[ Abstract for this presentation ] [ Proceedings Contents ] [ Schedule ] [ Abstracts ]

Predicting student views of the classroom: A Californian perspective

Tony Rickards
Curtin University of Technology
Perry den Brok
Utrecht University
Eric Bull
California State University
Darrell Fisher
Curtin University of Technology
WAIER logo
In this study, it was determined which factors influenced students' perceptions of their learning environment. Students' perceptions were measured with the What Is Happening In this Class questionnaire (WIHIC).

Data were analysed for a large sample of US (California) middle school classes. Several variables were included in the study, such as student and teacher gender, student ethnic background and socio-economic background (SES), student age, but also class and school variables, such as class ethnic composition, class size and school SES.

Hierarchical analyses of variance (multilevel analyses) were conducted to investigate separate and joint effects of these variables. In the presentation, we will report on effect sizes of these variables and on the amounts of variance they explained.

Results indicate that some scales of the WIHIC are more inclined to measure personal, idiosyncratic features of students' perceptions of their learning environments, whereas other scales contain more variance at the class level. Also, different variables affect different scale scores. A variable that consistently affected students' perceptions, regardless of the element of interest in the learning environment was student gender. On average, girls perceived their learning environment as more positive than did boys.


Rationale

Over 40 years ago, Mead and Metreaux (1957) asked children to describe their image of a scientist. Many other researchers (Chambers, 1983; Kahle, 1989; Mason, Kahle & Gardner, 1991) have collected and analysed drawings from around the world to see how teachers and students perceive teachers. Like these studies, the present study also tries to obtain an image of how students perceive their Science teachers and learning environment.

Data from the National Education Longitudinal Study of 1988 (NCES, 1992) showed that schools with large free lunch programs tended to offer science classes where experiments were conducted less than once a month. This means that schools with many students from low socio-economic status (SES) backgrounds could have students with very little exposure to quality science programs wherein depictions of science and scientist varieties are likely to be limited. Socio-economic status is a powerful factor in determining who succeeds in American schools. It may be the single, most powerful factor (Lynch, Atwater, Curley, Eccles, Lee, et al., 1996). Barba and Reynolds (1998) contend that the upward mobility rate for the USA has stabilised at about 3% per generation meaning that those born into poverty tend to remain in poverty as adults. Of college graduates earning a Bachelor's degree, 6% of students from low SES backgrounds achieved that status in 1994 compared with 41% of students from high socio-economic status.

Science is largely a result of our culture. One of the reasons students fail to excel in school is not that they come from deprived cultures, but because they come from cultures that are different from the school culture (Presseisen, 1988). An example would be when students are asked to draw a picture of a scientist, the depiction is often of a man working alone. This runs in conflict with many Hispanic children who have been raised in a manner that honours working in socialised groups (Tobias, 1990). Each culture has basic guidelines, which direct beliefs, and values that define how people act, judge, decide and solve problems in their life and in all situations (Anderson, 1988).

The first objective of this study was to validate the What Is Happening In this Class (WIHIC) questionnaire with a large sample of eighth-grade science classes in middle schools in the USA. The second objective was to investigate associations between school socio-economic and racial diversity factors and to students' perceptions of their classroom learning environments.

This study focussed on the socio-economic status and the representation of racial diversity at both the student and school level. We have chosen to use a classroom learning environment questionnaire, the What Is Happening In this Class (WIHIC) to determine what associations those factors have when it comes to individual students' perceptions of science classroom environments.

Theoretical framework: the What Is Happening In this Classroom (WIHIC) questionnaire

Developed by Fraser, Fisher and McRobbie (1996), the WIHIC measures high school students' perceptions of their classroom environment. The WIHIC measures a wide range of dimensions that are important to the present situation in classrooms. It includes relevant dimensions from past questionnaires and combines these with dimensions that measure aspects of constructivism and other emphases relevant to the environment of contemporary classrooms. A description for each scale in the WIHIC is presented in Table 1.

Table 1: Scale description for each scale in the
What Is Happening In This Classroom (WIHIC) Questionnaire.

WIHIC scaleThe extent to which...
Student cohesiveness...students are friendly and supportive of each other.
Teacher support... the teacher helps, befriends, and is interested in students.
Involvement... students have attentive interest, participate in class and are involved with other students in assessing the viability of new ideas.
Investigation..there is emphasis on the skills and of inquiry and their use in problem solving and investigation.
Task orientation... it is important to complete planned activities and stay on the subject matter.
Cooperation... students cooperate with each other during activities.
Equity... the teacher treats students equally, including distributing praise, question distribution and opportunities to be included in discussions.

The origi nal version of the WIHIC contained 90 items and nine scales, but was refined by both statistical analysis of data from 355 junior high school science students, and extensive interviewing of students about their views of their classroom environments in general, the wording and salience of individual items and their questionnaire responses (Fraser, Fisher & McRobbie 1996). Only 54 items in seven scales survived these procedures, although this set of items was expanded to 80 items in eight scales for the field testing of the second version of the WIHIC, which involved junior high school science classes in Australia and Taiwan. Whereas the Australian sample of 1,081 students in 50 classes responded to the original English version, a Taiwanese sample of 1,879 students in 50 classes responded to a Chinese version that had undergone careful procedures of translation and back translation (Huang & Fraser, 1997). This led to a final form of the WIHIC containing the seven eight-item scales.

Since its development, the WIHIC has been used to measure the psychosocial aspects of the classroom learning environment in various contexts. In some research, the questionnaire has been used without any modifications, and in others the questionnaire was adapted to suit the specific context. To date, the original questionnaire in English has been translated into Chinese for use in Taiwan (Aldridge & Fraser, 1997) and Singapore (Chionh & Fraser, 1998) and Korean for use in Korea (Kim, Fisher, & Fraser, 2000).

A study by Rawnsley and Fisher (1998) investigated associations between learning environments in mathematics classrooms and students' attitudes towards that subject in Australia using the WIHIC questionnaire. It was found that students developed more positive attitudes towards their mathematics in classes where the teacher was perceived to be highly supportive, equitable, and in which the teacher involved them in investigations. Chionh and Fraser (1998) used actual and preferred forms of the WIHIC to further validate the instrument and to investigate associations between actual classroom environment and the outcomes. The associations between five different outcome measure namely, examination results, self esteem, and three attitude scales and the seven actual classroom environment scales were investigated in geography and mathematics classrooms in Singapore and Australia. The study revealed that better examination scores were found in geography and mathematics classrooms where students perceived the environment as more cohesive. It was also found that self esteem and attitudes were more favourable in classrooms perceived as having more teacher support, task orientation and equity.

A study by Khoo and Fraser (1997) used a modified version of the WIHIC to measure classroom environment in evaluating adult computer courses. In investigating the differential effectiveness of computer courses for different gender, they found that males perceived significantly greater Involvement. At the same time, it was found that females perceived significantly higher levels of Equity in the computer classroom environment. In the Trainer Support dimension, sex and age interaction occurred in addition to a significant sex main effect. It was found that males perceived greater Trainer Support than females, but older females has more positive perceptions than younger females. Gender related differences in students' perceptions of their learning environment and teacher personal behaviour were explored by Kim, Fraser, and Fisher (2000). The study involved 543 grade 8 students in 12 different secondary schools in metropolitan and rural areas of Korea. Statistically significant differences were found between boys' and girls' on all seven scales. It was reported that boys perceived more Teacher Support, Involvement, Investigation, Task Orientation, and Equity than did girls.

In examining education systems in different contexts and cultures, there is a suggestion that there are some fundamental differences in approaches. Schools in Asia are more examination oriented and teachers are seen as authority figures. As such, students from an Asian background seem to perceive their teachers significantly more positively in the classroom compared with students from other cultural backgrounds (Fisher & Rickards, 1998; Rickards, 1998).

In an attempt to explore the potential of cross-cultural studies, Fraser and Aldridge (1998) examined classrooms in Australia and Taiwan using English and Chinese versions of the WIHIC. The results show that students in Australia consistently viewed their classroom environment more positively then students in Taiwan. Significant differences were detected on the WIHIC scales of Involvement, Investigation, Task Orientation, Cooperation, and Equity. This means that students in Australia perceived they are given more opportunity to get involved in the experiments and investigate scientific phenomena. They also have an opinion that teachers are cooperative and give an equal chance of participation to both genders. It appears that the education system in Taiwan is examination driven and teaching styles are adopted to suit that particular situation. It was also found that in Taiwan the most important element of being a good teacher was perceived as having good content knowledge, but in Australia, having a good interpersonal relation between teacher and students is considered the most important element in the education process. The study indicates that the WIHIC is useful for differentiating between cultural differences in the classroom environment and therefore might be suitable for a study on multicultural classes, as is the case for the sample used in the present study.

Research questions

This was the first large study in the USA using the WIHIC with eighth-grade science classes. While research exists for individual ethnicity and classroom performance (see previous section), it is not known whether or not research has been done concerning students within schools whose populations are defined as having high, medium, or low diversities of race, and what perceptions those students have of their science classes. Likewise, since Fraser (1998) has established that students' perceptions of their classroom environment can affect student achievement and attitude to class, it is important to determine if the WIHIC can be used to discriminate between those factors and associations that may influence perception. Also, examining students in low, medium, or high SES school environments is a unique approach different from looking at individual status. This resulted in the following research questions for the study:
  1. To what degree do students' perceptions of their learning environment in terms of WIHIC scales differentiate between schools and teachers? To what degree are these perceptions idiosyncratic?

  2. In what way are students' perceptions of their learning environment in terms of WIHIC scales determined by their cultural and socio-economic background, or by class and school representatives of these variables?

Method

Instrumentation

To assess students' perceptions of their learning environment, the WIHIC was administered to all students of participating classes and schools. The WIHIC contains 56 items that are answered on a five point Likert type Scale. The items refer to 7 scales. For each scale, Table 2 presents a typical item.

Table 2: Typical items for the WIHIC scales

ScaleTypical item
Student cohesivenessI work well with other class members.
Teacher supportThe teacher helps me when I have trouble with the work.
InvolvementI give my opinion during class discussions.
InvestigationI find out answers to questions by doing investigations.
Task orientationI know how much work I have to do.
CooperationWhen I work in groups in this class, there is teamwork.
EquityI am treated the same as other students in this class.

Since this study was the first to use the WIHIC on an American sample, several analyses were done to investigate its quality. First, it was determined whether scales had been measured reliably by computing a Cronbach's Alpha at the student and class (aggregated) level. The findings of these analyses are given in Table 3. Second, it was determined to what degree scales of the WIHIC were able to differentiate between classes by computing an intra-class coefficient. The intra-class coefficient represents the ratio between the amount of variance at the class (and school) level and the student level. Third, consistency of the WIHIC scales was determined by computing Multilevel Lambda (Snijders & Bosker, 1999). Lambda is based on both the reliability and intra-class correlation coefficients and represents the degree to which the instrument is capable in measuring consistently across classes.

From Table 3 it can be seen that all scales are very reliable. Reliability ranges between .77 (student cohesiveness) and .89 (teacher support; cooperation) at the student level, and between .78 (student cohesiveness) and .96 (teacher support) at the class level.

Table 3: Reliability (alpha), intra-class coefficients (ICC) and consistency (lambda) of WIHIC scales

Scalealpha (student)alpha (class)ICClambda
Student cohesiveness.77.78.021.34
Teacher support .89.96.177.84
Involvement .86.89.045.53
Investigation .88.92.056.53
Task orientation .84.84.021.34
Cooperation .86.91.067.63
Equity .89.93.124.77

Intra-class, correlations are rather low, ranging from .02 (student cohesiveness and task orientation) to .18 (teacher support). These findings suggest that several scales, such as student cohesiveness, task orientation, involvement, investigation and cooperation, are hardly able to distinguish between classes and/or schools, at least with respect to the sample used in this study. The teacher support and equity scales are most sensitive for indicating differences between classes. It seems as if most of the variance in the WIHIC scales (well up to over 90 percent) pertains to differences between individual students, rather than differences between classes or schools. These findings are rather low compared to those reported in other studies using the WIHIC or other learning environments instruments (eg. Fraser, 1998). The low intra-class correlations, however, might not only be the result of sample characteristics. The method employed in this study to check for discriminant validity was also different from those in earlier research (multilevel analysis instead of one way analysis of variance). Further research, using multilevel analysis on other USA samples should verify the stability of these findings.

Due to the low intra-class coefficients consistency is also low for a number of scales. Again, teacher support, equity and to some degree cooperation seem to have been measured rather consistently across classes, while this is less true for the other variables.

A further analysis of the quality of measurement involved a principal components factor analysis with varimax rotation. The results of this factor analysis (which are not presented in this paper) confirmed that the a priori seven factor structure was replicated, with all items having a factor loading greater than 0.34. Thus, the seven scale nature of the WIHIC was confirmed.

Finally, correlations between the WIHIC scales were computed, in order to see whether they referred to distinctively different aspects of the learning environment. These correlations are presented in Table 4. As can be seen, the scales seem to measure distinct aspects, but also show some overlap. This is particularly true for involvement and student support, for investigation and teacher support, and for task orientation and involvement. It was concluded that the scales represented different elements and could be treated as separate concepts for analyses.

Several other variables were included in this study. Students were asked to indicate their self perceived ethnic group membership (Latino, African-American, Asian, Native-American, White-American or Other), and their gender. Also, teachers were asked to indicate their gender. Socio-economic status at the school level was determined by examining free and reduced lunch percentages. Racial Diversity for each school was determined through county demographics, which listed ethnicity percentages for all schools within the county's jurisdiction. Racial diversity of the school was coded in terms of 5 categories, with a score of 1 referring to a percentage of between 0 and 20 percent of non-White students, a score of 5 referring to a percentage of between 80 and 100 percent non-white students. Similar percentage scores were also used for the school socio-economic variable.

Table 4: Correlations between WIHIC scales*


CohesivenessSupportInvolvementInvestigationTask orientationCooperationEquity
Cohesiveness
.37.43.27.38.54.27
Support.28
.45.38.31.32.38
Involvement.45 .81
.52.26.37.27
Investigation.20 .71.69
.33.33.22
Task orientation.37 .52.70.61
.42.40
Cooperation.44 .44.59.42.46
.39
Equity.31 .66.61.43.61.58
* Within class correlations are given above the diagonal, while correlations at the teacher-class level are given above the diagonal.

Analyses

To find an answer to the research questions hierarchical analysis of variance (multilevel analyses) was conducted, using ML3E. By our knowledge, this was the first WIHIC study to employ multilevel analysis. Multilevel analyses take into account that data have not been randomly sampled and allow for including multiple variables at different levels a t the same time in one analysis. Since it may be assumed that responses of students that share a similar history, experience and class background are more alike compared to those of students from different classes, regular analyses of variance tend to overestimate the effects of variables (eg. Hox, 1995).

In the analyses, three levels of variance were distinguished: a student level, a teacher/class level and a school level. Standard estimation procedures in multilevel analyses programs, such as Iterative Generalised Least Squares (IGLS), often produce biased estimates of coefficients and variance distribution, especially when small numbers of units are available at the higher levels (Luyten & De Jong, 1998). Because of the small number of schools and classes involved in this study, it was decided to use the Restricted Iterative Generalised Least Squares (RIGLS) method, which is suitable for small numbers of units at the highest levels (Goldstein, 1995).

Analyses were conducted in two steps and were done separately for each of the WIHIC scales. To answer the first research question, referring to the amounts of variance located at each of the three levels, we formulated an empty model (with no explanatory variables), that provided a scale mean for the sample and estimates of variance at the student, class/teacher and school level. The second step consisted of entering all explanatory variables into the models. Next, variables displaying non-significant relationships were removed from the models. For each of the significant variables, apart from regression coefficients and standard errors, we also computed effect sizes.

Variables entered in the second step of the analyses were:

Gender and the student ethnic background variables were entered as dummy variables (with boys representing the baseline and girls the score of 1; for the cultural groups 1 referred to the particular cultural group, White students were used as the baseline). The student gender and ethnic background variables were also used to create the class composition variables mentioned above.

Sample

The study originally involved a sample of 1,720 students from eight-grade science classes from 11 Californian schools. Teachers and schools participated on a voluntary basis. Unfortunately, student characteristics were only available for a part of this sample. Therefore, in the analyses only those respondents were included in the analyses that could be linked to student background variables. This resulted in an ultimate sample of 655 students.

The sample used was relatively heterogeneous in terms of ethnic makeup: 135 students (20.7 percent) indicated to perceive themselves as Latino/Hispanic, 104 students (15.9 percent) as Afro-American, 6 students (0.9 percent) as Native-American, 93 students (14.2 percent) as Asian, 229 students (35.0 percent) as White- or Caucasian-American and 88 students (13.4 percent) as Other. Of the sample, 319 students (48.7 percent) were male and 335 female. Most of the teachers were female (11 out of 18).

Table 5: Descriptive statistics for variables used in the study (n=655)

VariableMinimumMaximumMeanStandard
deviation
Student gender01--
Latino01--
African-Am.01--
Native-Am.01--
Asian01--
White-American01--
Class size1530263.25
% boys in class3573519
% Hispanics in class0822118.9
% Africans in class326166.84
% Natives in class01092.34
% Asians in class029147.69
% Others in class425146.55
Number of cultures364.96.64
Teacher gender01--
School SES242.97.73
School diversity253.481.15

None of the schools contained less than 20 percent non-white students or less than 20 percent students receiving a meal at school. While the percentages of non-White students varied between schools from 21 to 100 percent, most schools contained between 21 and 40 percent non-White students, almost a quarter of the schools contained between 81 and 100 percent non-white students. In a similar manner, the percentage of students receiving a meal at school varied between 21 and 80 percent in the sample, with most schools containing a percentage of between 41 and 60 students receiving a meal.

Table 5 presents minimum and maximum scores for each of the explanatory variables included in this study, as well as the average score and standard deviation found in the sample.

Results

The results for the empty models, providing the amount of variance present at the school, class and student levels are displayed in Table 6.

The outcomes presented in Table 6 reflect those given in Table 3. It can be seen that most of the variance is located at the student level, with some variance at the class level and hardly any variance at the school level. For the teacher support and equity scales, fair amounts of variance relate to class variables, while there is some distinction between schools with respect to cooperation. These results indicate that, while student perceptions are determined for the larger part by student characteristics, for some elements of the learning environment, there are also distinct differences between teachers or classes, and even between schools. Table 6 also shows that, on average, students perceive high amounts of student cohesion, task orientation, cooperation and equity, but low amounts of teacher support, involvement and investigation.

Table 6: Variance distribution for the WIHIC scales (empty model)

VariableMean (st. error)School (%)Class (%)Student (%)-2*Log-likelihood
Student cohesion3.95 (.04)0.281.1298.601181.85
Teacher support2.72 (.14)016.0583.951666.34
Involvement2.83 (.04)0.430.1499.431627.44
Investigation2.64 (.06)0.270.8098.931645.79
Task orientation4.13 (.05)03.4396.571302.70
Cooperation3.70 (.09)3.131.0495.831578.46
Equity3.57 (.17)023.8676.141668.73

The second research question deals with the amounts of variance that can be explained (or degree to which perceptions can be predicted) by the variables included in this study. Student, class and school characteristics as included in this study, are only associated with WIHIC scale scores to a limited extent. The outcomes relating to the effects of variables on the WIHIC scale scores are given in Table 7.

Student gender appears to be related to four scales: student cohesion, teacher support, task orientation and cooperation. For all of these scales girls have higher ratings than boys, indicating that they have a more favourable perception of the learning environment.

Student ethnicity is not related to any of the scale scores in itself, but the class makeup variables that are constructed out of these variables are. The percentage of Latino/Hispanic students is negatively related to the amount of cooperation perceived. This means that the more Hispanic students are present in the class, the less favourable the class perception of cooperation is.

Table 7: Outcomes of multilevel analyses on WIHIC scales
(significant explanatory variables; standard errors between brackets; effect size after forward slash)


Student
cohesion
Teacher
support
InvolvementInvestigationTask
orientation
CooperationEquity
Mean (= constant)3.34 (.25)2.64 (.17)3.04 (.10)1.81 (.32)4.05 (.06)4.44 (.18)3.57 (.17)
Effects
Student gender.17 (.05)
/ .0255
.14 (.07)
/ .014
N.S.N.S..15 (.05)
/ .020
.16 (.06)
/ .018
N.S.
% LatinoN.S.N.S.N.S.N.S.N.S.-1.07 (.20)
/ -.229
N.S.
% AfricanN.S.N.S.-1.32 (.53)
/ -.103
N.S.N.S.-2.51 (.60)
/ -.212
N.S.
% Indian2.46 (1.22)
/ .096
N.S.N.S.N.S.N.S.N.S.N.S.
% otherN.S.N.S.N.S.N.S.N.S.-1.36 (.58)
/ -.110
N.S.
no. of cultures.09 (.04)
/ .096
N.S.N.S.N.S.N.S.N.S.N.S.
class sizeN.S.N.S.N.S..03 (.01)
/ .115
N.S.N.S.N.S.
teach gender.13 (.06)
/ .140
N.S.N.S..23 (.09)
/ .141
N.S.N.S.N.S.
Total % of variance explained3.350.580.140.121.147.300.0
% of explained variance
school10000001000
teacher750.7100100071.40
student2.30.61.01.31.43.60
-2*log-likelihood1158.881661.731622.401637.151293.621541.821668.73
Note: N.S. = non-significant

The percentage of African-American students in class is negatively related to involvement and cooperation. The percentage of Native-American students in class is positively related to student cohesion. However, given the low number of Native students present in the sample, this finding only has limited significance. The percentage of other students in class is negatively related to cooperation. The number of cultures in a class is positively related to student cohesion: thus, the more different cultures in a class, the more student cohesion is perceived.

Class size is positively related to investigation: in larger classes students perceive more i nvestigation. Finally, teacher gender is related to student cohesion and investigation: for female teachers higher ratings are reported for these scales.

When looking at the effect sizes, it can be seen that teacher gender is relatively stronger associated to students' perceptions than student gender or class composition variables. Class size also seems quite important looking at its effect size. For cooperation, the percentage of Hispanic and African-American students is relatively important, when compared to student gender or the percentage of Other cultures in the class.

The models explain less than 7 percent of the total variance in each variable. This means that other variables than the ones used in the study are responsible for differences in student perceptions. Overall, the variables do explain much of the variance at the teacher/class level of most of the scales, and only small parts of the student variance. This means that the gender and ethnicity makeup of a class explain to a large degree how a class will perceive its learning environment. Variables hardly explain any variance for teacher support, involvement, investigation and equity, though some variance is explained for cooperation (7.3 percent) and student cohesion (3.4 percent). No interaction effects between variables were found.

Discussion

This research has provided further evidence on the validation of the WIHIC, which assesses seven scales of student perceptions of the classroom environment. The WIHIC for use with this sample was shown to be valid and reliable. However, its ability to distinguish between Californian multicultural classes, teachers or schools was found to be limited.

This study provides the first study of associations between gender, attitude, racial diversity, ethnic origin and socio-economic status in US eighth-grade classes in California. As a result, it has provided the first validation data for the WIHIC in USA science classes in California.

A number of interesting findings were reported. First, it was found that girls have a more favourable perception of their Science learning environment in terms of scales of the WIHIC than boys. This finding was somewhat surprising, since earlier studies using the WIHIC indicated the opposite (eg. Khoo & Fraser, 1997; Kim, Fisher & Fraser, 2000). Of course, different methods of analyses and the context or country of study might help to explain this. On the other hand, research with other learning environments instruments, such as the Questionnaire on Teacher Interaction (QTI) has also indicated that girls have a more favourable perception of their learning environment (eg. den Brok, Levy, Wubbels & Rodriguez, 2003; Goh & Fraser, 1995; Levy, den Brok, Wubbels & Brekelmans, 2003; Levy, Wubbels & Brekelmans, 1992; Wubbels & Levy, 1993).

Second, the fact that several class ethnic makeup variables displayed a significant effect was also a finding that had not been reported previously. Earlier work using the WIHIC never used such variables. A study investigating the effects of class makeup with the QTI did report that classes with many Asian-American students had more favourable perceptions of the learning environment. An important finding in the present study was the positive association between the number of ethnic groups in the classroom and their perception of student cohesion. Apparently, classes without any dominant groups but a high degree of diversity are important for a students' belonging. Chances are that students that are not part of a dominant group may feel themselves isolated.

Third, class size was positively related to investigation. This seems logical, as teachers have less time to help students on an individual basis in larger classes, which means students have to find things out more by themselves.

Finally, teacher gender was related to a number of scales, with classes taught by female teachers displaying a more favourable picture. Again, this finding has not been investigated in other WIHIC studies. However, similar patterns have been found in research using the QTI (eg. Levy, et al., 2003).

Limited amounts of variance were reported and explained for the class and school level. While this finding was contradictory to earlier WIHIC studies, it might be related to the differences in methodology and characteristics of this particular sample (see Instrumentation section). However, other studies employing multilevel analyses on learning environments instruments indicate similar findings (eg. den Brok, 2001; den Brok, Levy, Rodriguez & Wubbels, 2002; Levy, et al., 2003).

Unfortunately, the study was subject to a number of limitations, some of which have been mentioned before. No knowledge was available indicating to what extent the sample used was representative of the larger population of Californian teachers. Moreover, the sample was relatively small, in particular with respect to the number of classes and schools surveyed. Future research on larger and other USA samples will be necessary to verify the stability of our findings. Then, only some and very specific variables were included in our effort, while many other important variables were not taken along, such as motivation, achievement, teacher experience, et cetera. Again, incremental efforts will be necessary in the future to uncover the potentially important variables that can predict students' perception of their learning environment. Given the small percentages of variance found for the class level, the limited number of associations found for our explanatory variables is not surprising.

This study has several implications. First, it has been shown that class composition may be of importance in creating a suitable, safe and effective learning environment. Schools can affect students perceptions and school career to some degree by making sure that students are placed in such a way that no single ethnic group (or any group in terms of other student characteristics for that matter) is dominant in terms of numbers. Diversity might even be used as a tool to create a favourable, rich and cohesive learning environment. Second, teachers should realise that their efforts may perceived differently by different students (eg. girls or boys, students from different ethnic groups or socio-economic backgrounds). Knowledge on how perceptions are affected by these characteristics may be relevant to affirm certain groups in the classroom and provide knowledge on how a teacher comes across. For researchers it may be important to test whether and how their instruments are able to discriminate between classes and/or schools in their specific sample, even if such capability has been demonstrated before on other samples or in other contexts.

References

Aldridge, J. M. and Fraser, B. J. (1997). Examining science classroom environments in a cross-national study. Proceedings Western Australian Institute for Educational Research Forum 1997. http://www.waier.org.au/forums/1997/aldridge.html

Anderson, J. A. (1988). Cognitive styles and multicultural populations. Journal of Teacher Education, 39(1), 2-9.

Barba, R. H., & Reynolds, K. E. (1998). Towards an equitable learning environment in science for Hispanic students. In B. J. Fraser and K. G. London (Eds), International Handbook of Science Education. Kluwer Academic Publishers: 925-939.

Chambers, D. W. (1983). Stereotypic images of the scientist: The draw-a-scientist test. Science Education, 67(2), 255-265.

Chionh Y. H & Fraser, B.J. (1998). Validation and use of the "What Is Happening In This Class" Questionnaire in Singapore. Paper presented at Annual Meeting of the American Educational Research Association, San Diego, April 1997.

Den Brok, P. (2001). Teaching and student outcomes. Utrecht, the Netherlands: W. C. C.

Den Brok, P., Levy, J., Rodriguez, R., & Wubbels, T. (2002). Perceptions of Asian-American and Hispanic-American teachers and their students on interpersonal communication style. Teaching and Teacher Education, 18 (4), 447-467.

Den Brok, P., Levy, J., Wubbels, T., & Rodriguez, M. (2003). Cultural influences on students' perceptions of videotaped lessons. International Journal of Intercultural Relations, 27(3), 355-374.

Fisher, D.L., & Rickards, T. (1998). Cultural background and gender differences in science teacher-student classroom interactions: Associations with student attitude and achievement. In L.Y.Pak., L. Ferrer & M. Quigley (Eds.), Science, Mathematics and Teacher Education for National Development (pp. 55-56). Brunei: Universiti Brunei Darussalam.

Fraser, B. J. (1998). Science learning environments: Assessments effects and determinants. In B. J. Fraser & K. Tobin (Eds), International Handbook of Science Education. Dordrecht, The Netherlands, Kluwer Academic Publishers. 1: 527-564.

Fraser, B.J., & Aldridge, J.M. (1998). The potential of cross-national studies of science classrooms: An example involving Australia and Taiwan. In L.Y.Pak., L. Ferrer. & M. Quigley (Eds), Science, Mathematics and Technical Education for National Development (pp. 76-83). Brunei: Universiti Brunei Darussalam.

Fraser, B.J., Fisher, D.L. & McRobbie, C.J. (1996, April). Development, validation and use of personal and class forms of a new classroom environment instrument. Paper presented at the annual meeting of the American Educational Research Association, New York, USA.

Goh, S. C., & Fraser, B. J. (1995, April). Learning environment and student outcomes in primary mathematics classrooms in Singapore. Paper presented at the annual meeting of the American Educational Research Association, San Francisco.

Goldstein, H. (1995). Multilevel statistical models. London: Edward Arnold.

Huang, I. & Fraser, B.J. (1997). The Development of a Questionnaire for Assessing Student Perceptions of Classroom Climate in Taiwan and Australia. Paper presented at the annual meeting of the National Association for Research in Science Teaching, Chicago, IL.

Kahle, J. B. (1989). Images of scientists: Gender issues in science classrooms. Perth: Key Centre for Science and Mathematics Education, Curtin University of Technology.

Khoo, H.S. & Fraser, B.J. (1997). Using classroom environment dimensions in the evaluation of adult computer courses in Singapore. Paper presented at Annual Meeting of the American Educational Research Association, Chicago, March 1997.

Kim, H., Fisher, D. & Fraser, B. (2000). Classroom environment and teacher interpersonal behaviour in secondary science classes in Korea. Evaluation and Research in Education, 14, 3-22.

Levy, J., den Brok, P., Wubbels, T., & Brekelmans, M. (2003). Students' perceptions of the interpersonal aspect of the learning environment. Learning Environments Research, 6 (1), 5-36.

Levy, J., Wubbels, T., & Brekelmans, M. (1992). Student and teacher characteristics and perceptions of teacher communication style. Journal of Classroom Interaction, 27, 23-39.

Luyten, H., & De Jong, R. (1998). Parallel classes: Differences and similarities, teacher effects and school effects in secondary schools. School Effectiveness and School Improvement, 9, 437-473.

Lynch, S., Atwater, M., Cawley, J., Eccles, J., Lee, O., Marrett, C., Rojas-Medlin, D., Secada, W., Stefanivh, G. & Willetto, A. (1996). An equity blueprint for project 2061. Washington, DC: American Association for the Advancement of Science.

Mason, C. L., Kahle, J. B. & Gardner, A. L. (1991). Draw-a-scientist test: Future implications. School Science and Mathematics, 91(5), 193-198.

Mead, M. & Metreaux, R. (1957). The image of the scientist among high school children. Science 126, 384-389.

NCES (1992). National education longitudinal study of 1988: A profile of American eighth-grade mathematics and science instruction. Washington, DC: Department of Education, Office of Educational Research and Improvement.

Presseisen, B. Z. (1988). Focus on the at risk learner: An introduction. In B. Presseisen (Ed), At Risk Students and Thinking: Perspectives from Research. Washington, DC: National Education Association/Research for Better Schools: 11-18.

Rawnsley, D. & Fisher, D.L. (1998,). Learning environments in mathematics classrooms and their associations with students' attitudes and learning. A paper presented at the Australian Association for Research in Education Conference, Adelaide, December.

Rickards, A. W. J. (1998). The relationship of teacher-student interpersonal behaviour with student sex, cultural background and student outcomes. Science and Mathematics Education Centre. Perth, Curtin University of Technology: 207.

Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel analysis: An introduction to basic and advanced multilevel modeling. London: Sage Publications.

Tobias, S. (1990). They're not dumb, they're different: Stalking the second tier. Tucson, AZ, Research Corporation - A Foundation for the Advancement of Science.

Wubbels, Th. & Levy, J. (1993). Do you know what you look like? London: The Falmer Press.

Contact person: t.rickards@curtin.edu.au

Please cite as: Rickards, T., den Brok, P., Bull, E. and Fisher, D. (2003). Predicting student views of the classroom: A Californian perspective. Proceedings Western Australian Institute for Educational Research Forum 2003. http://www.waier.org.au/forums/2003/rickards-1.html


[ Abstract for this presentation ] [ Proceedings Contents ] [ Schedule ] [ Abstracts ]
Created 1 Sep 2003. Last revised 19 May 2006. URL: http://www.waier.org.au/forums/2003/rickards-1.html
The Forum Proceedings are © Western Australian Institute for Educational Research. However
the copyright for each individual article remains with the authors of the article.
HTML: Roger Atkinson and Clare McBeath