Attitudes and Knowledge on University Sustainability: Variations among Students of the Educational Research and Sustainable Development MSc Programmes Table of Contents Abstract 1. Introduction 1 2. Operationalization of constructs 2 2.1 Knowledge of sustainability 3 2.2 Attitudes to sustainability 3 3. Hypotheses 4 4. Methodology 5 4.1 Research design 5 4.2 Participants 5 4.3 Instrument Design 6 4.4 Validity of the instruments 7 4.5 Reliability of the knowledge scale 7 4.6 Reliability of the attitude scale 8 4.7 Data collection 8
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Attitudes and Knowledge on University Sustainability: Variations among Students of the Educational Research and Sustainable
Development MSc Programmes
Table of Contents
Abstract1. Introduction 12. Operationalization of constructs 2
2.1 Knowledge of sustainability 32.2 Attitudes to sustainability 3
3. Hypotheses 44. Methodology 5
4.1 Research design 54.2 Participants 54.3 Instrument Design 64.4 Validity of the instruments 74.5 Reliability of the knowledge scale 74.6 Reliability of the attitude scale 84.7 Data collection 84.8 Limitations 94.9 Ethical issues 9
5. Presentation of results 96. Discussion 227. Conclusion 238. References 259. Appendix A: Item Total Statistics 2810. Appendix B: Histograms, Boxplots, Detrended Q-Q Plots 2911. Appendix C: Descriptives 3112. Appendix D: Cover Letter to Participants 3213. Appendix E: On-line Questionnaire 3314. Appendix F: Ethical Approval Form 35
Abstract
This study aims to explore the sustainability attitudes and knowledge of students from two different postgraduate programmes at the University of Exeter; namely, the MSc in Educational Research and the MSc in Sustainable Development. A questionnaire was administered and quantitative data was collected from 20 respondents. Responses were analysed using the statistical package SPSS21. The statistical analysis indicates a moderate correlation between attitudes and knowledge on university sustainability. The findings suggest that attitudes on sustainability were affected by the participants’ programme of study, while they did not appear to be influenced by the participants’ gender. The model predicting sustainability knowledge from a combination of attitudes on sustainability, programme of study and gender has reached statistical significance.
1. Introduction
The end of 2005-2014, which was defined by the United Nations as the Decade of Education for Sustainable Development finds Higher Education Institutions (HEIs) attempting to include sustainability in their curricula, policies and practice. ‘Greening’ the campus and the curriculum is dictated not only by governmental policies and funding councils, but also by students’ environmental sensitivity and prospective employers, who need to safeguard their corporate social responsibility profiles (Sterling, 2011).
As a result, higher education providers strive to improve their sustainability credentials and performance (Sterling, 2011). The University of Exeter is also committed to providing every student with the opportunity to learn about sustainability and promises to equip its graduates with skills for the green economy. It has participated in the Rio +20 conference and has signed the Rio +20 Treaty on Sustainable Development in Higher Education. Optional sustainability modules are now part of the curriculum across all disciplines while the university fosters world class sustainability research with all members of the teaching staff receiving training on sustainable development (University of Exeter, 2014)
While sustainability initiatives are blooming in most universities, the impact of this activity has not been formally estimated. Harraway et al. (2012) claim that universities generally lack formal processes to benchmark environmental performance and identify the impact it might have on student experience. Responding to this emerging need, the Higher Education Funding Council for England (HEFCE) commissioned in 2008 an assessment of sustainable development research, teaching, estates and corporate management at English universities. This review aimed at establishing a baseline in the sector, against which progress could be measured (Policy Studies Institute, 2008, p. vi).
1
A basic obstacle to this endeavor was the lack of a consistent and comprehensive definition of sustainability in higher education (Policy Studies Institute, 2008). Summers (2010) claims that given the inherent dilemmas surrounding sustainability, it is not surprising that the notion of education for sustainability is similarly open to debate. Sustainability in higher education (HE) is likewise subject to many interpretations as the practice it underlies is still evolving. In the literature there seems to be no single conceptualization of university sustainability (Kawaga, 2007; Moody and Hartel, 2007; Emanuel and Adams, 2011; Wachholz, Artzt and Chene, 2012). In 2013, Stephen Sterling attempts a comprehensive definition of university sustainability in the book The Sustainable University, which however has not yet been widely established.
Due to this lack of an established definition of sustainability in Higher Education (HE), the majority of the literature focuses on university students’ perceptions of sustainability in general (Zwickle et al., 2013; Stubbs and Cocklin, 2008) but not university sustainability in particular, as a phenomenon taking place within the time and space of HE. Gradually researchers are beginning to address certain aspects of university sustainability. For instance, Moody and Hartel (2007) examine students’ attitudes to sustainability in the university curriculum while Jones et al. (2013) explore students’ opinions of sustainability in university management. A three-year longitudinal study (2010-2012) by the National Union of Students and Higher Education Academy (Dryason, Bone and Agombar, 2012) addresses students’ attitudes to university sustainability. However, there seems to be lack of research on both the attitudes and knowledge of students on university sustainability holistically.
The current study aims to explore students’ attitudes and knowledge on the emerging field of university sustainability in terms of teaching, research, estates and corporate management, as specified by HEFCE. The survey focuses on students of two postgraduate programmes at the University of Exeter; the MSc in Educational Research and the MSc in Sustainable Development. The Sustainable Development students are by definition more exposed to sustainability discourse when compared with the Educational Research students. Previous literature presents programme of study as a determining variable of students’ sustainability attitudes (Ewert and Baker, 2001; Wysor, 1983 ; Harraway et al. 2012; Hodgkinson and Innes, 2001). The present survey thus hopes to identify whether the programme of study will have an impact on respondents’ sustainability attitudes.
Previous research has also highlighted gender as a predicting variable for sustainability attitudes. Most of the studies portray women as more pro-environmental than their male counterparts (Wachholz et al, 2014 ; Zelezny, Chua and Aldrich, 2010; Shephard et al., 2009) with only a few studies depicting men as more environmentally conscious (Arcury, Johnson and Scollay, 1986). Since gender had been extensively reported as a predictor of sustainability attitudes, it was decided that it would be included as a variable in the present study.
2
Therefore, the influence of gender and programme of study on students’ sustainability attitudes was considered worth-examining. The relationship between knowledge of and attitudes to sustainability was also explored. The outcome of the study could provide an additional insight in the way students respond to the emerging field of university sustainability. The study could also be contributive to the growing body of literature, which explores ways of enhancing university students’ sustainability literacy and consequently improving academic institutions’ sustainability performance.
2. Operationalization of the study’s constructs
The main constructs in this study are knowledge of university sustainability and attitudes to university sustainability. The instruments are operationalized as follows:
2.1. Knowledge of sustainable development
Knowledge of university sustainability was explored through 10 statements. Participants were given three options on a Likert scale (1 = agree, 2 = unsure, 3 = disagree). High scores indicated a high level of knowledge on university sustainability while low scores indicated a low level of knowledge on the issue.
2.2. Attitudes towards sustainable development
Attitudes towards sustainability were tested through 10 statements. Participants were asked to state their agreement on a 5 point-Likert scale (1 = strongly agree, 2 = agree, 3 = neutral, 4 = disagree, 5 = strongly disagree). High scores indicated positive attitudes towards sustainability whereas low scores indicated a less positive attitude towards sustainability.
Background information on the programme of study and gender was required in the last two items of the questionnaire.
3. The hypotheses
Hypothesis 1 The relationship between knowledge of and attitudes towards university sustainability.
H1: There will be a significant correlation between the respondents’ knowledge of and attitudes towards university sustainability.
H01: There will be no significant correlation between the respondent’s knowledge of and attitudes towards university sustainability.
3
Independent Variables 1. Knowledge of university sustainability2. Attitudes towards university sustainability
Statistical Test Pearson product-moment correlation coefficient if data meet parametric requirements(Alternatively, Spearman rank order correlation)
Hypothesis 2 The effect of respondents’ programme of study on attitudes to university sustainability.
H2: There will be a significant difference in attitudes towards university sustainability between the Sustainable Development and Educational Research students.
H02: There will be no significant difference in attitudes towards university sustainability between Sustainable Development and Educational Research students.
Independent Variables Programme of study (MSc in Educational Research and MSc in Sustainability)
Dependent Variables Attitudes towards university sustainability
Statistical TestIndependent-samples t-test if data meetparametric requirements(Alternatively, Mann -Whitney test )
Hypothesis 3 The effect of respondents’ gender on attitudes towards university sustainability.
H3. There will be a significant difference in attitudes towards university sustainability between male and female respondents.
H03. There will be no significant difference in attitudes towards university sustainability between male and female respondents.
4
Independent Variables Gender (male or female)
Dependent Variables Attitudes towards university sustainability
Statistical TestIndependent-samples t-test if data meetparametric requirements(Alternatively, Mann -Whitney test)
Hypothesis 4 The prediction of knowledge on sustainability from a combination of information on attitudes towards sustainability on gender and programme of study.
H4: Knowledge on sustainability will be predictable from a combination of information on attitudes towards university sustainability, gender and programme of study.
H04: Knowledge on sustainability will not be predictable from combination of information on attitudes towards university sustainability, gender and programme of study.
Independent Variables1. Attitudes towards sustainability2. Gender3. Programme of Study
Dependent Variables Knowledge on university sustainability
Statistical Test Standard multiple regression(No non-parametric alternative)
4. Methodology
This section offers information on research design, the study participants, the instrument developed, the data collection process as well as on limitations and ethical considerations.
4.1 Research Design
The survey research method is chosen for this study as it provides standardized measurement consistent across all respondents, enabling statistical analysis of the data collected (Fowler, 2009). Four hypotheses were formulated consulting previous literature in the field. A questionnaire was administered to collect data from the sample. The responses were be coded, and negatively worded items were reversed
5
before entering the data in the statistical pakage SPSS21. Preliminary analyses are conducted to check parametric assumptions. The reliability of the measuring instrument is checked using Cronbach’s alpha. Statistical tests are used as outlined in the hypotheses section. If data meet parametric requirements, the statistical methods to be used include Pearson product-moment correlation coefficient, independent samples t-tests and standard multiple regression. The outcome of the analysis is discussed prior to the conclusion.
4.2. Participants
Even though random sampling is considered ideal for surveys, this small scale enquiry utilizes a convenience sample due to accessibility limitations. The population of the study comprises of students from two different postgraduate taught programmes at the University of Exeter. The first programme is the MSc in Educational Research where students receive training in social science research methods. As the focus of the course is on research techniques, sustainability is not a part of the curriculum. The second programme is the MSc in Sustainable Development. As the title indicates, there is full immersion of students into the discourse of sustainability, which is achieved through a holistic and multidisciplinary approach.
The sample of the study consisted of 20 participants; 12 (65%) from the Educational Research MSc and 8 (35%) from the Sustainable Development MSc 2013/14 (Figure 1). Gender is equally distributed in the sample, with 50% of the respondents being female and 50% male (Figure 2).
Figure 1 Pie chart for MSc programme Figure 2 Pie chart for gender
6
4.4. Instrumentation
An on-line questionnaire was considered as the most suitable instrument for this survey as I was allowed (after asking for permission) to the group emails of the students attending the two postgraduate programmes. Moreover, the questionnaire was preferred as it provides a ‘rapid turnaround in data collection’ (Creswell, 146). Finally, related studies measuring sustainability in higher education followed a similar approach (Shephard, 2009; Harraway et al., 2012).
The instrument included 10 items measuring attitudes towards sustainability, 10 items measuring knowledge towards sustainability and 2 items selecting information on gender and programme of study. The respondents were asked to state their level of agreement or disagreement with the statements using a 5-point and a 3-point Likert scale.
The questionnaire was designed with the aim of maximizing response rate. According to Muijs, this can be achieved by keeping the questionnaire relatively short and
7
attractive (2004, p.43). A coloured theme was applied to make it more appealing while only 22 items were included with questionnaire brevity in mind. To enhance readability, items were kept as short as possible consisting of less than 20 words each. As the ordering of the items can affect motivation, initial questions should ‘impose minimal respondent burdens’ and ‘build rapport’ (Krosnick and Presser, 2010, p. 291). Therefore, the attitude scale was placed first as the use of the first person singular encourages participant identification and no right or wrong answers are entailed in the attitude statements, which makes them less stressful. The knowledge scale was placed next while background information items were placed at the end of the questionnaire, as they require less mental effort to fill in and are less likely to suffer from fatigue effects.
A Likert scale was chosen for both the knowledge and attitude scale responses, as it tends to perform well for a reliable and rough ordering of respondents with regard to a specific attitude (Oppenheim, 1992, p.141). The 5-point scale used for the attitude section was changed into a 3-point scale at the knowledge section in order to encourage respondent motivation in this more demanding part of the questionnaire. As Krosnick and Presser (2010) support each added point on a scale requires an additional interpretation a person must make. However, instead of the brief agree-disagree option, a mid-point was added to provide for respondents who genuinely didn’t know the answer. Since the discourse surrounding university sustainability is still new, an increased number of neutral responses was anticipated. Moreover, according to O’Muircheartaigh et al. (1999) ‘adding midpoints to rating scales improve[s] the reliability and validity of ratings’ (cited in Krosnick and Presser, 2010, p. 274).
Attention was given to a balanced combination of pro- and anti- sustainability items, while negative statements were kept to a minimum. Leading questions were avoided. Nevertheless, in order to stimulate intuitive responses, words that evoke feelings and emotions like ‘admire’ and ‘highly value’ appear in the attitude section in accordance with Oppenheim’s suggestion. (1992, p. 114).
4.5. Validity of the instrument
The content validity of the instrument was addressed through an extensive search of the literature, including examination of journal articles, policy documents and publications on sustainability in higher education. Particularly, documentation from the Higher Education Funding Council for England (HEFCE) on university sustainability was used as a basis for the knowledge scale items. While research by the Higher Education Academy on student attitudes to sustainability served as a source for the attitude scale statements.
Apart from an extensive literature search, content validity can be also be tested through professional judgments about the representativeness of an instrument (Cohen, Manion and Morrison, 2007). The opinion of experts in the field of university sustainability was sought and positive feedback was received regarding its coverage. A suggestion was made to change the answer format of the knowledge scale which
8
initially comprised of an agree/disagree option. A 3-point Likert- type scale format was adopted instead. One item was identified as a leading question and the wording was changed accordingly. The term ‘environmental’ in one of the questions was identified as problematic and was replaced by ‘sustainable’.
The face validity of the questionnaire was also checked by asking respondents whether the instrument looks valid to them (Muijs, 2004). A pilot test was carried out with four postgraduate students and the wording of two statements which were identified as ambiguous was altered.
4.6. Reliability of the knowledge scale
Reliability shows if an instrument can be interpreted consistently in various situations (Field, 2013). An aspect of reliability that can be assessed is internal consistency, which checks whether all the items are measuring the same attribute. The usual way to measure internal consistency is Cronbach’s alpha, a statistic that calculates the average correlation among all the scale items (Pallant, 2010). As seen from Table 1 the Cronbach’s alpha coefficient for the knowledge scale was .69 which is very close to the value of .7, set as a criterion for reliability. According to Pallant (2010), Cronbach’s alpha values are quite sensitive to the number of items in a scale, with small scales (ten items or less) yielding low Cronbach’s alpha values. This happens as the top part of the equation for alpha includes the number of items squared (Field, 2013). Therefore, as the number of the items on a scale increases, the alpha will increase. It is possible to get a large value of alpha because you have a lot of items on a scale, not because your scale is reliable (Field, 2013). Conversely, it is possible to get a relatively low value of alpha because there are a few items on a scale, not because the scale is unreliable. Bearing in mind that the current scale consists of only ten items, the value of .69 may not indicate lack of reliability.
Checking the Corrected Item-Total Correlation column, there were some values below three (Appendix A, Figure 1). Yet, if deleted they would not dramatically increase the Cronbach’s alpha and were thus retained.
Table 1: Knowledge of Sustainability Scale
Reliability Statistics
Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items
.688 .703 10
4.7. Reliability of the attitude scale
The Cronbach’s alpha for the attitude scale was .83 (Table 2), which suggests good internal consistency reliability. In the Corrected Item-Total Correlations column (Appendix A, Figure 2) there was one item below three, which however would not improve the value of Cronbach’s Alpha if deleted and was thus retained.
9
Table 2: Attitude to Sustainability Scale
Reliability Statistics
Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items
.828 .842 10
4.8. Data collection
Created on Google drive, an on-line file storage and synchronization service, the questionnaire was sent to the group emails of the Sustainable Development and Educational Research postgraduate programmes at the University of Exeter. It was administered on-line on the 20th of June 2014 by an administrative employee. A cover letter (Appendix C) accompanied the questionnaire where respondents were offered information on the purpose and structure of the instrument. Participants submitted their responses electronically and all answers were recorded by Google drive. The answers were automatically summarized on a Google spreadsheet, which was used for the subsequent coding of the data. The emails of the respondents would not appear on the Drive, protecting the anonymity of the participants. The questionnaire was sent to the 49 students of the two MSc programmes and 20 returned it, which gives a response rate of 41 %.
4.9. Limitations
One of the limitations of the study was the convenience sample chosen for the collection of the data, which did not allow generalizability of the findings to the wider population of the postgraduate students at the University of Exeter. The relatively low response rate ( 41% ) was also a limitation as it could incur ‘response bias’, which is the effect of non-responses on survey estimates (Fowler, 2002). Finally, additional qualitative data might have contributed to a more comprehensive exploration of students’ responses to university sustainability.
4.9. Ethical Issues
The ESRC Framework for Research Ethics provides basic principles the researcher should comply with in order to conduct ethical research. The first one is that research should be designed to ensure integrity and quality. In the present study time was spent to develop a valid questionnaire and the best possible effort was made to analyse and interpret the data using appropriate methods. Another principle is that research staff and subjects are fully informed about the purpose and methods of the research (ESRC, 2012, p.2). The module coordinator was informed by signing an ethical approval form prior to the data collection stage, which presented the aim, methodology and population of this study. Participants were made aware of the purpose and structure of the instrument in the cover letter (Appendix B) accompanying the on-line questionnaire.
The next important prerequisite of ethical research is confidentiality and anonymity. The on-line questionnaire in this study was sent to the group emails by an administrative employee, allowing the researcher no access to the identity of the
10
respondents. Returning the questionnaire on a public folder (Google drive) guaranteed anonymous treatment of all information, as the email addresses of respondents would not appear anywhere on the form. Participation in the survey was voluntary as the 41% response rate exemplifies, while the confidential and anonymous treatment of all information did not incur any harm to the participants. The last principle set out by the ESRC is the independence and impartiality of the research. In the current study this principle was safeguarded by the honest and impartial analysis and interpretation of the survey results.
5. Presentation of Results
Normality of the attitude score
An assessment of the normality of the data will determine whether the hypotheses will be checked using parametric or non-parametric statistics. Normality can be assessed graphically and numerically (Pallant, 2010). On first sight, the Histogram (Appendix B, Figure 1) did not exhibit a perfectly bell-shaped curve, which one can argue was due to the small bin sizes. Since the bin size width is inversely proportional to the sample size (Jenkinson, 2010) and bearing in mind that my sample consisted of only 20 cases, it could be argued that larger bin sizes than the ones proposed by SPSS (Appendix B, Figure 1) would be more suitable. As observed from the Histogram (Figure 3), when the bin size increased from 2 to 5 the bell shaped curve was far more prominent, reflecting the scores on the Stem-and-Leaf diagram (Figure 5). This also accorded with the Sig. values of the statistical tests which indicated normality. The Shapiro-Wilk result was consulted for this sample as it consisted of less than 50 respondents (Cole, 2013, p.24) and the Sig. value p = .88 > .05 indicated non-significance (Table 3). Moreover, a reasonably straight line could be claimed in the Normal Q-Q Plot (Figure 4), while no outliers appeared in the Boxplot (Appendix B, Figure 2).
Normal distribution of the scores can also be checked by looking at the skewness and kurtosis values (Appendix C, Figure 1), which are not further from zero. Field (2013) suggests converting these values into z values by dividing by their standard error and if the resulting scores are greater than 1.96 then they are significant. In this case the z scores were: for skewness z = .35 and for kurtosis z = .55. Both values were smaller than 1.96 which implies that the scores were normally distributed.
Table 3: Normality Test for Attitude towards Sustainability
Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
Total Attitude Score .103 20 .200* .976 20 .875
*. This is a lower bound of the true significance.
a. Lilliefors Significance Correction
Figure 3: Histogram for Attitude towards Sustainability
11
Figure 4: Normal Q-Q plot for Attitude towards Sustainability
The Stem-and-Leaf diagram for the Knowledge Score (Figure 8) presented a fairly normal distribution, which however was not reflected in the initial Histogram (Appendix B, Figure 4). Following the argument described in the previous section, the bin width was changed from 1 to 5 creating the bell shaped curve seen in the Histogram of Figure 6. This Histogram reflected the distribution of the Stem-and-Leaf diagram (Figure 8) and was aligned with the values distributed along the diagonal in the Normal Q-Q plot (Figure 7). Moreover, the Histogram (Figure 6) was in accordance with the Shapiro-Wilk test result p = .74 > .05, which indicated non-significance and normality (Table 4).
Moreover, the skewness and kurtosis values (Appendix C, Figure 2) when turned into z scores (skewness = .34 and kurtosis = .57) were not greater than 1.96, which indicated non-significance and thus normality (Field, 2013). Finally, the boxplot (Appendix B, Figure 5) revealed no outliers.
Table 4: Normality Test for Knowledge on Sustainability
Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
Total Knowledge Score .125 20 .200* .969 20 .739
*. This is a lower bound of the true significance.
a. Lilliefors Significance Correction
Figure 6: Histogram for Knowledge on Sustainability
13
Figure 7: Normal Q-Q Plot for Knowledge on Sustainability
Stem width: 10.00 Each leaf: 1 case(s)Hypotheses Testing
Parametric statistics were used to test the hypotheses, since the main assumption of normality had been met. The alpha value was set to .05 for all tests.
Hypothesis 1 The relationship between knowledge of and attitudes towards university sustainability.
The relationship between knowledge of and attitudes towards sustainability was investigated using Pearson’s product moment correlation coefficient. Preliminary analyses were performed to ensure no violation of the assumption of normality while a scatterplot was created to check the assumptions of linearity and homoscedasticity. As seen from the scatterplot (Figure 9) there is a positive correlation between the variables which looks roughly linear. From the Correlations Table (Table 5), one could conclude that there was a statistically significant, moderate, positive correlation between the two variables r = .68, n = 20, p = .001 < .05, with knowledge and attitudes on sustainability sharing 47% of their variance.
Figure 9: Scatterplot for Knowledge and Attitudes
14
Table 5: Correlations between Knowledge and Attitude
Correlations
Total Attitude Score Total Knowledge Score
Total Attitude Score
Pearson Correlation 1 .684**
Sig. (2-tailed) .001
N 20 20
Total Knowledge Score
Pearson Correlation .684** 1
Sig. (2-tailed) .001
N 20 20
**. Correlation is significant at the 0.01 level (2-tailed).
Therefore, hypothesis H1 was accepted and the null hypothesis H01 was rejected.
H1: There will be a significant correlation between the respondents’ knowledge of and attitudes towards university sustainability. (ACCEPTED)
H01: There will be no significant correlation between respondent’s knowledge of and attitudes towards university sustainability. (REJECTED)
Hypothesis 2 The effect of respondents’ programme of study on attitude to university sustainability.
Participants were divided into two groups according to their programme of study. (Group 1: Educational Research MSc, Group 2: Sustainable Development MSc). It was postulated that Sustainable Development students would have more positive attitudes to university sustainability when compared to Educational Research students. An independent samples t-test was used to compare attitudes to sustainability between the two groups. The results of Levene’s test for equality of variances showed that the
15
Sig. value was .90. This is larger than the cut-off point of .05 and shows that the assumption of homogeneity of variances is tenable. The t-test for equality of means indicated that there is a significant difference in the scores of Sustainable Development students (M = 42.13, S.D. = 5.51) when compared with the scores of Educational Research students (M = 34.83, S.D. = 5.32, t(18) = 2.95, p = .008 < .05) (two-tailed). Thus, the assumption that Sustainable Development students would have more positive attitudes to sustainability is tenable.
To check the magnitude of the difference between the two groups the eta squared value for the t-test was calculated. The result, eta squared = .33 indicates a ‘very large’ effect (Cohen, 1988, cited in Cohen, Manion and Morrison, 2007, p. 522). Expressed as a percentage, 33% of the variance in the attitudes to sustainability score is explained by the programme of study.
Table 6: Group Statistics for Programme of Study and Attitude
Group Statistics
programme of study N Mean Std. Deviation Std. Error Mean
Total Attitude ScoreMSc in Educational Research 12 34.8333 5.32291 1.53659
MSc in Sustainable Development 8 42.1250 5.51459 1.94970
Table 7: Independent Samples Test for Programme of Study and Attitude
According to the findings, H2 was accepted while the null hypothesis H02 was
rejected.
H2: There will be a significant difference in attitudes towards university sustainability between the Sustainable Development and Educational Research students. (ACCEPTED)
H02: There will be no significant difference in attitudes towards university sustainability between Sustainable Development and Educational Research students. (REJECTED)
Hypothesis 3 Effect of the respondents’ gender on attitudes towards university sustainability.
Participants were divided into two groups according to their gender. (Group 1: females, Group 2: males). To compare the attitude towards sustainability scores for males and females, an independent-samples t-test was conducted. As the significance level for Levene’s test was .46 (larger than the cut-off point of .05), the assumption of equal variances had not been violated and the first line was read. The t-test for equality of means showed that there was not a statistically significant difference in attitude to sustainability scores for males (M = 37.80, SD = 5.75) and females (M = 37.70, SD = 7.31; t (18) = .034, p = .97 > .05 (two-tailed).
Table 8: Group statistics for gender and attitude
Group Statistics
gender N Mean Std. Deviation Std. Error Mean
Total Attitude ScoreFemale 10 37.7000 7.31893 2.31445
Male 10 37.8000 5.75036 1.81842
Table 9: Independent Samples Test for Gender and Attitude
According to the findings respondents’ gender did not affect their attitude to sustainability. Therefore, H3 is rejected and the null hypothesis H03 is accepted.
H3. There will be a significant difference in attitudes towards university sustainability between male and female respondents. (REJECTED)
H03. There will be no significant difference in attitudes towards university sustainability between male and female respondents. (ACCEPTED)
Hypothesis 4 The prediction of knowledge of sustainability from a combination of information on attitudes towards university sustainability on gender and on age groups.
Standard multiple regression was used to investigate the degree to which scores on knowledge of sustainability could be predicted from attitude to sustainability combined with gender and programme of study.
Bearing in mind that the sample size assumption for multiple regression had been violated since there were only 20 respondents, preliminary analyses were conducted to ensure conformity with the rest of the regression assumptions. The Pearson values in the correlations table (Table 10) indicated that no two values were perfectly correlated (greater than .70). Moreover, the tolerance values in the coefficients table (Table 11) were not less than .10, suggesting no violation of the multicollinearity
18
assumption. This was also supported by the VIF values, which were below the cut-off point of 10. No major deviation from normality was suggested by the distribution of points along the diagonal line of the Normal P-P plot (Figure 10). In the scatterplot all no values are above +3 or below -3, which indicated lack of outliers (Figure 12). The Mahalanobis distances further supported the lack of outliers as none of the values exceeded the critical value for three independent variables (3IVs; 16.27).
Having checked the assumptions, I proceeded to the evaluation of the model. The R² value was .62 (Table 12). Expressed as a percentage this means that the model explains 62 % of the variance in knowledge of sustainability (Pallant, 2010). Yet, when the sample is small as in my case (n = 20), the R square value in the sample tends to be a ‘rather optimistic overestimation of the true value’ (Pallant, 2010). The Adjusted R² value would be more reliable in this case and its value was 0.54. According to Muijs, if the value of the adjusted R square is > .5 there is a ‘strong fit’ of the data to the model (2004, p.163). The Beta coefficient indicated which of the independent variables was the best predictor of the knowledge scores. As seen from Table 11, attitude to sustainability (beta = .45, p = .03 < .5) and programme of study (beta = .40, p = .08 < .5) were statistically significant predictors of sustainability knowledge, while gender (beta = .09, p = .62 > .5) did not make a statistically significant unique contribution to predicting sustainability knowledge. Finally, the ANOVA (Table 13) indicated that the model has reached statistical significance p = .001 < .05.
Table 10: Correlations for Knowledge, Attitudes, Gender and Programme of Study
Total Knowledge
Score
Total Attitude
Score
Gender Programme of
Study
Pearson
Correlation
Total Knowledge
Score
1.000 .684 .259 .701
Total Attitude
Score
.684 1.000 .008 .572
Gender .259 .008 1.000 .408
Programme of
Study
.701 .572 .408 1.000
Sig. (1-tailed) Total Knowledge
Score
. .000 .135 .000
Total Attitude Score .000 . .487 .004
19
Gender .135 .487 . .037
Programme of
Study
.000 .004 .037 .
N
Total Knowledge
Score
20 20 20 20
Total Attitude
Score
20 20 20 20
Gender 20 20 20 20
Programme of
Study
20 20 20 20
Table 11: Coefficients
Model Unstandardized
Coefficients
Standardized
Coefficients
t Sig. Collinearity
Statistics
B Std. Error Beta Tolerance VIF
1
(Constant) 11.752 3.467 3.390 .004
Total Attitude
Score
.224 .098 .452 2.282 .036 .612 1.635
Gender .555 1.099 .090 .506 .620 .758 1.320
Programme of
Study
2.555 1.367 .405 1.869 .080 .510 1.962
a. Dependent Variable: Total Knowledge Score
Table 12: Model Summary
Model R R Square Adjusted R
Square
Std. Error of the
Estimate
Durbin-Watson
1 .785a .616 .545 2.13855 1.238
a. Predictors: (Constant), Programme of Study, Gender, Total Attitude
Score
b. Dependent Variable: Total Knowledge Score
20
Table 13: ANOVA
Model Sum of Squares Df Mean Square F Sig.
1
Regression 117.626 3 39.209 8.573 .001
Residual 73.174 16 4.573
Total 190.800 19
a. Dependent Variable: Total Knowledge Score
b. Predictors: (Constant), Programme of Study, Gender, Total Attitude Score
Figure 10: Normal P-P Plot of Regression Standardised Residual
Figure 11: Scatterplot of Regression Standardised Residual
21
According to the results H4 is rejected and the null hypothesis H04 is accepted.
H4: Knowledge of university sustainability will be predictable from a combination of information on attitudes towards sustainability, gender and programme of study. (ACCEPTED)
H04: Knowledge of university sustainability will not be predictable from a combination of information on attitudes towards sustainability, gender and programme of study. (REJECTED)
6. Discussion
The results of the statistical analysis showed positive attitudes of students towards sustainability; the mean attitude score was 38 out of 50 (Appendix C, Figure 1). Also, the respondents were found knowledgeable on university sustainability issues with a mean score of 25 out of 30 (Appendix C, Figure 2). Ewert and Baker offer ‘social desirability’ response bias as a possible explanation for high scores on sustainability assessment tests (2001, p. 702). They posit that sustainability related issues attract such responses as university students select answers that are socially desirable. On the other hand, high scores could also illustrate respondents’ genuine awareness of sustainability issues within their educational institution.
Below, the results of the hypotheses testing will be discussed in the reverse order starting from the last and moving on to the first hypothesis.
The multiple regression conducted in order to test the fourth hypothesis has reached statistical significance. Therefore, a combination of sustainability attitudes, programme of study and gender seems to be able to predict knowledge. This outcome mirrors results from previous research where academic major, gender and attitudes to
22
sustainability were seen to influence knowledge of sustainability (Ewert and Baker, 2001; Shephard et al., 2009). However, as Tabachnick and Fidel (2007) claim, regression analysis presents relationships among variables, but does not necessarily demonstrate causality, since a correlation between variables could stem from many other sources including the influence of unmeasured variables. For instance, past experience or the social group an individual belongs to might also influence sustainability knowledge.
The testing of hypothesis three indicated that there is no significant difference in the mean scores on sustainability attitudes for males and females. Previous research on gender and sustainability has produced mixed results, with some studies depicting women as more pro-environmental (Harraway et al, 2012; Ewert and Baker, 2001) while others illustrate males as more environmentally conscious (Kawaga, 2007). In the present study gender did not emerge as a determining variable of attitudes to sustainability. However, this conclusion is only tentative as it is not known how non-respondents might have had affected this gender effect result.
The testing of the second hypothesis indicates that the programme of study influences attitudes to sustainability. This finding is consistent with results of previous research. For instance, Harraway et al. (2012) in their longitudinal survey of 364 students at the university of Otago, New Zealand report that the program of study was significant in predicting sustainability attitude scores. Ewert and Baker (2001) have also found that there is a relationship between academic major and the attitudes an individual has regarding the environment. Hodgkinson and Innes (2001) and Wysor (1983) report that economics and business students held less positive beliefs about the environment, compared to environmental studies students. Interpreting how the programme of study impacts on students’ sustainability worldview is not an easy task, as a correlation does always indicate causality. Ewert and Baker (2001) support that the ambient beliefs and practices surrounding an academic discipline (e.g classmates, type of work, literature, professorial attitudes etc) might serve as a determining factor. Still, links between academic affiliation and sustainability attitudes need to be further examined by future research.
The testing of the first hypothesis illustrates that there is a correlation between knowledge of and attitudes to sustainability. This outcome is in accordance with the findings of the International Institute of Sustainable Development survey, where a strong correlation was detected between the knowledge and the attitude scale (Michalos et al., 2009). The knowledge - attitude correlation is also supported by the research of Arbuthnot and Lingg (1975), who claim that knowledge can act as a mediating variable for sustainability attitudes and behaviours. Ewert and Baker (2001) further report that there is a significant relationship between knowledge and expressed environmental attitudes. According to the above findings, it seems that the more knowledgeable the students on sustainability the more positive their attitudes to it.
This correlation could have implications for higher education institutions eager to enhance their students’ attitudes to sustainability. However, since attitudes are a latent construct that cannot be measured directly, extra qualitative data might have offered a more comprehensive insight into students’ views on university sustainability.
23
Finally, since students are regarded as a relatively under-researched group concerning sustainability in higher education (Kawaga, 2007) and as student voice is gradually becoming a crucial component for HEIs’ development with the increase of tuition fees, a wider exploration of their sustainability attitudes, knowledge and needs might provide fertile ground for future research.
7. Conclusion
The aim of this small-scale study was to explore the relationship between the sustainability knowledge and attitudes among postgraduate students of two different MSc programmes at the University of Exeter. The convenience sample and response rate limitations do not allow generalizability of the findings to a larger population and any conclusions drawn can only be tentative. The analysis revealed a positive correlation between respondents’ knowledge of and attitudes to sustainability. Moreover, the programme of study was seen to influence attitudes to sustainability. Participants’ gender on the other hand, did not appear to significantly affect attitudes to sustainability contradicting previous research that portrayed gender as a determining variable.
With the commitment of universities to sustainability in view, the findings of this study might underline the significance of infusing sustainability content across the higher education curriculum so as to improve students’ knowledge and attitudes on sustainability. Further research in this area might enhance higher education institutions’ responsibility or response-ability (Sterling, 2013) to global sustainability challenges.
8. References
24
Arbuthnot, J. and Lingg, S. 1975. A Comparison of French and American Environmental Behaviours, Knowledge and Attitudes. International Journal of Psychology. 10, pp. 275-281.
Arcury, T., Johnson, T. and Scollay, J. 1986. Ecological Worldview and Environmental Knowledge: The “New Environmental Paradigm.” Journal of Environmental Education. 17, pp. 35-40.
Cohen L., Manion L. and Morrison K. 2007. Research Methods in Education. 6th ed. New York: Routledge.
Cole, Davis. 2013. SPSS Step by Step: Essentials for Social and Political Science. Bristol: The Policy Press.
Council of the European Union. 2008. Council Conclusions on Education for Sustainable Development. [pdf] Available at: <http://consilium.europa.eu/uedocs/cms_data/docs/pressdata/en/educ/117855.pdf > [Accessed 10 April 2014].
Creswell, J.W. 2009. Research Design: Qualitative, Quantitative and Mixed Methods Approaches. 3rd ed. London: Sage.
Dryason, R., Bone, E., and Agombar, J., 2012. Student Attitudes towards and Skills for Sustainable Development. [pdf] York: The Higher Education Academy. Available at:<http://www.heacademy.ac.uk/assets/documents/esd/student_attitudes_towards_and_skills_for_sustainable_development.pdf> [Accessed 30 May 2014]
ESRC. 2014. Sustainability.[online resource] Available at: <http://www.esrc.ac.uk/research/research-topics/environment-and-energy-research/sustainability/> [Accessed 3 June 2014].
Economic and Social Research Council (ESRC). 2012. ESRC Framework for Research Ethics (FRE) [pdf] Available at: < http://www.esrc.ac.uk/_images/framework-for-research-ethics-09-12_tcm8-4586.pdf> [Accessed 1 June 2014]
Emanuel, R. and Adams, J.N., 2011. College Students’ Perceptions of Campus Sustainability. International Journal of Sustainability in Higher Education. 12(1), pp. 79-92.
Ewert A. and Baker D. 2001. Standing for Where you Sit: An Exploratory Analysis of the Relationship between Academic Major and Environment Beliefs. Environment and Behaviour. 33: pp. 687-707.
Fowler, F. J. 2002. Survey Research Methods. 3rd ed. Thousand Oaks, CA: Sage.
Harraway, J., Broughton-Ansin, F., Deaker, L., Jowett, T. and Shephard, K., 2012. Exploring the use of the Revised New Ecological Paradigm Scale (NEP) to Monitor the Development of Students’ Ecological Worldviews. The Journal of Environmental Education. 43 (3), pp. 177-191.
Hawcroft, L.J. and Milfont, T.L. 2010. The Use and Abuse of the New Environmental Paradigm Scale over the Last 30 Years: A meta-analysis. Journal of Environmental Psychology. 30. pp. 143-158.
Hodgkinson, S. and Innes, J. 2001. The Attitudinal Influence of Career Orientation in 1st-year University Students: Environmental Attitudes as a Function of Degree choice. The Journal of Environmental Education, 32 (3), pp. 37-40.
Jenkinson, M. 2010. Histogram Bin Size. [online] University of Oxford: FMRIB Centre. Available at: <http://www.fmrib.ox.ac.uk/analysis/techrep/tr00mj2/tr00mj2/node24.html> [Accessed 10 July 2014]
Jones, N., Roumeliotis, S., Iosifides, T., Hatziantoniou, M., Sfakianaki, E., Tsigianni, E., Thivaiou, K., Biliraki, A., Evangelinos, K., 2013. Students’ Perceptions on Environmental Management of HEIs and the Role of Social Capital. International Journal of Sustainability in Higher Education.14 (3), pp. 278-290.
Kawaga, F. 2007. Dissonance in Students’ Perceptions of Sustainable Development and Sustainability. International Journal of Sustainability in Higher Education. 8(3), pp. 317-338.
Krosnick, J.A. and Presser, S. 2010. Question and Questionnaire Design. In: Marsden, P.V. and Wright J.D. eds., 2010. Handbook of Survey Research. 2nd ed. Bingley: Emerald.
Michalos, A., Creech, H., McDonald C., Kahkle, P.M. 2009. Measuring Knowledge, Attitudes and Behaviours towards Sustainable Development: Two Exploratory Studies. [pdf] Manitoba: International Institute for Sustainable Development. Available at: <http://www.iisd.org/pdf/2011/measuring_knowledge_sd_tenth_grade_mb.pdf > [Accessed 2 May 2014]
Moody, G. L. and Hartel, P.G. 2007. Evaluating an Environmental Literacy Requirement Chosen as a Method to Produce Environmentally Literate University Students. International Journal of Sustainability in Higher Education. 8(3), pp. 355-370.
Muijs, D. 2004. Doing Quantitative Research in Education with SPSS. London: Sage.
Oppenheim, A. 1992. Questionnaire Design, Interviewing and Attitude Measurement. London: Continuum.
Pallant, J. 2007. SPSS Survival Manual. Maidenhead: Open University Press.
Policy Studies Institute. 2008. HEFCE strategic review of sustainable development in higher education in England. [pdf] Higher Education Funding Council for England. Available at: <http://www.hefce.ac.uk/data/year/2008/hefcestrategicreviewofsustainabledevelopmentinhighereducationinengland/> [Accessed 5 April 2014].
Shephard, K., Mann, S., Smith, N., and Deaker, L. 2009. Benchmarking the Environmental Values and Attitudes of Students in New Zealand’s Post-compulsory Education. Environmental Education Research. 15 (5), pp. 517-587.
Sterling, S. 2011. The Future Fit Framework; An Introductory Guide to Teaching and Learning for Sustainability in HE. [pdf] Higher Education Academy. Available at: <http://www.heacademy.ac.uk/assets/documents/esd/Future_Fit_270412_1435.pdf > [Accessed 10 March 2014].
Sterling, S, Maxey, L and Luna, H. 2013. The Sustainable University – Progress and Prospects. Abingdon: Routledge.
Stubbs, W. and Cocklin, C., 2008. Teaching Sustainability to Business Students: Shifting Mindsets. International Journal of Sustainability in Higher Education. 9(3), pp. 206-221.
Summers, M., Corney, G. and Childs, A. 2010. Student Teachers’ Conceptions of Sustainable Development: the Starting Points of Geographers and Scientists. Educational Research. 46(2), pp.163-182.
Tabachnick, B. G. and Fidell, L.S. Using Multivariate Statistics. 5th ed. Boston: Pearson.
University of Exeter. 2014. Environmental Sustainability Strategy 2010-2015 [pdf] University of Exeter. Available at: <http://www.exeter.ac.uk/media/universityofexeter/campusservices/sustainability/pdf/2013_Environmental_Sustainability_Strategy.pdf> [Accessed 15 April 2014].
Wachholz, S. Artz, N. and Chene, D. 2012. Warming to the Idea: University Students’ Knowledge and Attitudes about Climate Change. International Journal of Sustainability in Higher Education. 15(2), pp. 128-141.
Wysor, M. S. 1983. Comparing College Students’ Environmental Perceptions and Attitudes: A Methodological Investigation. Environment and Behaviour. 15, pp. 615 -645.
Zelezny, L.C., Chua, P., Aldrich, C. 2000. Elaborating on Gender Differences in Environmentalism. Journal of Social Issues. 56(3). pp.443-457.
Zwickle, A., Koontz, T. M., Slagle, K. M. and Bruskotter, J.T. (forthcoming) Assessing Sustainability Knowledge of a Student Population: Developing a Tool to Measure Knowledge in the Environmental, Economic, and Social Domains. [pdf] Available at: < http://www.aashe.org/files/resources/student-research/2009/full_text_-_adam_zwickle.pdf > [Accessed 5 July 2014]
knowledge on sustainability 1 22.0000 8.000 .498 .718 .638
knowledge on sustainability 2 22.0000 8.316 .508 .647 .642
knowledge on sustainability 3 22.0500 8.892 .217 .283 .688
knowledge on sustainability 4 22.5000 7.105 .551 .698 .618
knowledge on sustainability 5 21.8000 9.221 .262 .348 .679
knowledge on sustainability 6 21.9500 8.155 .589 .810 .631
knowledge on sustainability 7 22.7000 9.168 .082 .462 .719
knowledge on sustainability 8 22.2000 7.958 .356 .534 .665
knowledge on sustainability 9 22.1500 8.345 .381 .774 .659
knowledge on sustainability 10 22.0500 8.997 .187 .596 .693
Figure 2: Attitude Scale Item-Total Statistics
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance
if Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's Alpha
if Item Deleted
attitude to sustainability 134.2500 33.250 .518 .559 .779
attitude to sustainability 234.4500 31.103 .622 .575 .765
attitude to sustainability 333.3500 34.029 .732 .814 .766
attitude to sustainability 433.8500 34.239 .377 .624 .798
attitude to sustainability 533.8500 32.345 .758 .893 .757
attitude to sustainability 633.8500 33.187 .436 .437 .791
attitude to sustainability 733.6000 33.305 .690 .843 .765
attitude to sustainability 834.4000 36.779 .329 .480 .799
attitude to sustainability 933.4000 34.253 .390 .632 .796
attitude to sustainability 10 34.7500 37.250 .170 .402 .821
29
Appendix B: Histograms, Boxplots, Detrended Normal Q-Q Plots
Figure 1: Histogram for Attitude Score
Figure 2: Boxplot for Attitude towards Sustainability
Figure 3: Detrended Normal Q-Q Plot for Total Attitude Score
30
Figure 4: Histogram for Knowledge Score
Figure 5: Histogram for Knowledge on Sustainability Issues
Figure 6: Detrended Normal Q-Q Plot for Total Knowledge Score
31
Appendix C: Descriptives
Figure 1: Attitude Scale DescriptivesDescriptives
Statistic Std. Error
Total Attitude Score
Mean 37.7500 1.43247
95% Confidence Interval for
Mean
Lower Bound 34.7518
Upper Bound 40.7482
5% Trimmed Mean 37.6667
Median 38.5000
Variance 41.039
Std. Deviation 6.40621
Minimum 27.00
Maximum 50.00
Range 23.00
Interquartile Range 9.50
Skewness -.018 .512
Kurtosis -.543 .992
Figure 2: Knowledge Scale Descriptives
Descriptives
Statistic Std. Error
Total Knowledge Score
Mean 24.6000 .70859
95% Confidence Interval for
Mean
Lower Bound 23.1169
Upper Bound 26.0831
5% Trimmed Mean 24.6667
Median 24.5000
Variance 10.042
Std. Deviation 3.16893
Minimum 18.00
Maximum 30.00
Range 12.00
Interquartile Range 3.75
Skewness -.176 .512
Kurtosis -.057 .992
32
Appendix D: Cover Letter to Participants
Dear Colleagues
May I ask you please to fill in a short questionnaire on university sustainability knowledge and attitudes among postgraduate students of two MSc programmes (Educational Research and Sustainable Development).
The survey consists of 22 multiple choice statements; items 1-10 explore attitudes while items 10-20 investigate knowledge on university sustainability with the last two items asking for background information.
The questionnaire takes approximately 5 minutes to fill in and your answers will be treated confidentially and anonymously.