杨维

The purpose of the present study was to examine the relationship between self-concept and academic self-concept， parent education， parent and student attitudes towards school， study habits and student achievement in a group of junior high students. This study investigated whether student achievement scores varied significantly by ratings for various aspects of self-concept as well as global self-concept， parent educational levels and gender. As well， the study examined how student and parental attitudes towards school correlated with each other. A third aspect of the study was an exploration of the best predictors for achievement in math， English and overall average academic.

Sixty-nine grade six students participated in this study， Student responses to the Battle （2002） Culture Free Self-Esteem Inventory and researcher developed scales to examine student attitudes towards school and study habits were collected. Student achievement scores on English and math tests for the last two terms of grade six were obtained from the cumulative records of the students. A parental attitudes towards school questionnaire developed by the researcher and a demographic data form designed to obtain selected background information including parental education attainment were completed by consenting parents and returned to the school along with consent forms for their children to participate in the study.

ANOVA， t-test， correlational and regressional analysis were performed on the data and produced the following major findings. Student groups with above-average Global and Academic Self-Concept achieved significantly higher scores in English and math than groups with lower Global and Academic Self-Concept ratings. However， gender had little or no effect on the achievement levels. Relative to the correlations among the variables， Global Self-Concept and Academic Self-Concept ranked among the top three with a positive correlation with the achievement variables. A surprising finding was that parent attitudes towards school was negatively correlated with parent education. Of all variables entered for predicting the achievement levels of the students， Academic Self-Concept turned out to be the strongest predictor for English and also the only predictor for the math and the average academic achievement scores. Parent education was the second strongest in predicting the English scores. All the other variables including Global Self-Concept played little or no part in predicting the performances on any of the achievement measures.

INTRODCUTION

Over the years， educational researchers have been striving to find factors that are related to students academic achievement. Of those important factors correlated with achievement level of students， self-concept turned out to be a very significant non-cognitive variable for explaining a certain amount of variance in academic achievement. （Mayo， Ethel & Gibbs， 1997）. Academic achievement is highly susceptible to a variety of cognitive and non-cognitive variables （Finn & Rock， 1997； Ford， 1996； Slate， Jones & Dawson， 1993）. Since research found that parental education level， parental attitudes towards school， student attitudes towards school and study habits impact on the final achievement level of students， it is reasonable to conduct a multiple-regression analysis to produce a list of independent variables that may affect the academic achievement and to rank them in ascending order in terms of their influence. Therefore， this study obtained quantitative data from responses to a group of standardized and self-developed questionnaires that aim to measure all variables included on 5-point rating scales. The study also investigates the correlations among these variables and the portion of the variance in academic achievement that can be accounted for by various independent variables. This research study may have important theoretical and practical educational applications. Since it has been shown by the previous research that there is a moderate， yet significant correlation between global self-concept and academic achievement， the results of this correlational study may provide new insight on global and academic self-concept and how they are correlated with students achievement level.

This study will also help examine whether the academic dimension of self-concept can serve as a better predictor of academic achievement. The research will continue testing the theories that view self-concept on a multi-dimensional continuum.

Since some other student variables and parent variables are included， the study will help identify a number of factors such as parent attitudes towards school that might help explain part of the variance in students academic achievement. The study should produce a rank order of all these variables involved in the research in the hope of identifying the most significant factors responsible for most of the variance in students level of achievement. A somewhat unique focus for the current study is the inclusion of parental attitudes.

The purpose of this study is to identify variables correlated with students achievement level. This research study will also generate a list of significant predictors ranked in descending order of predictive ability that explain most of the variance in students level of achievement.

METHODOLOGY

Research participants consisting of 69 grade 6 students were drawn from a junior high school in a middle size town in a mixed urban/rural setting. Although intellectual ability is deemed as an important factor that impacts the student achievement level， it can be reasonably assumed for the general student population at the junior high level that most of them would fall within the normal range of intelligence， as along as they had no experience with additional educational support or in repeating grades. Data were gathered on the Demographic Data Form regarding whether a student has received special educational support and/or repeated a grade. This information was used in subsequent data analyses.

The 15-minute questionnaire by Battle （1990） titled Culture-Free Self-Esteem Inventory （CFSEI） consisting of four cluster scales （Academic， General， Parental/Home， and Social） that are intended to measure general self-esteem and several particular areas of self-esteem of participants was administered to the selected samples. CFSEI is regarded as a reliable and valid instrument that can accurately assess a participants perception of self-worth in a broad sense and across different domains of self-esteem. In addition， three questionnaires developed by the researcher were administered to the sample. The first questionnaire aims at measuring several major aspects of school characteristics， including students internal interests in school subjects. The second questionnaire was created to evaluate how parents feel about school in general and important functions of school education. It covers areas such as the role of school in shaping developmental changes in adolescents， the relevance of education received in school to career requirements， etc. The above mentioned two questionnaires have the purpose of measuring students attitudes towards school and parents attitudes towards school. The third questionnaire was created for measuring students study habits. This questionnaire seeks to understand a wide range of components that are considered to be important to academic achievement. It covers students persistence in solving hard problems， study techniques， metacognitive control over learning process， time management and general attitudes towards learning.

Once permission has been granted from the regional school board， contact was made with several school principals who showed interest in this study. Once the principal agreed to allow the students in the school to participate， discussion were held with teacher prospects about the nature and purposes of the research study and information letters were presented to them as well. Those who decided to permit their students to participate in the study were then asked to pass out to every student in their classes an information package that contained two information letters. One was for youth and one was for their parents/guardians； also included was a consent form and a Parental Attitudes Towards School questionnaire. A basic demographic data form was also completed along with the consent forms for the purposes of obtaining parent educational level and other information such as special program support. Upon completion of the consent forms， the students were asked to return their finished consent forms， demographic data form and the Parental Attitudes Towards School questionnaire in an envelope provided to a drop box in the principals office at the school. Once the consent forms were gathered， a mutually agreed-upon time was arranged in consultation with each teacher involved for survey administration. The participating students completed the questionnaires in one designated room over approximately a one hour and half period of time. Following receipt of consent， student achievement records in math and English were gathered from the school.

Data analyses were performed on SPSS to generate inferential statistical data for research purposes. First， since many variables were measured on the rating scales with different ranges of scores， raw scores on each measure were converted into standardized scores. Then， a number of ANOVA and t-tests were run to sort out significant mean differences in the achievement variables and some other interested variables across subgroups within categorical variables such as Global Self-Concept groups， Parental Education and gender. As well， correlations among variables were calculated to understand their interactive relations and potential abilities to predict the achievement scores. Lastly， multiple regression analyses were performed for the purpose of producing the rank order of predictors for each achievement variable and for each of the other variables concerned.

RESULTS

Data analyses were performed in response to the following ten research questions. Research question 1 what is the relationship between performance on average academic achievement， math and English tests and subscale scores on Culture-Free Self-Esteem Inventory （CFSEI）？ The three achievement scores of the new above-average group were significantly higher than those of the new below-average group for Academic Self-Concept and General Self-Concept. For Parental/Home Self-Concept， the new average group achieved significantly lower scores on three achievement variables when compared to the new above-average group. The higher the new group divisions within Personal Self-Concept， the higher the average academic achievement and math scores were. There was no relationship between Social Self-Concept and three achievement scores. Research question 2 What is the relationship between performance on average academic achievement， math and English tests and parental education？ The scores in English were significant lower for the low Parent Education group compared to the high Parent Education group. However， no significant mean differences in average academic scores and math scores were found in the group comparisons. Research question 3 What is the relationship between performance on average academic achievement， math， English and gender？ The data analysis results showed that no significant mean differences in average academic achievement， English and math were found between the female group and the male group. Research question 4 What is the relationship among average academic achievement， English achievement， math achievement， Academic Self-Concept and gender？ The results showed that Academic Self-Concept exerted gendered effects on the two gender groups. It had a positive effect on all achievement variables for females， but not for males. In a word， Gender alone did not impact the achievement variables. Research question 5： What is the relationship among performance on Parental Attitudes Towards School questionnaire， Student Attitudes Towards School questionnaire and Student Study Habits questionnaire and Parental Education？ Parental Education negatively impacted on Parental Attitudes Towards School as the low group scored significantly lower on Parental Attitudes Towards School questionnaire when compared to the average and high groups. There were no relationships among Parent Education， Student Attitudes Towards School and Student Study Habits. Research question 6： Is there a correlation between Parent Education and Parental Attitudes Towards School？ Which of the following variables such as Student Attitudes Towards School are correlated with Student Study Habits？ What， if any are the intercorrelations among those achievement variables？ The only variable found to correlate with Parental Attitudes Towards School was Parent Education. Parent Education has a positive impact on Parental Attitudes Towards School. Student Study Habits was moderately correlated with Student Attitudes Towards School only and was also positively impacted by this variable. Average academic score was more highly correlated with math score than with English score. Research question 7： Which of the five CFSEI subscales or Global Self-Concept is most closely correlated with average academic score， English score and Math score？ What is the ranking of variables involved in descending order of correlation with each of the student achievement variables？ Academic Self-Concept turned out to most correlate with three achievement variables. Academic Self-Concept and Global Self-Concept ranked top two on the list of self-concept related variables in descending order of correlation with average academic achievement and math achievement， respectively. However， Academic Self-Concept， General Self-Concept and Global Self-Concept ranked on the top of the list of self-concept related variables in order of descending correlation with English. Research question 8： How do Parent Education and Student Attitudes Towards School correlate with the 5 CFSEI subscales and Global Self-Concept and with three achievement variables？ Based on the results of the data analyses， Parent Education was not correlated with Academic Self-Concept and was only correlated with English， but not with the other achievement variables. Student Attitudes Towards School was moderately correlated with Parent/Home Self-Concept， Academic Self-Concept and Global Self-Concept only. Also， Student Attitudes Towards School was uncorrelated with all achievement variables. Research question 9： What is the relationship of student characteristic variables and family-oriented variables in the achievement of students？ Also， which variable， Global Self-Concept or Academic Self-Concept serves as a better predictor of average academic score， English score and math score respectively？ Academic Self-Concept was the only significant predictor for average academic score and math score and significantly accounted for around 20% of the variance in the two respective achievement scores. Academic Self-Concept and Parent Education stayed as the only significant predictors for English score and significantly explained about 15% of the variance in English score. Global Self-Concept had no say in predicting three achievement scores. Research question 10： Do any of the variables predict Student Attitudes Towards School？ Also， what is the relationship between Parental Attitudes Towards School and Parent Education？ Student Study Habits and Parent/Home Self-Concept positively predicted Student Attitudes Towards School. Parent Education was the only significant predictor for Parental Attitudes Towards School and also negatively predicted it.

SUMMARY AND FINDINGS

In the section below， key findings from this study will be shared.

ANOVA tests revealed that the average academic scores and math scores of the very low， low and below-average Global Self-Esteem Quotient （GSEQ） groups were significantly lower than those of the above-average group. Also， the scores in English were significantly lower for the low GSEQ group compared to the above-average group. All three achievement scores （average academic， English and math） were significantly lower for the low GSEQ group compared to the average GSEQ group. Results from ANOVA analyses showed for Academic Self-Concept， the average academic score， English score and math score of the average academic Self-Concept group were significantly lower than those of the above-average group. For General Self-Concept， the below-average General Self-Concept group achieved significantly lower scores in average academic achievement， English and math when compared to the above-average group. For Parent/Home Self-Concept， the average Parent/Home Self-Concept group achieved significantly lower scores in average academic achievement， English and math when compared to the above average-group. For Personal Self-Concept， the average academic score and math score of the low Personal Self-Concept were significantly lower than those of the above-average group. For Social Self-Concept， the scores in English were significantly lower for the below-average Social Self-Concept group compared to the low group.

For Parent Education and Achievement， The scores in English were significantly lower for the low Parent Education group compared to the high Parent Education group. However， no significant mean differences in English scores were found in the other group comparisons and no significant mean differences in academic achievement scores and math scores were found in any group comparisons.

ANOVA test results indicated that there were no significant mean differences in average academic achievement， English and math between the female group and the male group. ANOVA results showed that the female average Academic Self-Concept group achieved significantly lower scores in average academic achievement， English and math when compared to the female above-average Academic Self-Concept group. Also， the math score was significantly lower for the female average Academic Self-Concept group compared to both the male average and above-average groups.

ANOVA results indicated that the scores on Parental Attitudes Towards School for the some High School group， High School Graduation group and Some University/College group were significantly higher than that of the Graduate Education And Above group. The data analyses results showed that Parental Attitudes Towards School was moderately correlated with Parent Education only and in a negative way （r=-.399， ɑ=.001）. Also， Student Study Habits was moderately correlated with Student Attitudes Towards School only and in a positive way（r=.471， ɑ=.000）.

Academic Self-Concept was most closely correlated with average academic achievement （r=.441， ɑ=.000）， English （r=.315， ɑ=.008） and math （r=.482， ɑ=.000） respectively than any of the other subscales of the CFSEI-3. When all the variables were ranked in descending order of correlation for the Z-scores and 100 point scores， Academic Self-Concept， followed by Global Self-Concept， were the highest ranking variables for the average academic scores and math scores. Moderate correlation of Parent Education with General Self-Concept， Social Self-Concept and Global Self-Concept measured in Z-scores were found， Student Attitudes Towards School was most closely correlated with Student Study Habits. Only Parent/Home Self-Concept， Academic Self-Concept and Global Self-Concept were found to be correlated with Student Attitudes Towards School.

Academic Self-Concept was the only significant predictor of average academic achievement. Academic Self-Concept and Parent Education were the only significant predictors of English achievement. Also， Academic Self-Concept was the only significant predictor of math achievement. Student Study Habits， Parent/Home Self-Concept plus Personal Self-Concept were significant predictors of Student Attitudes Towards School. Parent education acted as the only significant negative predictor of Parental Attitudes Towards School and significantly explained a small amount of variance in Parental Attitudes Towards School.

DISCUSSION

Significant mean differences in some or all three achievement scores were found among certain group comparisons for Global Self-Esteem Quotient. The findings of this study established that global self-concept had an impact on the achievement variables. Of the CFSEI-3 subscales， patterns among Academic Self-Concept and General Self-Concept with the new below average groups scoring significantly lower on the average academic， English and math achievement tests than the new above-average groups were found. Academic Self-Concept， Global Self-Concept and Parent/Home Self-Concept were most closely correlated with the achievement variables， followed by General Self-Concept， and then by Personal Self-Concept.

In regards to Parent Education and Achievement， the only significant finding was that the English score of the low Parent Education group was significantly lower than that of the high Parent Education group. The analyses showed that Parental Education was only correlated with the English score， but not with the average academic score and math score. The final results showed that no significant mean differences in average academic scores， English scores and math scores were found between female group and male group. This tells us that gender had no effect on the achievement variables.

The research results showed that Parental Attitudes Towards School was not susceptible to the student related variables such as Student Attitudes Towards School or dimensions of students self-concept. Parent Education was the only significant predictor of Parental Attitudes Towards School. Previous research indicated that Academic Self-Concept was more closely associated with academic achievement than was overall self-concept （Marsh 1990b）. This finding was again confirmed by this study as Academic Self-Concept enjoyed the highest correlation of all with three respective achievement measures. The research results showed that each CFSEI-3 subscale variable was more closely correlated with Global Self-Concept than with any other one of the CFSEI-3 subscale variables. This proves that self-concept consists of multi-dimensional subsets and each subset is relatively independent of the others （Bracken， 1996）.

Parent Education was not correlated with Academic Self-Concept （r=.117， ɑ=.347， 2-tailed）， plus the moderate correlation between Parent Education and English （r=.238， ɑ=.020， 2-tailed）， Parent Education plays a role as one of the predictors for English in conjunction with Academic Self-Concept. Student Attitudes Towards School was not correlated with average academic achievement， English and math. This contradicted the findings indicating negative student attitudes towards school was associated with underachievement in school （Reis & McCoach， 2000）.

Academic Self-Concept was top listed on each ranking list of predictors for each achievement variable. This is not surprising as Bloom （1976） also ranked Academic Self-Concept the strongest non-cognitive variable of all for predicting academic achievement. Global Self-Concept failed to predict either one of the achievement variables due to the results that it was also most highly correlated with all 5 CFSEI-3 subscale variables.

From the analysis results， the researcher concludes that only Parent/Home Self-Concept and Student Study Habits should be counted as strong predictors of Student Attitudes Towards School. It was interesting to note that parental attitudes towards school was not susceptible to the changes in the other variables， such as student achievement scores.

Based on the researching findings above， the researcher puts forward some recommendations for schools as follows. First， schools may run more sessions on the enhancement of Academic Self-Concept65 in combination with extra academic help to boost student achievement levels as Academic Self-Concept was a significant positive predictor for the achievement variables. Second， schools may want to add more Academic Self-Concept enhancement programs and activities specifically targeted to females as the female achievement levels were susceptible to how they feel about themselves academically. Third， schools may encourage the less educated parents to be more actively involved in the school related affairs so that they can contribute to the continued growth of the school system. Fourth， more resources and efforts should be shifted to the betterment of math achievement rather than to the betterment of English achievement if schools intend to improve the overall achievement level of students since the average academic score was more closely correlated with math achievement score than with English achievement score. Fifth， although parent education also had a share in predicting English scores， it did not predict the average academic score and math score. The results may imply that either the parents with high education levels had a good command of English， but not of math， or were more adept at helping with the language study of their children. If this is the case， the school may deliver off-school programs on math skills to interested students if their parents were not much of help with their study of math. Sixth， Student Study Habits and Parental/Home Self-Concept were identified as two strong predictors of Student Attitudes Towards School， which was moderately correlated with Academic Self-Concept， a strong predictor of the student achievement variables. Therefore， schools might set up workshops on developing proper study habits. Meanwhile， parents could frequently communicate with their children to create a more comfortable home environment in which both sides feel free to clear up any confusion. Improvement in performance on Student Study Habits and Parental/Home Self-Concept measuring scales could strengthen the relationship between these variables and scores on Student Attitudes Towards School.