Вы находитесь на странице: 1из 13

Public Understanding of Science

http://pus.sagepub.com

The gender gap in science attitudes, parental and peer influences: Changes between 1987-88 and
1997-98
Glynis M. Breakwell and Toby Robertson
Public Understanding of Science 2001; 10; 71
DOI: 10.1088/0963-6625/10/1/305

The online version of this article can be found at:


http://pus.sagepub.com/cgi/content/abstract/10/1/71

Published by:

http://www.sagepublications.com

Additional services and information for Public Understanding of Science can be found at:

Email Alerts: http://pus.sagepub.com/cgi/alerts

Subscriptions: http://pus.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.co.uk/journalsPermissions.nav

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


INSTITUTE OF PHYSICS PUBLISHING PUBLIC UNDERSTANDING OF SCIENCE
Public Understand. Sci. 10 (2001) 71–82 www.iop.org/Journals/pu PII: S0963-6625(01)19793-6

The gender gap in science attitudes, parental and peer


influences: changes between 1987–88 and 1997–98
Glynis M. Breakwell and Toby Robertson

This study examines whether differences in attitudes towards science between


males and females (aged 11–14 years) in the United Kingdom have changed
over a 10-year period. The study replicated in 1997–98 a questionnaire survey
first conducted in 1987–88, drawing samples from the same schools used in the
initial research. Data from the two surveys were compared. Both surveys found
that females, in comparison to males, liked science at school less, reported that
they performed worse in science, participated in fewer extracurricular scientific
activities, and had more negative attitudes to science in general. There was a
significant main effect for year of survey upon liking of science at school,
and performance in science at school, with a marked decline in both over
the period. However, there was no significant change in attitudes to science
in general. There were no significant interaction effects between gender and
survey year, suggesting no changes in gender differentials across the 10-year
period. There were, however, changes over time in the factors predicting liking
of and attitudes towards science. Mothers’ perceived support for science is
particularly important in predicting attitudes and involvement of children. In
1997–98, mothers were perceived to be significantly more supportive of science.

1. Introduction

This paper focuses on changes over a period of 10 years in the UK in the gender differences in
attitudes towards science. Girls have been consistently found to like science less and achieve
less in science subjects.1 This not only denies industry and academia a source of talent, but
also helps sustain inequality, as girls are unable to pursue highly valued jobs in science and
engineering. Concern for this situation has led to a focusing of research and intervention in
this area over the last two decades.
Much research has focused on teachers’ attitudes, behavior, and classroom techniques
as a source of possible bias and discouragement to girls or has concentrated on the gendered
image of science and scientists.2 Others have looked at parental and peer influences on the
development of positive or negative attitudes.3
Some commentators have claimed that gender differences in science attitudes and
performance are slowly decreasing, but doubt has been cast on this.4 Drawing a clear conclusion
has been hindered by diversity in methods and measures used. For instance, Archer and
Macrae base their assertions about the decline in gender differences in attitudes to science

0963-6625/01/010071+12$30.00 © 2001 IOP Publishing Ltd and The Science Museum Printed in the UK 71

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


72 G. M. Breakwell and T. Robertson

on respondents’ rating of school subjects on a masculine-feminine dimension whereas Archer


and McDonald interviewed 10–15 year old girls about their school subject likes and dislikes
to substantiate their claims.5 In contrast, Colley et al.’s data consisted of subject preference
rankings from boys and girls. The differences in attitude target and measurement method (e.g.
ratings vs. rankings) may be subtle but they may also account for the diametrically opposed
conclusions drawn.
A more defensible estimate of change would be obtained if identical measures were used
and compared over a suitable period of time from as near identical samples as possible and
for this reason this study repeats a survey carried out in 1987–88 by Breakwell and a team of
researchers. That study collected data from 12 state sector schools selected from six regions
in England and Scotland. The data included attitudes to school science as well as to science
in general, extracurricular science activity and self-reported performance at school science.
Using regression techniques, a small but significant amount of variance in science attitudes was
accounted for by the parental and peer variables. While peer and parental variables were never
expected to be complete predictors of attitudes, there were some interesting relationships.
Mothers’ support for science was found to predict enjoyment of school science and self-
rated performance in science, whereas fathers’ support predicted positive attitudes to science
in general and extracurricular science activities. The data suggested that mothers might be
important in the perpetuation of gender differences in attitudes to science by offering more
encouragement to their sons. Peer influences were also found to be important as earlier claimed
by Keeves and subsequently by Talton and Simpson.6 Pupils who nominated as their peers those
who engaged in normally censured behavior (smoking, drinking and playing truant) were found
to like school science less. Moreover, pupils who nominated as their peers those who liked
science tended to like school science more themselves and claimed to perform better in it, had
more positive attitudes to science in general and took part in more extracurricular scientific
activities.
The new survey reported here obtained measures of these same mediating variables, thus
permitting not only comparison of individual measures then and now but also comparison of
relationships between variables then and now. An objective of this research was to provide
separate measures of liking of school science on the one hand and attitude towards science
in general on the other. Breakwell and Beardsell reported only a small correlation of 0.14
(N = 391) between the two, suggesting that they may tap quite different domains.7 Moreover,
while both measures were predicted by mothers’ and peers’ support of science, the two
measures were also predicted by additional distinctive variables. They showed liking of school
science was predicted by age, peer culture (the index of sanctioned peer behavior) and parental
liberalism (a measure of parental support for egalitarianism), while attitude towards science
in general was predicted by father’s support of science, class, and gender. Therefore, a further
important question addressed here is: to the extent there have been changes over time in science
attitudes, have the changes been similar for school science and science in general?
The measures obtained in both surveys were as follows: three demographic variables,
gender, class, and age. In the 1992 report, class was assessed via the Cambridge Coding System
of father’s occupation.8 Subjective self-reported class was also used but was unpublished. In
the 1997–98 survey, only the self-report measure was used since the objective coding system
had been shown to be unreliable for children reporting parental occupational status. Of course,
self-report may be totally misleading but it can be argued that for this age group subjective
socio-economic status is the only systematic method of assessing status when independent
evidence is not available. Five parental variables were assessed. These were father’s support
of science, mother’s support of science, parental support for education generally, degree of
parent/child activity sharing and parental liberalism. All were indexed through the child’s

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


The gender gap in science attitudes, parental and peer influences 73

self-report. There were also two peer-related measures, assessing whether respondents’ peers
liked science and how rule-breaking they perceived their peers to be. In addition to these
mediating variables, there were four measures relating to science: liking of school science,
self-reported performance at school science, attitudes towards science in the broadest sense
and extracurricular scientific activity.
Other than direct comparison of individual measures, relationships between measures
were examined by comparison of regression path models, a series of nested regressions being
performed on each data set separately. In the first regression, each of the parental and peer
variables was regressed in turn onto the demographic variables of age, class and gender. In the
second regression phase, liking of school science and self-reported performance were regressed
in turn on the demographic, parental and peer variables. In the last regression phase, attitude
towards science in general and extracurricular scientific activity were regressed in turn on the
school science variables and on parental, peer and demographic variables. The order of these
regressions is not supposed to represent any causal assumptions, in that somehow age causes
peer culture that causes liking of school science, which causes attitude towards science in
general. Rather, the order represents an assumption of temporal sequence.

2. Method

Procedure
All schools participating in the Breakwell et al. 1987–88 survey were invited to participate
in the 1997 survey. Five schools in the first survey declined to participate in the new one,
leaving seven schools in the new survey common to the first (two each from Northampton and
Coventry, and one each from Aberdeen, Rochdale and Swindon). The analyses reported here
entail comparisons of only those schools present in both surveys.
It could be argued that, while this procedure optimizes the likelihood that samples are
socio-demographically and educationally matched at the two survey points, it is possible that
changes in school ethos or recruitment area would make the samples dissimilar over time in
some systematically biased fashion. Unfortunately, it was not possible to determine whether
such changes had occurred. The most obvious technique to adopt would be to use postcodes
of the residences of the children participating in the two surveys to assess variance over time in
socio-economic status. This was not possible since the 1987–88 survey did not collect postcode
data. It was possible to establish that there had been no significant changes in the geographical
recruitment profiles of the individual schools over time. However, this does not mean that
those geographical areas from which they drew pupils had not changed in socio-economic or
cultural mix over the ten years. The self-assessment of class given by the respondents is the
only index available in this study for differences over time in the socio-economic status of the
samples. Findings concerning this subjective measure are given below.

Respondents
The questionnaire was administered to respondents by the participating schools during the
school day. Each school drew a representative sample of up to 250 11–16 year olds, (25 boys
and 25 girls from each of the five years). Eventual numbers varied a little from school to school,
resulting in a total of 1,400 respondents. Since the introduction of the National Curriculum
after 1988, all students in the UK now study science up to age 16. However, at the time of the
previous survey, pupils could abandon science at age 14, and for that reason, only the data from
11–14 year olds were analyzed. Accordingly, only the 11–14 year olds in the newer survey

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


74 G. M. Breakwell and T. Robertson

were used for comparison purposes in this report. Finally, participants with missing data
were excluded listwise, which means that failing to complete a single measure resulted in that
participant’s removal from all comparative analyses. This means that the regression analyses
could be comprehensive. This left a total of 408 participants. Breakwell and Beardsell reported
data for 391 participants from the 12 schools in their survey.9 For comparison purposes, this
number was reduced to the 132 participants from the seven schools common to both surveys
if the subjective socio-economic class measure is included (since many failed to answer this
question).

Measures
The measures were identical to those used in the 1987–88 survey (when they were derived
from larger pools of items on the basis of factor analysis). Using the full set of 1,540 responses
in the 1997–98 survey, the indices and their associated Cronbach’s alphas were as follows: No
test of the construct validity of the scales was attempted in this study. The Cronbach alpha
reported merely represents an estimate of inter-item reliability. In Appendix A, full details of
the items in each scale are provided.
• Attitudes towards science (Cronbach’s alpha = 0.59): Seven items reflecting the positive
attitude towards scientific change and involvement in change (e.g., “it is important to
know about science to get a good job;” “overall, science and technology have produced
more good than harm”). The Cronbach alpha for this scale was poor, though technically
acceptable for an attitude index, but it was retained in the analyses in order to maximize
comparability with the 1987–88 data set.
• Extracurricular scientific activity (Cronbach’s alpha = 0.72): Four items rating frequency
of the following activities: conducting a piece of scientific research or experiment out of
school, reading science magazines, attending a club to do some scientific activity, and
watching science programs on television.
• Measurements of liking of school science and perceived performance in school science:
Respondents were asked to list all the subjects they studied at school and rate on a five-
point scale how much they like them and then how good they were at them. Where more
than one science was listed, a mean rating was taken. This process yielded two separate
measures: one of liking and one of perceived performance.
• Scientific peers (Cronbach’s alpha = 0.66): Three items rating how much a respondent’s
peers like science subjects, do well in science and want to be scientists.
• “Delinquent” peers (Cronbach’s alpha = 0.77): Three items indexing peer involvement
in drinking alcohol regularly, smoking regularly, and playing truant sometimes.
• Parent/child activity sharing (Cronbach’s alpha = 0.69): Seven items reflecting the
frequency with which the respondent had done various activities with parents in the
previous six months, such as going to a sports event or making or repairing something
together.
• Mother’s support of science (Cronbach’s alpha = 0.70): Eight items indexing mothers’
(or stepmothers’) encouragement to succeed in mathematics and science (e.g., “would
like me to be a scientist, doctor or engineer,” “encourages me to do well in sciences.”
• Father’s support of science (Cronbach’s alpha = 0.80): This is similar in structure to the
scale for mother’s support of science.
• Parents’ support of school (Cronbach’s alpha = 0.63): 10 items indexing how far parents
are perceived to encourage the respondent to do well in school. For example, “Mum
rewards me for doing well in school,” “Mum often helps me with my school work,” “Dad
tells me how proud he is when I do well at school.”

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


The gender gap in science attitudes, parental and peer influences 75

• Parental “liberalism” (Cronbach’s alpha = 0.67): Five items concerning parents’ support
of equal opportunities and voting (e.g., “votes in most national or local elections,” “thinks
women should have the same education as men”). This is a measure concerned with
supporting democratic processes and egalitarianism.
• Social class: Respondents indicated whether they felt they belonged to the working class,
middle class or upper middle class, belonged to no class, belonged to some other class or
didn’t know. Only respondents who indicated one of the first three classes were included
in this analysis.
Identical data were collected, though not reported, in the earlier survey, and these were
used for comparison purposes.

3. Results

A two-way Multivariate Analysis of Variance (MANOVA) was carried out with survey year
(1987–88/1997–98) and gender (male/female) as independent variables and all of the indices
as dependent variables. Table 1 contains the means and significance levels.
1997–98 respondents were, on average, slightly older and rated their class higher than
those in 1987–88. The former result chiefly reflects the survey data being collected at a
later point in the 1997–98 school year than in the 1987–88 survey. To check whether these
differences could act as confounds for the rest of the analysis, they were entered as covariates
for the analysis of the remaining variables. However, the results were not changed qualitatively
or significantly so Table 1 reports age and class as dependent variables, not covariates.
Relative to the 1987–88 survey and according to respondents’ ratings, it appeared that in
1997–98 mothers were more supportive of science, parents were more supportive of education,
and parents and children shared significantly more activities. Despite this, school children had
more “delinquent” peers, liked science significantly less, performed significantly worse at
school science, and participated in significantly fewer science activities outside school.
The only gender effects significant across both surveys were females liking school science
less, perceiving themselves to perform less well in school science, having more negative
attitudes towards science in general and doing fewer science activities outside of school.
There were no significant interaction effects between gender and survey year, suggesting no
significant changes in gender differentials across the 10-year period.
To investigate how the relationships between variables may have changed the regression
path, models from both surveys were compared. If the subjective class variable was included,
respondent numbers for the 1987–88 survey were too low for a reliable re-analysis. When
regressions were performed, subjective class had no significant effect on any other variable.
And since there were more than 100 respondents removed from the analysis only because they
had not said they belonged to one of the three classes, this variable was dropped from the
regression analyses. This led to 239 respondents being included from the 1987–88 survey and
683 from the 1997–98 survey. Caution should obviously be exercised in interpreting these
results, since the effect of class has not been assessed.
The process of nested regressions, as used by Breakwell and Beardsell in 1992, was
used. In the first phase, the parental and peer variables were regressed in turn onto the
three demographic variables of gender and age. In the second phase, the school science
variables of liking and rated performance were regressed in turn onto all of the parental and
peer variables and the demographic variables. Lastly, attitudes towards science in general and
extracurricular science activities were regressed in turn on all the other variables. Regressions
on both surveys accounted for similarly low absolute amounts of variance though these were

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


76 G. M. Breakwell and T. Robertson

Table 1. Variable means by survey and gender.

Survey year Effect significance

Index Gender 1987–8 1997–8 Year Gender


Class male 1.78 1.95 *
female 1.71 1.86
Age male 12.74 13.15 ***
female 12.66 13.05
Mother’s support male 0.56 0.64 ***
of science female 0.49 0.62
Father’s support male 0.63 0.61
of science female 0.56 0.61
Parental support male 0.64 0.68 *
of school female 0.68 0.69
Parental ‘liberalism’ male 0.69 0.71
female 0.64 0.74
Parent/child male 1.85 2.00 **
activity sharing female 1.80 1.99
Scientific peers male 0.45 0.44
female 0.40 0.43
‘Delinquent’ peers male 0.16 0.32 ***
female 0.12 0.32
Liking of school male 4.31 3.67 *** *
science female 4.04 3.40
Performance at male 4.09 3.71 *** *
school science female 3.96 3.49
Attitudes towards male 2.32 2.34 ***
science female 2.15 2.13
Extra-curricular male 2.06 1.79 *** ***
science activities female 1.76 1.58
Numbers in male 64 202
each cell female 68 206

* p < 0.05.
** p < 0.01.
*** p < 0.0005.

significant statistically, so it seems likely that the parental and peer variables have as a whole
not altered their influence on school and outside school science attitudes during the period.
Figures 1 and 2 represent the path diagrams. Only standardized beta weights that proved
to be statistically significant were used in their construction. Beta weights are presented in the
figures. Those above a line refer to the 1997–98 surveys and those below (in italics) refer to
the 1987–88 survey. It should be noted that the paths found using the 1987–88 data were not
identical to those reported by Breakwell and Beardsell.10 The difference is due to the fact that
their analysis was of data from 12 schools, rather than the subset of seven used here, and class
is not included as an independent variable here.
Starting with Figure 1, age is related to reports concerning peers. With increasing age,
fewer respondents report that they have peers who are interested in science (though this is true
only for the 1997–98 sample) and more report that they have peers that smoke, drink, and are
truant. Moreover, both surveys found a path from scientific peers to liking of school science:

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


The gender gap in science attitudes, parental and peer influences 77

-0.093
-0.123

Father’s Support of Science


0.0
84

“Delinquent” Peers
-0.107 Liking of 1997-8: Adj R2 0.11
29 School Science 1987-8: Adj R2 0.109
0.1 6
0

0.08 03
0 2 -0.222
.
Parent/Child Activity Sharing

0.1
Age -0.132
7
.0

-0.01
-0

-0.095
0.0

4
09
67
Parental Liberalism

0.
8
0.0 Rated Performance 1997-8: Adj R2 0.1
-0. .077
-0

68
13

0.0 at Science 1987-8: Adj R2 0.069


5

Parental Support
Gender
-0.0

2
0.11
65

.105
-0.1 0
Mother’s Support of Science 6
-0.0

-0. 07
- 0 .1 5

07 0. 177
0.271

5
6

0.
75

0.167

Scientific Peers

-0.065
-0.16

Figure 1. Path diagram resulting from the nested regression analyses: liking of and perceived
performance in school science (including standardised beta coefficients in italics below the line for
1987–8 data and in regular font above the line for 1997–8 data).

The more scientific one’s peers, the more one likes science. This suggests that it is possible
that the relationship of age to liking of science may be largely moderated through the gradual
decline of scientific interests among peers, although a fairly small direct effect of age was also
found in the 1997–98 survey.
In both surveys, gender influenced both liking of school science and rated performance
at science, supporting the results of the MANOVA. Girls liked science less and perceived
their performance to be worse. Breakwell and Beardsell, using data from all 12 schools in
the original survey, had found no direct effect of gender on the school science variables, only
one mediated through mother’s support of science. One of the clear differences between the
two surveys in the current study is that in 1987–88 there was a path found between gender
and mothers’ support of science, with girls perceiving less support from their mother. This
path was not found in the more recent survey. Inspection of the means in Table 1 confirms
that girls and boys now have equivalent perceptions of their parents’ support of science. Both
perceive mothers as more supportive than in 1987–88. In relation to the fathers there has
been no statistically significant change. It is notable that in both surveys parent/child sharing
of activities declines with age. This is particularly relevant when we consider the effect of
activity-sharing on the degree of extracurricular scientific activity (indicated in Figure 2) and
could be a further dimension for a reduction of science interest with increasing age. Figure 2
also shows that both surveys found unmediated gender effects on both attitudes towards science

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


78 G. M. Breakwell and T. Robertson

Liking of
Father’s Support of Science School Science

0.1 112
0.
98
0.0
“Delinquent” Peers 0.1 6

0.0 .22
8

0
46
Parent/Child Activity Sharing 0.1 Attitude
0.1 Toward 1997-8: Adj R2 0.25
0.2 74 Science 1987-8: Adj R2 0.171
21
Parental Liberalism

0.271
Age -0.026 2
-0.21 7
-0.061 -0.19

Extracurricular
-0.1 Science 1997-8: Adj R2 0.232
Gender -0.079 1987-8: Adj R2 0.196
Activity
Parental Support 8
0.07
0.1 .173
67
0

74
0.0 142
Mother’s Support of Science

0.1 6
0.
1

8
0.10

07
0.0
0.1066
05
0.

Scientific Peers Perceived Performance


at Science

Figure 2. Path diagram resulting from the nested regression analyses: attitude toward science and
extramural science activity (including standardised beta coefficients in italics below the line for
1987–8 data and in regular font above the line for 1997–8 data).

and extracurricular scientific activity.


The two surveys both show that mothers’ support of science is directly predictive of
attitudes towards science and so is fathers’ support of science (but to a lesser degree).
One difference between the two surveys is that pupils’ liking of and perceived performance
in school science is much more related to attitudes and behavior outside school in 1997–98.
In the 1987–88 survey, paths were found between liking school science and attitude toward
science and between perceived performance and extracurricular scientific activity. In the
1997–98 survey, there are four significant paths, especially between liking of school science
and attitudes towards science in general and extracurricular scientific activities. Thus it would
seem to be the case that the linkages between school science and science in general became
closer between the surveys.
Two further observations can be made here. Firstly, age had a direct effect on
extracurricular scientific activity, even when all other variables were taken into account. As
in liking of school science, pupils tend to go off extracurricular science the older they get.
Secondly, in the 1997–98 survey, parental support for education in general had a positive
relationship with liking of school science and perceived performance in science but no direct
relationship with extent of extracurricular scientific activities or attitudes to science generally.
This suggests the possible effects of parental encouragement for the respondent at school do
not extend into a push toward particular attitudes towards science outside of school despite its
impacts on liking of science and performance in school.

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


The gender gap in science attitudes, parental and peer influences 79

4. Discussion

The amount of variance accounted for by the regressions was small even though the
relationships reported are statistically significant. This is not surprising given the multiplicity
of influences that should be expected to work on attitudes towards science and liking of
school science. For instance, teachers’ attitudes, behavior and classroom techniques, and
the masculine (and often anti-social) stereotype of the scientist are likely to have a role to play.
In fact, the amount of variance accounted for in the regressions reported here is not the crucial
issue. The central aim of this research was to investigate the possibility of changes over time
in a restricted set of variables.
Liking of and perceived performance in school science and extracurricular science
activities were all significantly lower in the 1997–98 survey. Attitudes towards science in
general remained the same. Amid this gloomy picture, one might take heart from the way
attitude towards science in general has apparently remained constant. However, even here the
strong path between liking of school science and attitudes suggests that reductions in liking of
school science may subsequently lead to a more negative attitude to science generally.
There were significant gender differences on all four variables across both surveys, with
no significant interactions that might indicate a narrowing or widening of gender differences.
Indeed, examining the means in Table 1, it is apparent that the two surveys find roughly
equivalent-sized gender differences, with perhaps perceived performance at school science
showing a slight widening. Of course, this study did not gather data on actual performance.
There has been evidence that females are performing better in school science than males in UK
national examinations in 2000. Consequently, this study may be revealing a gender difference
in willingness to claim positive performance in science, which is constant over time rather than
an objective difference in performance.
The finding that liking of school science has declined over the 10 years is puzzling.
Breakwell and Beardsell’s analysis suggested that liking of school science was positively
influenced by mother’s support of science and this was found to have significantly increased
in the more recent survey.11 Clearly, other countervailing influences are at work. The 1997–98
survey found that age negatively affected liking of school science and although the 1997–98
sample was marginally older, the size of difference in age and the beta weight of the path
(0.08) suggests this could not account for much of the difference in liking. Breakwell and
Beardsell also found a negative relationship between “delinquent” peers and liking of school
science, though it was not found in either the re-analysis of the subset of the 1987–88 data used
here or in the 1997–98 data. Nevertheless, “delinquent” peers significantly increased across
surveys and cannot be ruled out as a possible source of influence. A similar pattern obtained
for extracurricular scientific activity. Despite the significant increase in parent/child activity
sharing, which was significantly positively related to scientific activity outside of school, this
latter variable showed a significant decrease across surveys. Of the seven parental and peer
variables considered, six either increased in a way that should promote liking of school science
or did not change significantly, leaving only “delinquent” peers to account for the decrease in
liking of science. This suggests that positive changes in parental factors were swamped by
changes in other areas not measured by this study.
To conclude, no evidence was found to suggest that the gender gap in attitudes towards
science is closing. This was the case for all four of the science variables examined. The
amount of variance captured by the nested regressions suggested that demographic, parental
and peer influences are only small parts of the broader picture. Indeed, this is underlined
by the positive improvements in some of the parental variables while the science preference
measures decreased. Alternative explanations to account for these decreases might focus on

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


80 G. M. Breakwell and T. Robertson

the masculine and anti-social stereotypes of scientists, and changes in youth culture.

Appendix A. The scales used

1. Extracurricular Scientific Activity


Frequency of following activities:
i. Conduct a piece of scientific research or experiment outside of school
ii. Read science magazines
iii. Attend a club to do some scientific activity
iv. Watch science programs on TV
2. Father’s Support of Science
i. Dad expects me to do well in mathematics
ii. Dad encourages me to do well in mathematics
iii. Dad is good at mathematics
iv. Dad expects me to do well in sciences
v. Dad encourages me to do well in sciences
vi. Dad often helps me with school homework
vii. Dad expects me to complete a college/university course
viii. Dad knows a lot about sciences
ix. Dad wants me to learn about computers
x. Dad would like me to become a scientist, doctor or engineer
3. Mother’s Support of Science
i. Mum expects me to do well in mathematics
ii. Mum expects me to do well in sciences
iii. Mum encourages me to do well in sciences
iv. Mum encourages me to well in mathematics
v. Mum would like me to become a scientist, doctor or engineer
vi. Mum is good at mathematics
vii. Mum wants me to learn about computers
viii. Mum expects me to complete a college/university course
4. Parents’ Support of School
i. Dad rewards me for doing well at school
ii. Dad tells me how proud he is when I do well at school
iii. Mum rewards me for doing well at school
iv. Dad is very busy and does not spend much time with me
v. Mum often does not understand how I feel about things
vi. Dad often does not understand how I feel about things
vii. Mum tells me how proud she is when I do well at school
viii. Mum often helps me with my school work
ix. Mum is very busy and does not spend much time with me
x. Dad tells me he is confident in my ability
5. Parent/Child Shared Activities
i. Been to a play, concert or film
ii. Been to a sports event
iii. Worked on making or repairing something together
iv. Visited a museum, zoo, aquarium, botanical garden or planetarium together
v. Taken part in some recreation like tennis, skating or swimming

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


The gender gap in science attitudes, parental and peer influences 81

vi. Sang or played music together


vii. Attended a church or religious service
6. Scientific Peers
Most of my friends. . .
i. Like science subjects
ii. Do well in science subjects
iii. Want to be scientists, doctors, engineers or mathematicians
7. “Delinquent” Peers
Most of my friends. . .
i. Drink alcohol regularly
ii. Smoke cigarettes regularly
iii. Sometimes play truant
8. Attitudes to Science
i. I would like to have a job involving new technology
ii. I think a technical training will help me in the future
iii. It’s important to know about science to get a good job
iv. Overall, science and technology have produced more good than harm
v. My school thinks the sciences are important subjects
vi. I think it will be good if the speed of scientific change increases
vii. New inventions will always be found to deal with the harmful consequences of
technological development

References

1 B. J. Becker, “Gender and science achievement: a reanalysis of studies from two meta-analyses,” Journal of
Research in Science Teaching 26 (1989): 141–169; A. Colley, C. Coomber, and D. J. Hargreaves, “Gender
effects in school subject preferences: a research note,” Educational Studies 20 (1994): 13–18; A. Kelly, ed., The
Missing Half: girls and science education (Manchester: Manchester University Press, 1981); R. A. Schibeci,
“Attitudes to science: an update,” Studies in Science Education 11 (1984): 26–59; R. D. Simpson and J. S. Oliver,
“A summary of major influences on attitude toward and achievement in science,” Science Education 74 (1990):
1–18; M. Weinburgh, “Gender differences in student attitudes toward science: A meta-analysis of the literature
from 1970–1991,” Journal of Research in Science Teaching 32 (1995): 387–398; and H. Weinreich-Haste, “The
image of science,” in The Missing Half: girls and science education ed. A. Kelly (Manchester: Manchester
University Press, 1981).
2 J. B. Kahle, L. H. Parker, L. J. Rennie, and D. Riley, “Gender differences in science education: building a model,”
Educational Psychologist 28 (1993): 379–404; A. Kelly, “Sex stereotypes and school science: a three year
follow up,” Educational Studies 14 (1988): 151–163; R. E. Myers III and J. T. Fouts, “A cluster analysis of high
school science classroom environments and attitudes towards science,” Journal of Research in Science Teaching
29 (1992): 929–937; D. L. Pratt, “Responsibility for student success failure and observed verbal-behavior among
secondary science and mathematics teachers,” Journal of Research in Science Teaching 22 (1985): 809–816;
M. Spear, “Teachers’ attitudes towards girls and technology,” in Girl-friendly schooling, eds, J. White, R. Deem,
L. Kant and M. Cruikshank (London: Methuen, 1985) 237–268; J. Archer and M. Macrae, “Gender perceptions
of school subjects among 10–21 year-olds,” British Journal of Educational Psychology 61 (1991): 99–103;
J. Archer and S. Freedman, “Gender-stereotypic perceptions of academic disciplines,” British Journal of Educa-
tional Psychology 59 (1989): 306–313; S. Sjøberg, “Gender and the image of science,” Scandinavian Journal of
Education Research 32 (1988): 498–560; M. Sadker, Failing at Fairness: How America’s Schools Cheat Girls
(New York: C. Scribner’s Press, 1994).
3 G. Breakwell and S. Beardsell, “Gender, parental and peer influences upon science attitudes and activities,”
Public Understanding of Science 1 (1992): 183–197; A. C. Howe and B. Durr, “Using concrete materials and
peer interaction to enhance learning in chemistry,” Journal of Research in Science Teaching 19 (1982): 225–232;
J. Keeves, “The home, the school and achievement in mathematics and science,” Science Education 59 (1975):
439–460; and E. L. Talton and R. D. Simpson, “Relationships between peer and individual attitudes towards
science,” Science Education 69 (1985): 19–24.

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008


82 G. M. Breakwell and T. Robertson

4 J. Archer, “Gender stereotyping of school subjects,” The Psychologist 5 (1992): 66–69; J. Archer and M. Macrae,
“Gender perceptions of school subjects among 10–21 year-olds,” British Journal of Educational Psychology 61
(1991): 99–103; J. Archer and M. McDonald, “Gender roles and school subjects in adolescent girls,” Educational
Research 33 (1991): 55–64; A. Colley, C. Coomber, and D. J. Hargreaves, “Gender effects in school subject
preferences: a research note,” Educational Studies 20 (1994): 13–18; and J. Meece and G. Jones, “Girls
in mathematics and science: Constructivism as a feminist perspective,” The High School Journal 79 (1996):
242–248.
5 J. Archer and M. Macrae, “Gender perceptions of school subjects among 10–21 year-olds,” British Journal of
Educational Psychology 61 (1991): 99–103; and J. Archer and M. McDonald, “Gender roles and school subjects
in adolescent girls,” Educational Research 33 (1991): 55–64.
6 J. Keeves, “The home, the school and achievement in mathematics and science,” Science Education 59 (1975):
439–460; and E. L. Talton and R. D. Simpson, “Relationships between peer and individual attitudes towards
science,” Science Education 69 (1985): 19–24.
7 G. Breakwell and S. Beardsell, “Gender, parental and peer influences upon science attitudes and activities,” Public
Understanding of Science 1 (1992): 183–197.
8 K. Prandy, A. Stewart, and R. M. Blackburn, White Collar Work, (London: Macmillan, 1982).
9 G. Breakwell and S. Beardsell, “Gender, parental and peer influences upon science attitudes and activities,” Public
Understanding of Science 1 (1992): 183–197.
10 ibid.
11 ibid.

Author

Glynis Breakwell is Professor of Psychology at the University of Surrey, UK, and Toby
Robertson is Lecturer in Psychology at Brunel University, UK. Address correspondence to
Breakwell at SPERI, School of Human Sciences, University of Surrey, Guildford GU2 7XH,
UK.

Downloaded from http://pus.sagepub.com at Umea University Library on November 17, 2008

Вам также может понравиться