Analysis of Dataset in Biostatistics

This document provides an analysis of a dataset in the field of biostatistics. It includes the mean, standard deviation, minimum and maximum values for age, as well as the frequency of students in different age categories. The document also explores the descriptive statistics for demographic information of the students and tests for the difference in means based on gender, metropolitan background status, and study mode.

Biostatistics
Student Name:
Instructor Name:
Course Number:
16 April 2019

Assessment Task 1: Analysis of dataset
1. In your SPSS file, age (at time of enrolment) is given as a continuous variable.
a. What is the mean, standard deviation, minimum and maximum values for age?
Answer
Table 1 below presents the descriptive statistics for the variable age. As can be seen, the
mean age of the respondents was 20.50 years old with the oldest participant being 59
years old and the youngest participant being 16 years old. The standard deviation is 4.89.
Table 1: Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
AGE 38681 16 59 20.50 4.888
Valid N (listwise) 38681
b. What is the frequency (% and number) of students in each of these new age
categories?
Answer
In the next table we present the frequency for the newly created variable (age group).
As can be seen, majority (45.9%, n = 17760) are aged 18 years and below while
minority group (9.7%, n = 3755) were aged 26 years and above.
Table 2: Age group
Frequency Percent Valid Percent Cumulative Percent
Valid
18 yrs. Old and below 17760 45.9 45.9 45.9
19 to 21 yrs. old 11672 30.2 30.2 76.1
22 to 25 yrs. old 5494 14.2 14.2 90.3
26 yrs. Old and above 3755 9.7 9.7 100.0
Total 38681 100.0 100.0
2. Descriptive statistics for the demographic information of the students

Answer
Table 3: Demographic characteristics
Demographic characteristics Frequency (n) Percent (%)
State
NSW 15860 41.0
Victoria 13571 35.1
Queensland 7528 19.5
ACT 1722 4.5
Total 38681 100.0
Fees Status
Domestic 32238 83.3
International 6443 16.7
Total 38681 100.0
Cohort
2005 3259 8.4
2006 3615 9.3
2007 3944 10.2
2008 4086 10.6
2009 5010 13.0
2010 5687 14.7
2011 6383 16.5
2012 6697 17.3
Total 38681 100.0
Living Arrangement
At home 20840 53.9
College/student
accommodation
6850 17.7
Independently 10991 28.4
Total 38681 100.0
Age
< 18 yrs. old 17760 45.9
19-21 yrs. old 11672 30.2
22-25 yrs. old 5494 14.2
> 26 yrs. old 3755 9.7
Total 38681 100.0
Metro
Metro 27223 70.4

Non-metro 5015 13.0
Missing 6443 16.7
Total 38681 100.0
Degree Type
Single 34620 89.5
Double 4061 10.5
Total 38681 100.0
Faculty
Arts and Sciences 9004 23.3
Education 15038 38.9
Health Sciences 11729 30.3
Theology and Philosophy 588 1.5
Business 2322 6.0
Total 38681 100.0
Study Type
FT 34770 89.9
PT 3911 10.1
Total 38681 100.0
Table 3 above gives the demographic characteristics of the sample under study. Clearly
most of the participants were from NSW state followed by Victoria State. In terms of
cohort, most of the participants were in the cohort of 2012 while the least came from the
cohort of 2005. In terms of age, majority (45.9%, n = 17760) are aged 18 years and below
while minority group (9.7%, n = 3755) were aged 26 years and above.
3. Testing for the difference in means
Answer
a. Gender
Difference in mean aggression scores
Group Statistics
GENDER N Mean Std. Deviation Std. Error Mean

driver_agg Male 10449 7.52 4.335 .042
Female 28232 7.51 4.347 .026
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differen
ce
Std.
Error
Differen
ce
95% Confidence
Interval of the
Difference
Lower Upper
driver_agg
Equal variances
assumed
.117 .732 .083 38679 .934 .004 .050 -.093 .102
Equal variances
not assumed
.083 18713 .934 .004 .050 -.093 .102
From the above table we can see that the p-value for the t-test is 0.934 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean aggression scores are the same for both male and female participants. The
two groups don’t significantly differ in terms of the mean aggressions scores.
Difference in mean thrill seeking scores
Group Statistics
GENDER N Mean Std. Deviation Std. Error Mean
thrill Male 10449 5.00 1.217 .012
Female 28232 5.00 1.226 .007
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t Df Sig. (2-
tailed)
Mean
Differen
ce
Std.
Error
Differen
95% Confidence
Interval of the
Difference

ce Lower Upper
thrill
Equal variances
assumed
.847 .357 -.370 38679 .711 -.005 .014 -.033 .022
Equal variances
not assumed
-.371 18783 .710 -.005 .014 -.033 .022
From the above table we can see that the p-value for the t-test is 0.711 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean thrill seeking scores are the same for both male and female participants.
The two groups don’t significantly differ in terms of the mean thrill seeking scores.
Difference in mean risk acceptance scores
Group Statistics
GENDER N Mean Std. Deviation Std. Error Mean
risk_accep Male 10449 8.59 4.358 .043
Female 28232 8.51 4.357 .026
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differen
ce
Std.
Error
Differen
ce
95% Confidence
Interval of the
Difference
Lower Upper
risk_accep
Equal variances
assumed
.054 .817 1.571 38679 .116 .078 .050 -.019 .176
Equal variances
not assumed
1.571 18663 .116 .078 .050 -.019 .176

From the above table we can see that the p-value for the t-test is 0.116 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean risk acceptance scores are the same for both male and female participants.
The two groups (males and females) don’t significantly differ in terms of the mean risk
acceptance scores.
b. Metropolitan background status
Difference in mean aggression scores
Group Statistics
METRO N Mean Std. Deviation Std. Error Mean
driver_agg Metro 27223 7.53 4.345 .026
Non-metro 5015 7.49 4.303 .061
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differen
ce
Std.
Error
Differen
ce
95% Confidence
Interval of the
Difference
Lower Upper
driver_agg
Equal variances
assumed
1.060 .303 .714 32236 .475 .048 .067 -.083 .178
Equal variances
not assumed
.719 7029 .472 .048 .066 -.082 .177
From the above table we can see that the p-value for the t-test is 0.475 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean aggression scores are the same for both the metro and the non-metro

participants. The two groups (metro and non-metro) don’t significantly differ in terms of
the mean aggression scores.
Difference in mean thrill seeking scores
Group Statistics
METRO N Mean Std. Deviation Std. Error Mean
thrill Metro 27223 5.01 1.224 .007
Non-metro 5015 4.99 1.207 .017
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differen
ce
Std.
Error
Differen
ce
95% Confidence
Interval of the
Difference
Lower Upper
thrill
Equal variances
assumed
1.845 .174 .686 32236 .493 .013 .019 -.024 .050
Equal variances
not assumed
.692 7048.2 .489 .013 .019 -.024 .049
From the above table we can see that the p-value for the t-test is 0.493 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean thrill seeking scores are the same for both the metro and the non-metro
participants. The two groups (metro and non-metro) don’t significantly differ in terms of
the mean thrill seeking scores.
Difference in mean risk acceptance scores

Group Statistics
METRO N Mean Std. Deviation Std. Error Mean
risk_accep
Metro 27223 8.53 4.358 .026
Non-metro 5015 8.59 4.304 .061
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differen
ce
Std.
Error
Differen
ce
95% Confidence
Interval of the
Difference
Lower Upper
risk_accep
Equal variances
assumed
3.228 .072 -.866 32236 .386 -.058 .067 -.189 .073
Equal variances
not assumed
-.874 7040.4
76
.382 -.058 .066 -.188 .072
From the above table we can see that the p-value for the t-test is 0.493 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean risk acceptance scores are the same for both the metro and the non-metro
participants. The two groups (metro and non-metro) don’t significantly differ in terms of
the mean risk acceptance scores.
c. Study mode
Difference in mean aggression scores
Group Statistics
STUDY_MODE N Mean Std. Deviation Std. Error Mean
driver_agg FT 34770 7.51 4.344 .023
PT 3911 7.54 4.339 .069

Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differen
ce
Std.
Error
Differen
ce
95% Confidence
Interval of the
Difference
Lower Upper
driver_agg
Equal variances
assumed
.323 .570 -.309 38679 .757 -.023 .073 -.166 .121
Equal variances
not assumed
-.310 4834.4
53
.757 -.023 .073 -.166 .121
From the above table we can see that the p-value for the t-test is 0.757 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean aggression scores are the same for both the full time students and the part
time students. The two groups (Full time and part time students) don’t significantly differ
in terms of the mean aggression scores.
Difference in mean thrill seeking scores
Group Statistics
STUDY_MODE N Mean Std. Deviation Std. Error Mean
thrill FT 34770 5.00 1.224 .007
PT 3911 5.00 1.225 .020
Independent Samples Test
Levene's Test
for Equality of
Variances
t-test for Equality of Means

F Sig. t df Sig. (2-
tailed)
Mean
Differe
nce
Std.
Error
Differe
nce
95% Confidence
Interval of the
Difference
Lower Upper
thrill
Equal
variances
assumed
.222 .637 .132 38679 .895 .003 .021 -.038 .043
Equal
variances not
assumed
.132 4829.6
35
.895 .003 .021 -.038 .043
From the above table we can see that the p-value for the t-test is 0.895 (a value greater
than 5% level of significance), we therefore fail to reject the null hypothesis and conclude
that the mean thrill seeking scores are the same for both the full time students and the part
time students. The two groups (Full time and part time students) don’t significantly differ
in terms of the mean thrill seeking scores. .
Difference in mean risk acceptance scores
Group Statistics
STUDY_MODE N Mean Std. Deviation Std. Error Mean
risk_accep FT 34770 8.51 4.355 .023
PT 3911 8.68 4.374 .070
Independent Samples Test
Levene's Test
for Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differe
nce
Std.
Error
Differe
nce
95% Confidence
Interval of the
Difference
Lower Upper
risk_accep Equal variances
assumed
.045 .832 -2.269 38679 .023 -.167 .073 -.311 -.023

Equal variances
not assumed
-2.261 4823.7
06
.024 -.167 .074 -.311 -.022
From the above table we can see that the p-value for the t-test is 0.023 (a value less than
5% level of significance), we therefore reject the null hypothesis and conclude that the
mean risk acceptance scores are significantly different for the full time students and the
part time students. Part time students (M = 4.68, SD = 4.37) have slightly higher mean
risk acceptance as compared to the full time students (M = 8.51, SD = 4.36). The
difference of 0.167 was found to be significant hence resulting to significant differences
in the two groups (full time and part time students) in terms of the mean risk acceptance.
d. RTA in past 12 months (from follow-up survey)
Difference in mean aggression scores
Group Statistics
RTA_one_crash N Mean Std. Deviation Std. Error Mean
driver_agg No RTAs 33628 6.79 4.116 .022
One RTA or more 5053 12.34 2.218 .031
Independent Samples Test
Levene's Test
for Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differe
nce
Std.
Error
Differe
nce
95% Confidence
Interval of the
Difference
Lower Upper
driver_agg
Equal variances
assumed
3179.6
09
.000 -93.86 38679 .000 -5.552 .059 -5.668 -5.436
Equal variances
not assumed
-144.45 11183. .000 -5.552 .038 -5.627 -5.476

From the above table we can see that the p-value for the t-test is 0.000 (a value less than
5% level of significance), we therefore reject the null hypothesis and conclude that the
mean aggression scores are significantly different for the no RTAs and one RTA or more.
One RTA or more (M = 12.34, SD = 2.22) have higher mean aggression scores as
compared to the no RTA (M = 6.79, SD = 4.12). The difference of 5.55 was found to be
very significant hence resulting to significant differences in the two groups (no RTA and
one RTA or more) in terms of the mean aggression scores.
Difference in mean thrill seeking scores
Group Statistics
RTA_one_crash N Mean Std. Deviation Std. Error Mean
thrill No RTAs 33628 4.80 1.159 .006
One RTA or more 5053 6.34 .686 .010
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std.
Error
Differenc
e
95% Confidence
Interval of the
Difference
Lower Upper
thrill
Equal variances
assumed
1715.363 .000 -92.063 38679 .000 -1.539 .017 -1.572 -1.507
Equal variances not
assumed
-133.493 10036.7 .000 -1.539 .012 -1.562 -1.517
From the above table we can see that the p-value for the t-test is 0.000 (a value less than
5% level of significance), we therefore reject the null hypothesis and conclude that the

mean thrill seeking scores are significantly different for the no RTAs and one RTA or
more. One RTA or more (M = 6.34, SD = .69) have higher mean aggression scores as
compared to the no RTA (M = 4.80, SD = 1.16). The difference of 1.54 was found to be
very significant hence resulting to significant differences in the two groups (no RTA and
one RTA or more) in terms of the mean thrill seeking scores.
Difference in mean risk acceptance scores
Group Statistics
RTA_one_crash N Mean Std. Deviation Std. Error Mean
risk_accep No RTAs 33628 7.91 4.210 .023
One RTA or more 5053 12.68 2.748 .039
Independent Samples Test
Levene's Test for
Equality of
Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differe
nce
Std.
Error
Differe
nce
95% Confidence
Interval of the
Difference
Lower Upper
risk_accep
Equal
variances
assumed
1951.96 .000 -78.154 38679 .000 -4.775 .061 -4.895 -4.655
Equal
variances
not
assumed
-106.2 9076.2 .000 -4.775 .045 -4.863 -4.687
From the above table we can see that the p-value for the t-test is 0.000 (a value less than
5% level of significance), we therefore reject the null hypothesis and conclude that the
mean risk acceptance scores are significantly different for the no RTAs and one RTA or

more. One RTA or more (M = 12.68, SD = 2.75) have higher mean aggression scores as
compared to the no RTA (M = 7.91, SD = 4.21). The difference of 4.78 was found to be
very significant hence resulting to significant differences in the two groups (no RTA and
one RTA or more) in terms of the mean risk acceptance scores.
.
4. In terms of depression, is there a difference between the depressed and not depressed
groups according to their:
a. Gender
Answer
For this task, we ran a chi-square test of association. The results are presented in the
tables below;
GENDER * depression Cross tabulation
Count
depression Total
Not depressed Depressed
GENDER
Male 9365 1084 10449
Female 25355 2877 28232
Total 34720 3961 38681
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square .280a 1 .597
Continuity Correctionb .260 1 .610
Likelihood Ratio .279 1 .597
Fisher's Exact Test .598 .305
Linear-by-Linear Association .280 1 .597
N of Valid Cases 38681
a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 1070.00.
b. Computed only for a 2x2 table

From the Chi-square tests table, we can see the p-value to be 0.597 (a value greater than α
= 0.05), we therefore fail to reject the null hypothesis and come to a conclusion that there
exists no significant differences in having depression among the different gender groups
(males and females).
b. Metropolitan background status
Answer
For this task, we ran a chi-square test of association. The results are presented in the
tables below;
METRO * depression Cross tabulation
Count
depression Total
Not depressed Depressed
METRO Metro 24443 2780 27223
Non-metro 4495 520 5015
Total 28938 3300 32238
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square .114a 1 .736
Continuity Correctionb .097 1 .755
Likelihood Ratio .113 1 .737
Fisher's Exact Test .743 .378
Linear-by-Linear Association .114 1 .736
N of Valid Cases 32238
a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 513.35.
b. Computed only for a 2x2 table

From the Chi-square tests table, we can see the p-value to be 0.736 (a value greater than α
= 0.05), we therefore fail to reject the null hypothesis and come to a conclusion that there
exists no significant differences in having depression among the different metro groups
(metro and non-metro).
c. Study mode
Answer
For this task, we ran a chi-square test of association. The results are presented in the
tables below;
STUDY_MODE * depression Cross tabulation
Count
depression Total
Not depressed Depressed
STUDY_MODE
FT 31241 3529 34770
PT 3479 432 3911
Total 34720 3961 38681
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square 3.072a 1 .080
Continuity Correctionb 2.975 1 .085
Likelihood Ratio 3.011 1 .083
Fisher's Exact Test .084 .042
Linear-by-Linear Association 3.072 1 .080
N of Valid Cases 38681
a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 400.49.
b. Computed only for a 2x2 table
From the Chi-square tests table, we can see the p-value to be 0.080 (a value greater than α
= 0.05), we therefore fail to reject the null hypothesis and come to a conclusion that there

exists no significant differences in having depression among the different study mode
groups (full time and part time group).
d. Fee status
Answer
For this task, we ran a chi-square test of association. The results are presented in the
tables below;
FEE_STATUS * depression Cross tabulation
Count
depression Total
Not depressed Depressed
FEE_STATUS Domestic 28938 3300 32238
International 5782 661 6443
Total 34720 3961 38681
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square .003a 1 .956
Continuity Correctionb .001 1 .974
Likelihood Ratio .003 1 .956
Fisher's Exact Test .966 .487
Linear-by-Linear Association .003 1 .956
N of Valid Cases 38681
a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 659.77.
b. Computed only for a 2x2 table
From the Chi-square tests table, we can see the p-value to be 0.956 (a value greater than α
= 0.05), we therefore fail to reject the null hypothesis and come to a conclusion that there
exists no significant differences in having depression among the different fees staus
groups (domestic and international groups).

5. Using “RTA_one_crash” as the dependent variable, perform a binary logistic regression.
a. Provide a table of odds ratios, confidence intervals and significance values for each
predictor variable.
Answer
Variables in the Equation
B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)
Lower Upper
Step 1a
Age1 143.281 3 .000
Age1(1) .798 .094 72.777 1 .000 2.221 1.849 2.668
Age1(2) .565 .097 34.215 1 .000 1.760 1.456 2.127
Age1(3) .041 .106 .146 1 .702 1.041 .846 1.282
GENDER(1) .616 .052 138.597 1 .000 1.852 1.671 2.052
LIVING_ARRANGE 2.264 2 .322
LIVING_ARRANGE(1) -.093 .062 2.262 1 .133 .911 .807 1.029
LIVING_ARRANGE(2) -.042 .079 .287 1 .592 .959 .822 1.119
FEE_STATUS(1) -.524 .077 46.323 1 .000 .592 .509 .688
dist_driving -.057 .049 1.391 1 .238 .944 .858 1.039
driver_agg .626 .024 668.411 1 .000 1.871 1.784 1.962
thrill .551 .079 48.006 1 .000 1.734 1.484 2.027
risk_accep .615 .010 3981.108 1 .000 1.849 1.814 1.885
Constant -18.369 .356 2668.265 1 .000 .000
a. Variable(s) entered on step 1: Age1, GENDER, LIVING_ARRANGE, FEE_STATUS, dist_driving, driver_agg, thrill,
risk_accep.
b. How do you interpret the findings for living arrangements?
Answer
Living arrangement is given as a categorical variable. There are two categorical
variables for the living arrangement presented in the table. The first is related to the
college/student accommodation while the second is related to students who live
independently. The odds ratio for the students who live in college or in student

accommodation is -0.093. This means that students who live in college or student
accommodation have a lower probability of about 8.9% to have a road traffic accident
as compared to those students who live at home with their parents.
The odds ratio for the students who live in independently is -0.042. This means that
students who live independently have low probability of about 4.1% to have a road
traffic accident as compared to those students who live at home with their parents.
c. Provide a preliminary interpretation and conclusion about the influence of the
predictor variables on road traffic accidents. In particular, what is the effect of
changes (increase or decrease) in “driver aggression” on the odds of experiencing a
road traffic accident?
Answer
The aim of this section was to try and predict the chances of having a road traffic
accident. 8 variables were used to predict the chances of having a road traffic
accident. Results showed that 6 out of the 8 predictor variables included in the model
significantly predicts the chances of having a road traffic accident. The coefficient for
the driver aggression is positive. This means that driver aggression has positive odds
on predicting the road traffic accident and as such an increase in the driver aggression
would definitely result to an increase in the chances of road traffic accident.
6. Using “OB” (obesity at third-year follow-up) as the dependent variable, perform a binary
logistic regression.

a. Provide a table of odds ratios, confidence intervals and significance values for each
predictor variable.
Answer
Variables in the Equation
B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)
Lower Upper
Step 1a
Age1 73.524 3 .000
Age1(1) .465 .071 43.345 1 .000 1.592 1.386 1.828
Age1(2) .329 .073 20.138 1 .000 1.390 1.204 1.605
Age1(3) .083 .081 1.053 1 .305 1.087 .927 1.274
GENDER(1) .370 .039 90.949 1 .000 1.447 1.341 1.561
LIVING_ARRANGE 15.164 2 .001
LIVING_ARRANGE
(1)
-.097 .044 4.825 1 .028 .907 .832 .990
LIVING_ARRANGE
(2)
.079 .055 2.066 1 .151 1.083 .972 1.206
BL_owob(1) .030 .037 .663 1 .416 1.030 .959 1.107
depression(1) -2.695 .056 2322.441 1 .000 .068 .061 .075
edu_par -3.201 .062 2653.795 1 .000 .041 .036 .046
Constant 1.021 .081 157.509 1 .000 2.776
a. Variable(s) entered on step 1: Age1, GENDER, LIVING_ARRANGE, BL_owob, depression, edu_par.
b. Again, how do you interpret the findings for living arrangements?
Answer
Living arrangement is given as a categorical variable. There are two categorical
variables for the living arrangement presented in the table. The first is related to the
college/student accommodation while the second is related to students who live
independently. The odds ratio for the students who live in college or in student
accommodation is -0.097. This means that students who live in college or student

accommodation have low probability of about 9.3% to have obesity as compared to
those students who live at home with their parents.
The odds ratio for the students who live in independently is 0.079. This means that
students who live independently have a higher probability of about 8.3% to have
obesity as compared to those students who live at home with their parents.
c. Provide a preliminary interpretation and conclusion about the influence of the
predictor variables on obesity at follow-up. In particular, what is the effect of changes
(increase or decrease) in the number of parents with university education on the
likelihood of later obesity?
Answer
The aim of this section was to try and predict the chances of having later obesity. 6
variables were used to predict the chances of having a later obesity (OB). Results
showed that 5 out of the 6 predictor variables included in the model significantly
predicts the chances of having a road traffic accident. The coefficient for the number
of parents with university education is negative, i.e., -3.20. This means that number of
parents with university education has negative odds on predicting the chances of later
obesity (OB) and as such an increase in the number of parents with university
education would definitely result to a decrease in the chances of having later obesity.

Appendix
Age group
Frequency Percent Valid Percent Cumulative
Percent
Valid
< 18 yrs. old 17760 45.9 45.9 45.9
19-21 yrs. old 11672 30.2 30.2 76.1
22-25 yrs. old 5494 14.2 14.2 90.3
> 26 yrs. old 3755 9.7 9.7 100.0
Total 38681 100.0 100.0
cohort
Frequency Percent Valid Percent Cumulative
Percent
Valid 2005 3259 8.4 8.4 8.4
2006 3615 9.3 9.3 17.8
2007 3944 10.2 10.2 28.0
2008 4086 10.6 10.6 38.5
2009 5010 13.0 13.0 51.5
2010 5687 14.7 14.7 66.2

2011 6383 16.5 16.5 82.7
2012 6697 17.3 17.3 100.0
Total 38681 100.0 100.0
STATE
Frequency Percent Valid Percent Cumulative
Percent
Valid
NSW 15860 41.0 41.0 41.0
Victoria 13571 35.1 35.1 76.1
Queensland 7528 19.5 19.5 95.5
ACT 1722 4.5 4.5 100.0
Total 38681 100.0 100.0
Age group
Frequency Percent Valid Percent Cumulative
Percent
Valid
< 18 yrs. old 17760 45.9 45.9 45.9
19-21 yrs. old 11672 30.2 30.2 76.1
22-25 yrs. old 5494 14.2 14.2 90.3
> 26 yrs. old 3755 9.7 9.7 100.0
Total 38681 100.0 100.0
LIVING_ARRANGE
Frequency Percent Valid Percent Cumulative
Percent
Valid
At home 20840 53.9 53.9 53.9
College/student accom 6850 17.7 17.7 71.6
Independently 10991 28.4 28.4 100.0
Total 38681 100.0 100.0
FACULTY

Frequency Percent Valid Percent Cumulative
Percent
Valid
Arts and Sciences 9004 23.3 23.3 23.3
Education 15038 38.9 38.9 62.2
Health Sciences 11729 30.3 30.3 92.5
Theology and Philosophy 588 1.5 1.5 94.0
Business 2322 6.0 6.0 100.0
Total 38681 100.0 100.0
DEGREE_TYPE
Frequency Percent Valid Percent Cumulative
Percent
Valid
Single 34620 89.5 89.5 89.5
Double 4061 10.5 10.5 100.0
Total 38681 100.0 100.0
METRO
Frequency Percent Valid Percent Cumulative
Percent
Valid
Metro 27223 70.4 84.4 84.4
Non-metro 5015 13.0 15.6 100.0
Total 32238 83.3 100.0
Missing System 6443 16.7
Total 38681 100.0
STUDY_MODE
Frequency Percent Valid Percent Cumulative
Percent
Valid
FT 34770 89.9 89.9 89.9
PT 3911 10.1 10.1 100.0
Total 38681 100.0 100.0

FEE_STATUS
Frequency Percent Valid Percent Cumulative
Percent
Valid
Domestic 32238 83.3 83.3 83.3
International 6443 16.7 16.7 100.0
Total 38681 100.0 100.0
Group Statistics
GENDER N Mean Std. Deviation Std. Error Mean
thrill Male 10449 5.00 1.217 .012
Female 28232 5.00 1.226 .007
Independent Samples Test
Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Difference
Std. Error
Difference
95%
Inte
D
Lowe
thrill
Equal variances assumed .847 .357 -.370 38679 .711 -.005 .014 -.0
Equal variances not assumed -.371 18783.25
0
.710 -.005 .014 -.0
Group Statistics
METRO N Mean Std. Deviation Std. Error Mean
thrill Metro 27223 5.01 1.224 .007
Non-metro 5015 4.99 1.207 .017
Independent Samples Test

Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Difference
Std. Error
Difference
95%
Inte
Di
Lower
thrill Equal variances assumed 1.845 .174 .686 32236 .493 .013 .019 -.0
Equal variances not assumed .692 7048.178 .489 .013 .019 -.0
Group Statistics
STUDY_MODE N Mean Std. Deviation Std. Error Mean
thrill FT 34770 5.00 1.224 .007
PT 3911 5.00 1.225 .020
Independent Samples Test
Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Difference
Std. Error
Difference
95%
Inte
Di
Lower
thrill Equal variances assumed .222 .637 .132 38679 .895 .003 .021 -.0
Equal variances not assumed .132 4829.635 .895 .003 .021 -.0
Group Statistics
GENDER N Mean Std. Deviation Std. Error Mean
driver_agg Male 10449 7.52 4.335 .042
Female 28232 7.51 4.347 .026
Independent Samples Test

Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Difference
Std. Error
Difference
driver_agg Equal variances assumed .117 .732 .083 38679 .934 .004 .050
Equal variances not assumed .083 18712.803 .934 .004 .050
Group Statistics
GENDER N Mean Std. Deviation Std. Error Mean
risk_accep Male 10449 8.59 4.358 .043
Female 28232 8.51 4.357 .026
Independent Samples Test
Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
9
L
risk_accep
Equal variances assumed .054 .817 1.571 38679 .116 .078 .050
Equal variances not assumed 1.571 18663.18
0
.116 .078 .050
Group Statistics
METRO N Mean Std. Deviation Std. Error Mean
risk_accep Metro 27223 8.53 4.358 .026
Non-metro 5015 8.59 4.304 .061
Independent Samples Test

Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
9
Lo
risk_accep Equal variances assumed 3.228 .072 -.866 32236 .386 -.058 .067
Equal variances not assumed -.874 7040.476 .382 -.058 .066
Group Statistics
METRO N Mean Std. Deviation Std. Error Mean
driver_agg Metro 27223 7.53 4.345 .026
Non-metro 5015 7.49 4.303 .061
Independent Samples Test
Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
95
In
Low
driver_agg Equal variances assumed 1.060 .303 .714 32236 .475 .048 .067
Equal variances not assumed .719 7029.087 .472 .048 .066
Group Statistics
STUDY_MODE N Mean Std. Deviation Std. Error Mean
driver_agg FT 34770 7.51 4.344 .023
PT 3911 7.54 4.339 .069
Independent Samples Test

Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
95
Lo
driver_agg Equal variances assumed .323 .570 -.309 38679 .757 -.023 .073
Equal variances not assumed -.310 4834.453 .757 -.023 .073
Group Statistics
STUDY_MODE N Mean Std. Deviation Std. Error Mean
risk_accep FT 34770 8.51 4.355 .023
PT 3911 8.68 4.374 .070
Independent Samples Test
Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
L
risk_accep Equal variances assumed .045 .832 -2.269 38679 .023 -.167 .073
Equal variances not assumed -2.261 4823.706 .024 -.167 .074
Group Statistics
RTA_one_crash N Mean Std. Deviation Std. Error Mean
driver_agg No RTAs 33628 6.79 4.116 .022
One RTA or more 5053 12.34 2.218 .031
Independent Samples Test

Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
driver_agg
Equal variances assumed 3179.609 .000 -93.863 38679 .000 -5.552 .059
Equal variances not assumed -144.454 11183.46
6
.000 -5.552 .038
Group Statistics
RTA_one_crash N Mean Std. Deviation Std. Error Mean
thrill No RTAs 33628 4.80 1.159 .006
One RTA or more 5053 6.34 .686 .010
Independent Samples Test
Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
95% C
Interv
Diff
Lower
thrill
Equal variances assumed 1715.363 .000 -92.063 38679 .000 -1.539 .017 -1.572
Equal variances not assumed -133.493 10036.69
7
.000 -1.539 .012 -1.562
Group Statistics
RTA_one_crash N Mean Std. Deviation Std. Error Mean
risk_accep No RTAs 33628 7.91 4.210 .023
One RTA or more 5053 12.68 2.748 .039
Independent Samples Test

Levene's Test for
Equality of Variances
t-test for Equality of Means
F Sig. t df Sig. (2-
tailed)
Mean
Differenc
e
Std. Error
Differenc
e
L
risk_accep Equal variances assumed 1951.956 .000 -78.154 38679 .000 -4.775 .061
Equal variances not assumed -106.209 9076.181 .000 -4.775 .045
GENDER * depression Crosstabulation
Count
depression Total
Not depressed Depressed
GENDER Male 9365 1084 10449
Female 25355 2877 28232
Total 34720 3961 38681
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square .280a 1 .597
Continuity Correctionb .260 1 .610
Likelihood Ratio .279 1 .597
Fisher's Exact Test .598 .305
Linear-by-Linear Association .280 1 .597
N of Valid Cases 38681
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 1070.00.
b. Computed only for a 2x2 table
METRO * depression Crosstabulation
Count
depression Total
Not depressed Depressed

METRO Metro 24443 2780 27223
Non-metro 4495 520 5015
Total 28938 3300 32238
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square .114a 1 .736
Continuity Correctionb .097 1 .755
Likelihood Ratio .113 1 .737
Fisher's Exact Test .743 .378
Linear-by-Linear Association .114 1 .736
N of Valid Cases 32238
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 513.35.
b. Computed only for a 2x2 table
STUDY_MODE * depression Crosstabulation
Count
depression Total
Not depressed Depressed
STUDY_MODE FT 31241 3529 34770
PT 3479 432 3911
Total 34720 3961 38681
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square 3.072a 1 .080
Continuity Correctionb 2.975 1 .085
Likelihood Ratio 3.011 1 .083
Fisher's Exact Test .084 .042
Linear-by-Linear Association 3.072 1 .080
N of Valid Cases 38681
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 400.49.

b. Computed only for a 2x2 table
FEE_STATUS * depression Crosstabulation
Count
depression Total
Not depressed Depressed
FEE_STATUS Domestic 28938 3300 32238
International 5782 661 6443
Total 34720 3961 38681
Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
Exact Sig. (2-
sided)
Exact Sig. (1-
sided)
Pearson Chi-Square .003a 1 .956
Continuity Correctionb .001 1 .974
Likelihood Ratio .003 1 .956
Fisher's Exact Test .966 .487
Linear-by-Linear Association .003 1 .956
N of Valid Cases 38681
a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 659.77.
b. Computed only for a 2x2 table
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a
Age1 -.303 .026 135.691 1 .000 .739
GENDER -.617 .052 139.293 1 .000 .539
LIVING_ARRANGE .049 .030 2.638 1 .104 1.051
FEE_STATUS .523 .066 63.082 1 .000 1.688
dist_driving -.057 .049 1.353 1 .245 .945
driver_agg .625 .024 667.888 1 .000 1.869
thrill .551 .079 48.169 1 .000 1.735
risk_accep .614 .010 3980.888 1 .000 1.848
Constant -17.547 .336 2731.688 1 .000 .000

a. Variable(s) entered on step 1: Age1, GENDER, LIVING_ARRANGE, FEE_STATUS, dist_driving, driver_agg,
thrill, risk_accep.
Categorical Variables Codings
Frequency Parameter coding
(1) (2) (3)
Age group
< 18 yrs. old 17760 1.000 .000 .000
19-21 yrs. old 11672 .000 1.000 .000
22-25 yrs. old 5494 .000 .000 1.000
> 26 yrs. old 3755 .000 .000 .000
LIVING_ARRANGE
At home 20840 1.000 .000
College/student accom 6850 .000 1.000
Independently 10991 .000 .000
FEE_STATUS Domestic 32238 1.000
International 6443 .000
GENDER Male 10449 1.000
Female 28232 .000
Variables in the Equation
B S.E. Wald df Sig. Exp(B) 95% C.I.for
Lower
Step 1a Age1 143.281 3 .000
Age1(1) .798 .094 72.777 1 .000 2.221 1.849
Age1(2) .565 .097 34.215 1 .000 1.760 1.456
Age1(3) .041 .106 .146 1 .702 1.041 .846
GENDER(1) .616 .052 138.597 1 .000 1.852 1.671
LIVING_ARRANGE 2.264 2 .322
LIVING_ARRANGE(1) -.093 .062 2.262 1 .133 .911 .807
LIVING_ARRANGE(2) -.042 .079 .287 1 .592 .959 .822
FEE_STATUS(1) -.524 .077 46.323 1 .000 .592 .509
dist_driving -.057 .049 1.391 1 .238 .944 .858
driver_agg .626 .024 668.411 1 .000 1.871 1.784
thrill .551 .079 48.006 1 .000 1.734 1.484
risk_accep .615 .010 3981.108 1 .000 1.849 1.814

Constant -18.369 .356 2668.265 1 .000 .000
a. Variable(s) entered on step 1: Age1, GENDER, LIVING_ARRANGE, FEE_STATUS, dist_driving, driver_agg, thrill, risk_accep.
Categorical Variables Codings
Frequency Parameter coding
(1) (2) (3)
Age group
< 18 yrs. old 17760 1.000 .000 .000
19-21 yrs. old 11672 .000 1.000 .000
22-25 yrs. old 5494 .000 .000 1.000
> 26 yrs. old 3755 .000 .000 .000
LIVING_ARRANGE
At home 20840 1.000 .000
College/student accom 6850 .000 1.000
Independently 10991 .000 .000
GENDER Male 10449 1.000
Female 28232 .000
Categorical Variables Codings
Frequency Parameter coding
(1) (2) (3)
Age group
< 18 yrs. old 17760 1.000 .000 .000
19-21 yrs. old 11672 .000 1.000 .000
22-25 yrs. old 5494 .000 .000 1.000
> 26 yrs. old 3755 .000 .000 .000
LIVING_ARRANGE
At home 20840 1.000 .000
College/student accom 6850 .000 1.000
Independently 10991 .000 .000
OB Not obese 33628 1.000
Obese 5053 .000
GENDER Male 10449 1.000
Female 28232 .000
depression Not depressed 34720 1.000
Depressed 3961 .000
owob_par No obese parents 14455 1.000
At least one obese parent 24226 .000

Variables not in the Equation
Score df Sig.
Step 0 Variables
Age1 55.040 3 .000
Age1(1) 29.033 1 .000
Age1(2) .584 1 .445
Age1(3) 22.478 1 .000
GENDER(1) 65.962 1 .000
LIVING_ARRANGE 12.192 2 .002
LIVING_ARRANGE(1) 1.083 1 .298
LIVING_ARRANGE(2) 11.598 1 .001
OB(1) 38681.000 1 .000
depression(1) 2796.202 1 .000
edu_par 4913.920 1 .000
owob_par(1) 3315.468 1 .000
Overall Statistics 38681.000 10 .000