Analysis of Dataset in Biostatistics

This document provides an analysis of a dataset in Biostatistics. It includes the mean, standard deviation, minimum and maximum values for age, as well as the frequency of students in different age categories. It also includes descriptive statistics for demographic information and testing for differences 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
The average age of the ACU cohort is 20.5 years old with a standard deviation of 4.89.
The minimum and maximum ages are 16 years old and 59 years old respectively.
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
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
2. Descriptive statistics for the demographic information of the students
Answer
Demographic characteristics Frequency (n) Percent (%)
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
State
NSW 15860 41.0
Victoria 13571 35.1
Queensland 7528 19.5
ACT 1722 4.5
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
Living Arrangement
At home 20840 53.9
College/student
accommodation
6850 17.7
Independently 10991 28.4
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
Degree Type
Single 34620 89.5
Double 4061 10.5
Total 38681 100.0
Metro
Metro 27223 70.4
Non-metro 5015 13.0

Missing 6443 16.7
Total 38681 100.0
Study Type
FT 34770 89.9
PT 3911 10.1
Total 38681 100.0
Fees Status
Domestic 32238 83.3
International 6443 16.7
Total 38681 100.0
The above table gives the frequencies for the demographic characteristics. As can be
seen, majority were the 2012 cohorts (17.3%, n = 6697) and most of the participants
came from NSW state (41.0%, n = 15860). In terms of age, majority were aged either 18
years old or less (45.9%, n = 17760). Most common living arrangement was living at
home (53.9%, n = 20840) and majority of the students were in the education faculty
(38.9%, n = 15038) with most students taking single degree programs (89.5%, n =
34620).
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 18712.8
03
.934 .004 .050 -.093 .102
An independent samples t-test was performed to compare the mean aggression scores for
the males and the female respondents. Results showed that the males (M = 7.52, SD =
4.34, N = 10449) had no significant difference in terms of the mean aggression scores
when compared to the females (M = 7.51, SD = 4.35, N = 28232), t (38679) = .083, p
> .05, two-tailed. The mean aggression scores for the males and the females are the same.
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
ce
95% Confidence
Interval of the
Difference
Lower Upper
thrill
Equal variances
assumed
.847 .357 -.370 38679 .711 -.005 .014 -.033 .022
Equal variances
not assumed
-.371 18783.2
50
.710 -.005 .014 -.033 .022
An independent samples t-test was performed to compare the mean thrill seeking scores
for the males and the female respondents. Results showed that the males (M = 5.00, SD =
1.22, N = 10449) had no significant difference in terms of the mean thrill seeking scores
when compared to the females (M = 5.00, SD = 1.23, N = 28232), t (38679) = .711, p
> .05, two-tailed. The mean thrill seeking scores for the males and the females are the
same.
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 18663.1
80
.116 .078 .050 -.019 .176
An independent samples t-test was performed to compare the mean risk acceptance
scores for the males and the female respondents. Results showed that the males (M =
8.59, SD = 4.36, N = 10449) had no significant difference in terms of the mean risk
acceptance scores when compared to the females (M = 8.51, SD = 4.36, N = 28232), t
(38679) = .116, p > .05, two-tailed. The mean risk acceptance scores for the males and
the females are the same.
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 7029.08
7
.472 .048 .066 -.082 .177
An independent samples t-test was performed to compare the mean aggression scores for
the metro and the non-metro respondents. Results showed that the metro (M = 7.53, SD =
4.35, N = 27223) had no significant difference in terms of the mean aggression scores
when compared to the non-metro (M = 7.49, SD = 4.30, N = 5015), t (32236) = .714, p
> .05, two-tailed. The mean aggression scores for the metro and the non-metro are the
same.
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
An independent samples t-test was performed to compare the mean thrill seeking scores
for the metro and the non-metro respondents. Results showed that the metro (M = 5.01,

SD = 1.22, N = 27223) had no significant difference in terms of the mean thrill seeking
scores when compared to the non-metro (M = 4.99, SD = 1.21, N = 5015), t (32236)
= .686, p > .05, two-tailed. The mean thrill seeking scores for the metro and the non-
metro are the same.
Difference in mean thrill seeking 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
An independent samples t-test was performed to compare the mean risk acceptance
scores for the metro and the non-metro respondents. Results showed that the metro (M =
8.53, SD = 4.36, N = 27223) had no significant difference in terms of the mean risk

acceptance scores when compared to the non-metro (M = 8.59, SD = 4.30, N = 5015), t
(32236) = -.866, p > .05, two-tailed. The mean risk acceptance scores for the metro and
the non-metro are the same.
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
An independent samples t-test was performed to compare the mean aggression scores for
the full time (FT) and the part time (PT) respondents. Results showed that the Full time
(M = 7.51, SD = 4.34, N = 34770) had no significant difference in terms of the mean
aggression scores when compared to the part time (M = 7.54, SD = 4.34, N = 3911), t
(38679) = .895, p > .05, two-tailed. The mean aggression scores for the full time and the
part time are the same.

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
An independent samples t-test was performed to compare the mean thrill seeking scores
for the full time (FT) and the part time (PT) respondents. Results showed that the Full
time (M = 5.00, SD = 1.22, N = 34770) had no significant difference in terms of the mean
thrill seeking scores when compared to the part time (M = 5.00, SD = 1.23, N = 3911), t
(38679) = .895, p > .05, two-tailed. The mean thrill seeking scores for the full time and
the part time are the same.
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
An independent samples t-test was performed to compare the mean risk acceptance
scores for the full time (FT) and the part time (PT) respondents. Results showed that the
Full time (M = 8.51, SD = 4.36, N = 34770) had no significant difference in terms of the
mean risk acceptance scores when compared to the part time (M = 8.68, SD = 4.37, N =
3911), t (38679) = .895, p > .05, two-tailed. The mean risk acceptance scores for the full
time and the part time are the same.
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
An independent samples t-test was performed to compare the mean aggression scores for
the no RTAs and one RTA or more. Results showed that the no RTAs (M = 6.79, SD =
4.12, N = 33628) had significantly lower mean aggression scores as compared to one
RTA or more (M = 12.34, SD = 2.22, N = 5053), t (38679) = -93.86, p < .05, two-tailed.
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
An independent samples t-test was performed to compare the mean thrill seeking scores
for the no RTAs and one RTA or more. Results showed that the no RTAs (M = 4.80, SD
= 1.16, N = 33628) had significantly lower mean aggression scores as compared to one
RTA or more (M = 6.34, SD = .69, N = 5053), t (38679) = -92.06, p < .05, two-tailed.
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
An independent samples t-test was performed to compare the mean risk acceptance
scores for the no RTAs and one RTA or more. Results showed that the no RTAs (M =

7.91, SD = 4.21, N = 33628) had significantly lower mean aggression scores as compared
to one RTA or more (M = 12.68, SD = 2.75, N = 5053), t (38679) = -78.15, p < .05, two-
tailed.
4. In terms of depression, is there a difference between the depressed and not depressed
groups according to their:
a. Gender
Answer
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
A Chi-square test of association was performed to check on whether there is significant
association between gender and depression status. Results showed that there is no
significant association between gender and depression status ( χ2 ( 1 ) =.28 , p=.597).

b. Metropolitan background status
Answer
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

A Chi-square test of association was performed to check on whether there is significant
association between metropolitan background status and depression status. Results
showed that there is no significant association between metropolitan background status
and depression status ( χ2 ( 1 ) =.114 , p=.736).
c. Study mode
Answer
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
A Chi-square test of association was performed to check on whether there is significant
association between study mode and depression status. Results showed that there is no
significant association between study mode and depression status ( χ2 ( 1 ) =3.07 , p=.080).
d. Fee status
Answer
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
A Chi-square test of association was performed to check on whether there is significant
association between fees status and depression status. Results showed that there is no
significant association between fees status and depression status ( χ2 ( 1 )=.003 , p=.956).
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
The coefficient for living arrangement(1) which is the College/student
accommodation is -0.093; this shows that students who live in college or student
accommodation are less likely to have obesity as compared to those who live at home
by about 8.9%. Also the coefficient for living arrangement(2) which is independently

is -0.042; this shows that students who live independently are less likely to have
obesity as compared to those who live at home by about 4.1%.
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
In overall, the results of the logistic regression shows that 6 out of the 8 predictor
variables were significantly predicting the likelihood of road traffic accident (RTA).
The insignificant predictors were living arrangement and the driving distance.
The coefficient of driver aggression is 0.626; this implies that an increase in the
driver aggression scores results to an increase in the odds of experiencing a 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
The coefficient for living arrangement(1) which is the College/student
accommodation is -0.097; this shows that students who live in college or student
accommodation are less likely to have obesity as compared to those who live at home
by about 9.3%. On the other hand, the coefficient for living arrangement(2) which is
independently is 0.079; this shows that students who live independently are more
likely to have obesity as compared to those who live at home by about 8.3%.
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
In overall, the results of the logistic regression shows that only Overweight or obese
at baseline as a predictor variable was insignificant in predicting the obesity at third-
year follow-up (OB). The rest of predictors were significant in the model.

The coefficient of number of parents with university education is -3.201; this implies
that an increase in the number of parents with university education results to a
decrease in the odds of experiencing later obesity.