Introduction:
This report is an analysis of a sample of 49 high school students (25 females and 24
males) on primarily their mathematics achievement and attitudes towards mathematics. The
students in the sample are all college bound. The important variables in this study are the
achievement variables, sat which is the student’s performance on SAT math and math which is
the student’s performance on the math final exam. There are three variables measuring student’s
attitudes towards math (known as the attitude variables), the questions asked are, level of interest
in math, how useful is math for your future career and likely to study math in college. These last
three variable are all categorically measured from 1-low to 5-high. The last two variables are
tvhrs which is the hours of evening TV in a typical week, and gender. The teachers and
administrators in the school from which the sample is taken have recently become interested in
knowing the patterns of achievement and attitude and also how they might be related to gender.
This report will look at the relationships between these variables, the best predictors for
mathematics achievement, the relationship between TV habits and mathematics achievement and
Relationships among these variable can be analyzed using the Pearson correlations test
and are shown in Table 1. The null hypothesis is that there is no correlation in the population for
achievement variables i.e. SAT math and Math final exam. The null hypothesis is rejected at the
0.05 significance level and the correlation is strong 0.78 (p < 0.001). Students who tend to score
higher in SAT math tend to score higher in math final exam and students who tend to score lower
population for attitudes. The null hypothesis using a 0.05 level of significance is rejected in all
three cases, interest in math and useful for future has strong correlation 0.68 (p < 0.001), interest
in math and study in college has strong correlation 0.65 (p < 0.001), and useful for future and
The null hypothesis is that there is no correlation in the population for attitude variables
as related to achievement. The null hypothesis using a 0.05 level of significance is rejected in all
cases. Interest in math and math final exam has moderate correlation 0.51 (p < 0.001), useful for
future and math final exam has moderate correlation 0.53 (p < 0.001), and study in college and
math final exam has moderate correlation 0.43 (p < 0.01). The results of the correlation tests
between attitude variables and the other achievement variable Sat math are similarly significant
Other important correlations using the 0.05 level of significance in the population
amongst the variables are the following. Hours of TV per week has moderate negative
correlation with achievement, for math final exam -0.46 (p < 0.01) and for sat math -0.37 (p <
0.01). Hours of TV per week has weak negative correlation with a single attitude variable, useful
for future -.032 (p < 0.05). Gender has a weak correlation 0.33 (p < 0.05) of being male and Sat
math performance.
For further analyses into the relationships and patterns between variables, it would be a
good idea to combine the three attitude variables into a single composite variable. There is a
statistical basis for doing so that measures internal consistency using Cronbach’s coefficient
alpha. Item analysis shows that level of interest in math has a 0.72 correlation with the other two
attitude variables, how useful for your future has a 0.77 correlation with the other two attitude
variables, likely to study math in college has a 0.75 correlation with the other two attitude
variables. All three attitude variables combined have a strong Cronbach’s alpha coefficient of
0.86. There is a strong correlation amongst attitude variables and students who report higher
scores in attitude on one attitude variable tend to have higher scores on the other two attitude
variables. Students who have lower scores on one attitude variable tend to have lower scores on
the other two attitude variables. Therefore it would be useful to use a composite variable for
attitudes.
After analyzing the relationships between the variables, the next step is to find the best
predictors of mathematics achievement using regression analysis. The best predictor for SAT
math scores is scores on the math final exam as it has the strongest correlation for SAT amongst
all the other variables 0.78 (see table 2.). The null hypothesis is that there is no slope in the
population for SAT math scores as predicted by math final exam. Using a 0.05 level of
significance the null hypothesis is rejected p < 0.001. The value of r-square was 0.60, indicating
that 60% of the variation in SAT scores was associated with variation in math final exam score.
The regression equation shows that, predicted SAT scores = 102 + 6(math). On average students
who differed by 1 point on the math exam, differed by 6 points on the SAT exam.
The best predictor for math scores leaving SAT scores aside is overall math attitude as it
has the stronger correlation 0.55. The null hypothesis is that there is no slope in the population
for math scores as predicted by overall math attitude. Using a 0.05 level of significance the null
hypothesis is rejected p < 0.001. The value of r-square was 0.3, indicating that 30% of the
variation in math scores was associated with variation in overall math attitude. The regression
equation shows that, predicted math scores = 63 + 5(aveq). On average students who differed by
1 point on overall math attitude, differed by 5 points on the math final exam.
The people in the school are very interested in knowing how students’ TV habits may be
achievement the outcome of math final exam will be used because between the two achievement
variables, it has a stronger correlation of -0.46. The null hypothesis is that there is no slope in the
population for math scores as predicted by hours of TV per week. Using a 0.05 level of
significance the null hypothesis is rejected p = 0.001. The value of r-square was 0.21, indicating
that 21% of the variation in math scores was associated with variation in hours of TV per week.
The regression equation shows that, predicted math scores = 87 – 0.5(Hours of TV). On average
students who differed by 1 hour of TV viewing, differed by -0.5 points on the math final exam.
This last section of the report will focus on the role of gender in the patterns of achievement and
attitude. The research question being addressed is, are the outcomes of boys and girls similar? A
t-test approach will be utilized to compare the boys and girls. The F-test was used to check the
For the achievement variable SAT math, girls in the sample had a mean score of 551
points (see Table 3.), whereas boys had a mean score of 588 points. The null hypothesis is that
there is no difference in the population means of SAT math scores for boys and girls. The null
hypothesis is rejected at the 0.05 level of significance (t(47) = -2.41, p = 0.02, d = -0.69). Boys in
the population had a higher mean score for SAT math. The high effect size suggests that on
average girls score -0.69 deviations lower than boys in the population.
Using a t-test for math exam, the null hypothesis is that there is no difference in the
population means of math exam scores for boys and girls. The null hypothesis cannot be rejected
at the 0.05 level of significance (t(47) = -1.46, p = 0.15, d = -0.41). There may be no difference
in the population means of boys and girls in math exam scores. The results are summarized in
Table 3.
Using a t-test for attitude towards math (the composite variable for overall attitude), the
null hypothesis is that there is no difference in the attitudes towards math in the population for
boys and girls. The null hypothesis cannot be rejected at the 0.05 level of significance (t(47) = -
1.64, p = 0.11, d = -0.47). There may be no difference in the attitudes towards math for boys and
Using a t-test for TV viewing per week, the null hypothesis is that there is no difference
in the TV viewing per week in the population for boys and girls. The null hypothesis cannot be
rejected at the 0.05 level of significance (t(47) = 0.6, p = 0.55, d = 0.18). There may be no
difference in the TV viewing per week for boys and girls in the population.
Summary Report:
This report is an analysis of a sample of 49 high school students (25 females and 24
males) on primarily their mathematics achievement and attitudes towards mathematics. This
report analyzes the relationships between the given variables, provides the best predictions for
mathematics achievement, the relationship between TV habits and mathematics achievement and
The results found a strong association between the achievement variables as students who
scored higher on the SAT math scored higher on the math final exam, and students who scored
lower on SAT math scored lower on the math final exam. There was a strong association in the
three attitude variables as students who had a more positive attitude on one variable had a more
positive attitude on the other two variables as well. There was a moderate association between
The best prediction for SAT math scores was scores on the math final exam. On average
students who differed by 1 point on the math exam, differed by 6 points on the SAT exam. The
best prediction for math scores leaving SAT scores aside was overall math attitude. On average
students who differed by 1 point on overall math attitude, differed by 5 points on the math final
exam. On average students who differed by 1 extra hour of TV viewing, differed by -0.5 points
For the variables of achievement, attitude and TV viewing, being female did not pose a
statistically significant difference in results as compared to being male. Except in the SAT math
SAT math --
Overall
attitude TV
SAT math --
Table 3. Comparisons of Girls and Boys on math achievement, math attitudes, and
Girls Boys
(SD)
(SD)
Note. Effect sizes represent the difference in the means divided by the pooled standard
deviation.