interpret, or use the information in new situations. Whereas, the use of question type (essay or multiple-choice) of high cognitive levels ( UG ) plays an important role in helping students to understand the text because they generally elicit long complex, and deep responses; they require students to comprehend, interpret, and use the information properly (pp.12-13 ).

Hopkins & Stanely (1981) summarized the results of some studies (e.g. Hoffman, 1962; Lafave, 1966) which showed that objective tests (multiple - choice) can measure mostly knowledge of facts, whereas essay tests can measure more complex, higher levels of under- standing and have a great effect on learning (See, Hopkins & Stanely, 1981, p. 202 ).

Concerning the interaction among question levels and students’ ability, Sanders (1973) found significant interactions which indicated that low ability students performed better on remember questions; whereas, high ability students performed better on application level. Sanders used 72 college students, and 40 item multiple-choice immediate and delayed test measuring relevant and irrelevant information (See, Darwazeh,1996, for more details). Memory (1983) supported the above result. He used a posttest containing 29 questions written on literal (remember ) and application levels. Results indicated that there was a significant interaction which indicated that high ability students performed better on the post test when they are written on high levels (application).

Darwazeh (1982) also reached similar results, when she investigated the interaction between students’ ability and the position of adjunct questions. She used 181 tenth-grade students, and a 19-item short-answer (essay) questions measured remember (RI, RG) and application levels of learning. Results showed that high ability students performed better on high level (UG) post-passage questions, while, low ability students performed better on low level pre-passage questions.

Summary TOP
= = = = = = =
Although there were few studies which addressed the interaction between the type of questions (essay versus multiple-choice ) and the level of questions (RI, RG, and UG ) and its interact with students’ ability, we can conclude from the above studies (e.g., Darwazeh, 1992; Hancock, 1994; Hopkins & Stanely, 1981) that essay questions have great effects on students’ achievement when they were written on high cognitive levels (UG). On the other hand, some researchers (Anderson & Biddle, 1974; Foos & Fisher, 1988) indicated that multiple-choice questions have great effect on students’ achievement when they are written on low levels of learning. Researchers (e.g., Darwazeh,1982; Memory 1983; Sanders, 1973) found that high ability students performed better when they received essay high level postquestions; whereas, low ability students performed better when they received low level prequestions.

General Conclusion of the Research Studies TOP
= = = = = = = = = = = = = = = = = = = = = = = =
From the above studies, the auther found that:
1. Some researchers indicated that essay type questions have more effect on students’ learning than do multiple-choice ones especially on higher level of learning (e.g., Arrasmith, Sheehan & Applebaum, 1984; Hambleton & Murphy, 1992; Lukhele, Thissen, and Wainer, 1994; Rodrick & Anderson; etc). Others, indicated that multiple-choice type questions have more effect on students’ achievement than essay ones especially on lower level of learning (e.g., Bridgman & Lewis,1994; Duchastel & Nungester, 1982; Perkins, 1984; etc.). 

2. Some researchers found that higher level questions (e.g., use a generality ) have greater effect on students’ learning than lower level questions (e.g., Al-Nayef, 1989; Kneip & Grossman, 1979; Redfield, 1981; Rickards & Vesta, 1974; Watts & Anderson, 1971). Few researchers found that lower level questions have greater effect on students’ achievement than higher level ones (e.g., Felkar & Dapra, 1975; Perkins et al.,1990; Samson et al., 1987).

3. Some researchers (e.g., Anderson & Biddle, 1974; Darwazeh, 1992; Hancock, 1994; Hopkins & Stanely,1981 ) indicated that essay questions have greater effect on student achievement when they are written on high levels of learning (e.g., application, evaluation etc.), whereas few researchers (e.g., Anderson & Biddle, 1974; Foos & Fisher, 1988; Haffman, 1962; Lafave, 1966) indicated that multiple-choice questions have greater effect on student’s achievement when they are written on remember levels of learning (remember instance and remember generality). 

After analyzing the previous studies on adjunct questions and their effects on students’ achievement, the researcher did not find any study that addressed the effect of question types ( essay versus multiple-choice ) and levels based on "Merrill’s" Taxonomy of learning ( RI, RG, and UG ) and their interaction with students’ ability except Darwazeh’s studies (1982, 1984). Besides, the researcher believes that the present study will shed light on how to use adjunct questions during instruction in terms of their types, levels, and their interaction with students’ ability levels. Since this topic hasn’t yet been covered by interested researchers, the researcher intended to conduct this study to see whether type and level of adjunct questions have an effect on students’ academic achievement, with a consideration of students’ ability.