The paper aims to verify the results of various studies conducted in different parts of the World about the academic self concept of secondary school students, specially keeping the comparison of academic self-concept among male and female students’. The literature shows that boys have more confidence in their own abilities than girls in many areas, including the evaluation of their own academic capacities, or ‘academic self-concept’ (Colwill, 1982). Keeping this result in mind I want to verify the outcomes in my context as I understand that research outcomes can be differed context to context. While literature indicate that boys are likely to rate their abilities more highly than girls in subjects that are perceived as ‘masculine’ such as mathematics (Joffe and Foxman, 1988, Marsh, 1989, Marsh and Yeung, 1998, Wilgenbusch and Merrell, 1999).

Rationale

There can be certain reasons that convinced me to conduct this study but following is the most driving factor for this research.

1. I have observed many students (boys & girls) during my teaching career, having multiple capacities in Math, Science and other subjects. Sometimes female on high capacities and sometime male on high capacities. These capacities vary in both genders, which the research papers also indicate that female students are simply not good at Math and Science as compared to male students. This notion could be an imported phenomenon and can be a sweeping statement without testing it in our context. It is purely not tested in our context yet. So, why not to test for verification?

Significance of the Study

This study initially will help ‘me’ to understand the differences among male and female about academic self-concept, which will contribute to improve my leadership practices specifically planning in future, keeping my own context. These findings will also help management and teachers to understand the level of academic self-concept in secondary school students.

There have been mixed findings on the effect of students’ age on the gender gap. Some studies have found that the gender gap in self-concept increases during teenage years, while others have found that the gap is either constant or reduces somewhat during the secondary school years (Eccles, 1987, Jacobs, et al., 2002). My research will show a facet of testing phase in Pakistan to verify and validate these findings.

Literature Review

Self Concept

By self, we generally mean the conscious reflection of one's own being or identity, as an entity separate from other or from the environment. There are a variety of ways to think about the self. There are two widely used terms; self-concept and self-esteem. Self-concept, which is the focus of this research paper, is the cognitive or thinking aspect of self and generally refers to the learned beliefs, attitudes and opinions that each person holds to be true about his or her personal existence (Purkey, 1988). Franken (1994) states that "there is a great deal of research which shows that the self-concept is, perhaps, the basis for all motivated behaviors. It is the self-concept that gives rise to possible selves and creates the motivation for behavior (p. 443).

I understand that academic self-concept is an important aspect in education as the importance of self-perception for the growth and development of children has been demonstrated in studies showing how self-efficacy can enhance or harm the level of cognitive functioning and performance (Bandura, 1989). Byrne (1984) has also concluded that ‘self-concept’ is a multidimensional construct, having one general facet and several specific facets, one of which is ‘academic self-concept’. The term ‘academic self-concept’ can be characterized by two elements consistent with the Shavelson model (Strein, 1993). First, academic self-concept reflects descriptive (e.g., I like math) as well as evaluative (e.g. I am good at math) aspects of self-perception. This survey is focusing on this aspect of self-concept. Second, self-perceptions associated with academic self-concept tend to focus on scholastic competence, rather than attitudes.

Self Concept – Male V Female

Boys have more confidence in their own abilities than girls in many areas, including the evaluation of their own academic abilities, or ‘academic self-concept’ (Colwill, 1982). Boys are usually rating their abilities more highly than girls in subjects that are perceived as ‘masculine’ such as mathematics (Joffe and Foxman, 1988, Marsh, 1989, Marsh and Yeung, 1998, Wilgenbusch and Merrell, 1999). Some studies have found that the gender gap in self-concept increases during teenage, while others have found that the gap is either constant or reduces somewhat during the secondary school years (Eccles, 1987, Jacobs, et al., 2002).

Jonsson (1999) has used a concept of ‘relative advantage’ to examine girls’ lower rates of participation in math and sciences, despite roughly equal levels of prior attainment in math. The point is that girls’ high attainment in other subjects could prevent them from pursuing math and science, even though their math and science attainment is equal to that of their male peers.

Statement of hypotheses/research question

Through literature review and my own understanding regarding academic self-concept following research question emergies, because it is expected to find boys having higher academic self concepts in mathematics, and science than girls. So the main question of this research is, “what is the difference in academic self concept of boys and girls at secondary school level about mathematics, science and other subjects?”

RESEARCH DESIGN

Methodology

I choose the quantitative research paradigm and survey design for my study as it is capable to address my research question which requires studying the academic self-concept of two gender based groups. Because survey designs are used to gain data on attitudes (May, 2003) and deals more directly with the nature of people’s thoughts, opinions and perceptions (Shanghnessy & Zecheister 1997). The mode of administration of this study was a direct administered session with an adopted questionnaire of Marsh, 2007. Within the survey design, cross sectional survey was used because collecting data at a single point in time within limited time was sufficient to answer the research questions (Kosecoff, 1999).

Setting

I conducted this survey in two private schools of Karachi. The schools were categorized as single gender-schooling. One school dedicated for male while another for female only.

Research Participants

The key to good sampling is finding ways to give all population members an equal chance of being selected and one of these is to use the probability methods for choosing a sample (Fowler, 2001) while for this research, the participants were the convenient samples as convenience sample is one that is simply available to the researcher by virtue of its accessibility (Bryman, 2004) and stratified as male and female of class 8^th between the ages of 13 to 16 with mean age 14 years. Total 54 students were taken as samples (Boys 23 = 42.6%, girls 31 = 57.4%).

Procedures

The questionnaire, I adapted was consisted of 18 items which corresponded to three sub categories of self-concept about Math, Science and other subjects. The questionnaire (Marsh, 2007), which was divided in four portions i.e. demographic data (age, class, school & gender), Science Self Concept, Mathematics Self Concept and finally other subjects self-concept. The questionnaire was having an 8 rating scale tool starting from ‘Very False’ towards ‘Very True’. Each section of self concept was having 6 statements except the demographic section. The questionnaire had a simple and clear wording according to the level of my research participants, as Robinson (2002) states that, “a good questionnaire not only provides a valid measure of the research question but also gets the cooperation of the respondents, and illicit accurate information” (p.242).

It was a direct administration to all the students in their science classroom. In the class, first of all, I introduced myself and explained briefly the title and the purpose. I told them that I was also a student like them which seemed to put them at ease. I also stated that I was interested in learning about their self concept about Math, Science and other subjects. In this way, I was also able to engage students right away which helped me to maximize my response rate. I emphasized that as a fellow student I want their own responses to the statements, because it was not a test so there were no right or wrong answers but honest opinions. For the convenience and better understanding of students I had elaborated the rating scale on a separate chart which was displayed in the class for handy reference.

The direct administration of the questionnaire helped me to get the information immediately with a highest response rate (100%). The questionnaires were completed by students in half an hour; hence, this mode of administration was cost effective in terms of time. At the end of the class I thanked all the students for their participation as well as their teachers, especially head teacher, for granting permission.

Ethical Consideration

Researchers must pay attention to the ethical principals as the term research ethics indicates a moral enterprise between the researcher and the research participants (Vazir, 2004). It refers to the question of right or wrong, and it is conforming to the standards of conduct specified in research (Fraenkel & Wallen, 2006).

So, in terms of ethical consideration, our worthy facilitators had taken the permission from the school administration about conduction of my survey and students were also informed, while during administration of the questionnaire, I also took formal permission to conduct the survey from the class teacher and students. Therefore I didn’t face any problem in this regard. I had informed the research participants about the purpose of the research and expectations from them.

I also assured my respondents that the information provided by them would be used for the study purposes only. The privacy of the records would be maintained and no one would have access to them before analyzing and report-writing.

DATA ANALYSIS

Data Entry

All the responses collected from 53 students were fed into SPSS[1], which according to Gaur & Gaur (2006) is the most reliable software for analyzing quantitative data. I created a database by defining the variable, attaching labels and entering the pre-coded responses. The code for missing data was ’99’. But luckily there was not a single missing response.

Data cleaning and Analysis

After data entry, data cleaning is as important as the need to proof-read text for errors (Robson, 2002). I carried out data cleaning process by checking the data for errors which could have been made while keying-in the data. After simple eye-balling of the data for any visible abnormality, I conducted descriptive statistics analysis to explore each variable separately, while for data analysis, first of all I had to convert three question responses as all the questions of the questionnaire were in positive sense, while three questions were in negative, so using the convert command in SPSS the answers were converted to ensure a reliable result. Then to check the percentage and mean value of participants’ age and gender, descriptive statistics and frequency was used to analyze the data. To check the validity of all entries, whether the entries are with in the specified range of rating scale or not, I again checked the entire variables in the frequency table, which showed an accurate filling of the questionnaire. Then mean was calculated in groups like mean score of Science, Math and other subjects, which gave a big picture of data and its dispersion verification.

To test the normal distribution of the data, I used KS[2] test, which verified that data is accurate and it falls in normal distribution for comparison between male and female students according to their academic self concept, so now it was clear to have a comparison of both starts and to do so, I used t-test to determine the significant difference. After performing t-test, I analyzed the data to get findings.

Findings

Result of KS test

The summary in Table 1 below shows that in all the cases difference is not significant which means that the data falls in normal distribution.

TABLE 1:

Tests of Normality - KS Test

	Gender	Kolmogorov-Smirnov^a			Shapiro-Wilk
	Gender	Statistic	Df	Sig.	Statistic	df	Sig.
Mean Science	Boy	0.11	23	0.20^*	0.96	23	.676
Mean Science	Girl	0.11	31	0.20^*	0.94	31	.131
Mean Math	Boy	0.15	23	0.14	0.95	23	.449
Mean Math	Girl	0.14	31	0.09	0.94	31	.082
Mean Other sub	Boy	0.15	23	0.16	0.94	23	.194
Mean Other sub	Girl	0.14	31	0.11	0.95	31	.179
a. Lilliefors Significance Correction
*. This is a lower bound of the true significance.

Result of descriptive statistics and T-Test

TABLE 2: Group Descriptive Statistics Group Statistics
	Gender	N	Mean	Std. Deviation	Results of t-test
Mean Science	Boy	23	5.95	1.05	[t(52)=- 0.001; P= 0.99]
Mean Science	Girl	31	5.95	1.48	[t(52)=- 0.001; P= 0.99]
Mean Math	Boy	23	5.61	1.35	[t(52)=-0.72; P=0.47]
Mean Math	Girl	31	5.90	1.51	[t(52)=-0.72; P=0.47]
Mean Other Sub	Boy	23	5.71	1.33	[t(52)=-2.76; P=0.008]
Mean Other Sub	Girl	31	6.54	.880	[t(52)=-2.76; P=0.008]

The above table shows that on the average the self concept of girls in all subjects is slightly higher (M=6.54, SD=.88) than boys (M=5.71, SD=1.33). The difference is significant. The table also shows that self concept of girls in math is also slightly higher (M=5.90, SD=1.51) than boys (M=6.61, SD=1.35) however, the difference is not significant. Moreover, the table also illustrates that self concept of girls and boys in Science is almost same (M=5.95, SD=1.48 and M=5.95, SD=1.05 respectively). The difference is not significant.

The result surprisingly differ with the findings of researches done in different places, (which I have quoted earlier) which show that boys have the high academic self-concept than girls, but my research shows a different story. This survey shows that there is no any major difference among male and female in the group statistics, although female students have shown slightly higher difference in Science, Math and other subjects as compared to male. Science Mean (Male = 5.95, Female = 5.95) Math Mean (Male = 5.61, Female = 5.90) other subjects mean (Male = 5.71, Female = 6.54).

FIGURE 2:

Gender wise Science Mean

FIGURE 3:

Gender wise Math Mean

FIGURE 4:

Gender wise Other Subjects mean

The above Block Diagrams in Figure 2, 3 & 4 show the result of gender wise academic self concept in Science, Mathematics and all other subjects. In Science there is no any outlier, and the median in girls’ data lies above six on the scale, while boys’ median lies below point six on the Likert scale. Likewise in Mathematics two boys (case 4 and 22) are outliers. These two outliers fall between 2 and 3 on the Likert Scale, while boys are also bellow than girls. Finally the result of other subjects shows that one girl is outlier which is case 30. This outlier lies at point 4 on Likert scale and median of girls is higher than boys.

DISCUSSION AND Conclusion

This is in line with my expectation that boys would have higher self-concepts than girls, both in general and in stereotypically ‘masculine’ subjects, such as math keeping the researches and their findings in mind. I was expecting continuity in my findings, but now it is a different story. So based on the findings of this research I would urge that the earlier research results which have shown an another story than mine may be caused by some reasons like, context, single gender-schooling, educational environment, syllabus, basic education, facilities provided to students, because these are the basic elements to raise or to drop down the self conception among students. It is very much evident that boys were having a high self concept in Math and Science while girls were having high self concept in other subjects. My research shows that in other subjects girls are still on high position, while they are getting towards high conception in math and Science as well.

REFERENCES

Bandura, A. (1989). Self-efficacy: the exercise of control. New York: Freeman.

Bryman, A. (2004). Social Research Methodsl. (2^nd ed.). New York: Oxford University

Byrne, B. (1984). The general/academic self-concept nomological network: A review of construct validation research. Review of Educational Research, 54, 427-456. Callingham,

Eccles, J. S. (1987) Gender roles and achievement patterns: An expectancy value perspective, in: Reinish, J. M., Rosenbaum, L. A. and Saunders, S. A. Masculinity/femininity: Basic Perspectives (New York, Oxford University Press).

Franken, R. (1994). Human motivation (3rd ed.). Pacific Grove, CA: Brooks/Cole Publishing Co.

Jacobs, J. E., Lanza, S., Osgood, D. W., Eccles, J. S. and Wigfield, A. (2002) Changes in children's self-competence and values: Gender and domain differences across grades one through twelve, Child Development 73. 509-27.

Joffe, L. and Foxman, D. (1988) Attitudes and Gender Differences (Slough, NFER-Nelson).

Jonsson, J. O. (1999) Explaining Sex Differences in Educational Choice: An Empirical Assessment of a Rational Choice Model, European Sociological Review 15, (4). 391-404.

Marsh, H. W., Craven, R. G. & Debus, R.. L. (1998). Structure, stability, and development of young children’s self-concepts: A multicohort-multioccasion study. Child Development, 69, 1030 –1053.

May, T. (2003).Social research: issues, methods and process. (2^nd ed.). Buckingham: Open University Press

-Purkey, W. (1988). An overview of self-concept theory for counselors. ERIC Clearinghouse on Counseling and Personnel Services, Ann Arbor, Mich. (An ERIC/CAPS Digest: ED304630)

Strein, W. (1993). Advances in research on academic self-concept: Implications for school psychology. School Psychology Review, 22, 273-284.

Vazir, N. (2004). Research ethics, significance, application and obligation to the practice of research. Journal of Education, 7(1, 2) 3-11.

Wilgenbusch, T. and Merrell, K. W. (1999) Gender differences in self-concept among children and adolescents: A meta-analysis of multidimensional studies, School Psychology Quarterly 14, (2). 101-20.

Shanghnessy, J.J & Zecheister, E.B. (1997). Research methods in Physiology. Singapore: Pan Publishers

Robinson, C. (2002). The real world research (2nd ed.). Malden: Blackwell Publishing.

APPENDIXES

Tests of Normality
	Are you a boy or Girl?	Kolmogorov-Smirnov^a			Shapiro-Wilk
	Are you a boy or Girl?	Statistic	df	Sig.	Statistic	df	Sig.
Mean_Science	Boy	.112	23	.200^*	.969	23	.676
Mean_Science	Girl	.112	31	.200^*	.947	31	.131
Mean_Maths	Boy	.157	23	.149	.959	23	.449
Mean_Maths	Girl	.145	31	.096	.940	31	.082
Mean_All_Sub	Boy	.155	23	.162	.942	23	.194
Mean_All_Sub	Girl	.141	31	.119	.952	31	.179
a. Lilliefors Significance Correction
*. This is a lower bound of the true significance.

One-Sample Kolmogorov-Smirnov Test
		Mean_Science	Mean_Maths	Mean_All_Sub	Are you a boy or Girl?
N		54	54	54	54
Normal Parameters^a	Mean	5.9568	5.7809	6.1883	1.57
Normal Parameters^a	Std. Deviation	1.30721	1.44095	1.16196	.499
Most Extreme Differences	Absolute	.065	.106	.122	.377
	Positive	.059	.062	.063	.301
	Negative	-.065	-.106	-.122	-.377
Kolmogorov-Smirnov Z		.480	.776	.898	2.773
Asymp. Sig. (2-tailed)		.975	.584	.395	.000
a. Test distribution is Normal.

Test Statistics
		Mean_Maths
Mann-Whitney U		306.000
Wilcoxon W		582.000
Z		-.884
Asymp. Sig. (2-tailed)		.376
a. Grouping Variable: Are you a boy or Girl?
Descriptive
	Are you a boy or Girl?				Statistic	Std. Error
Mean_Science	Boy		Mean		5.9565	.22060
			95% Confidence Interval for Mean	Lower Bound	5.4990
				Upper Bound	6.4140
			5% Trimmed Mean		5.9654
			Median		5.8333
			Variance		1.119
			Std. Deviation		1.05794
			Minimum		3.83
			Maximum		7.83
			Range		4.00
			Interquartile Range		1.83
			Skewness		.112	.481
			Kurtosis		-.627	.935
	Girl		Mean		5.9570	.26628
			95% Confidence Interval for Mean	Lower Bound	5.4132
				Upper Bound	6.5008
			5% Trimmed Mean		6.0137
			Median		6.1667
			Variance		2.198
			Std. Deviation		1.48260
			Minimum		2.83
			Maximum		8.00
			Range		5.17
			Interquartile Range		2.17
			Skewness		-.543	.421
			Kurtosis		-.578	.821
Mean_Maths	Boy		Mean		5.6159	.28280
			95% Confidence Interval for Mean	Lower Bound	5.0294
				Upper Bound	6.2024
			5% Trimmed Mean		5.6590
			Median		5.8333
			Variance		1.839
			Std. Deviation		1.35627
			Minimum		2.50
			Maximum		8.00
			Range		5.50
			Interquartile Range		1.67
			Skewness		-.671	.481
			Kurtosis		.399	.935
	Girl		Mean		5.9032	.27137
			95% Confidence Interval for Mean	Lower Bound	5.3490
				Upper Bound	6.4574
			5% Trimmed Mean		5.9633
			Median		6.3333
			Variance		2.283
			Std. Deviation		1.51093
			Minimum		2.67
			Maximum		8.00
			Range		5.33
			Interquartile Range		2.67
			Skewness		-.492	.421
			Kurtosis		-.709	.821
Mean_All_Sub	Boy		Mean		5.7101	.27806
			95% Confidence Interval for Mean	Lower Bound	5.1335
				Upper Bound	6.2868
			5% Trimmed Mean		5.7725
			Median		5.6667
			Variance		1.778
			Std. Deviation		1.33354
			Minimum		2.33
			Maximum		7.83
			Range		5.50
			Interquartile Range		2.33
			Skewness		-.549	.481
			Kurtosis		.135	.935
	Girl		Mean		6.5430	.15820
			95% Confidence Interval for Mean	Lower Bound	6.2199
				Upper Bound	6.8661
			5% Trimmed Mean		6.5842
			Median		6.3333
			Variance		.776
			Std. Deviation		.88083
			Minimum		4.17
			Maximum		8.00
			Range		3.83
			Interquartile Range		1.17
			Skewness		-.465	.421
			Kurtosis		.602	.821