Statistical Analysis And Findings

1. Provide a table of odds ratios, confidence intervals and significance values for each predictor variable.
2. How do you interpret the findings for living arrangements?
3. 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?

1. Provide a table of odds ratios, confidence intervals and significance values for each predictor variable.
2. Again, how do you interpret the findings for living arrangements?
3. 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?

Distinguish between different statistical tests, especially in terms of application and interpretation

Develop a sound statistical approach to the analysis and interpretation of public health data and communicate findings in an academic-standard output

Critique public health research on the basis of its statistical methods, analysis and interpretation

Question one

1. Age descriptive statistics

Descriptive Statistics
	N	Minimum	Maximum	Mean	Std. Deviation
AGE	38681	16	59	20.50	4.889
Valid N (listwise)	38681

Table 1

It can be observed from the table above that the mean age for the participants was 20.5. The youngest participant was 16 years old while the oldest participant was 59 years old.

1. Frequency for new age category

Statistics
Age category
N	Valid	38681
	Missing	0

Table 2

Age category
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	18 years	11881	30.7	30.7	30.7
	19 - 21 years	11666	30.2	30.2	60.9
	22 - 25 years	5494	14.2	14.2	75.1
	26 or more	3755	9.7	9.7	84.8
	system missing	5885	15.2	15.2	100.0
	Total	38681	100.0	100.0

Table 3

The table above table shows the frequency of age groups. Participants who were 18 years old were 11,881 representing 30.7%. This was followed closely by those within the age of 19 to 21 years. They were 11,666 representing 30.2%. Those who were 26 years old and above were 3,755, representing 9.7% of the total.

Question two

Age descriptive statistics

Descriptive Statistics
	N	Sum	Mean
AGE	38681	792845	20.50
Valid N (listwise)	38681

Table 4

The mean age of the participants was 20.5 while the sum total of their age was 792,845 years.

State descriptive

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

Table 5

From the table above, it can be observed that 41% of the participants come from NSW, 35.1% come from Victoria, and 19.5% come from Queensland while the minority of the participants (4.5%) comes from ACT.

GENDER
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	Male	10449	27.0	27.0	27.0
	Female	28232	73.0	73.0	100.0
	Total	38681	100.0	100.0

Table 6

It can be observed that majority of the participants were females. They were 28,232 in number and represented 73%. The rest were females who were 10,449 representing 27%.

LIVING_ARRANGE
		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	At home	20840	53.9	53.9	53.9
	College/student accommodation	6850	17.7	17.7	71.6
	Independently	10991	28.4	28.4	100.0
	Total	38681	100.0	100.0

Table 7

The table above shows the how participants are accommodated. It can be observed that 53.9% (20,840) were being accommodated from their homes. 17.7% (6,850) were accommodated at the college while 28.4% (10,991) had their own independent accommodation.

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

Table 8

The table above shows the distribution of the student participants based on their faculties. Majority of them came from the faculty of education (15,038) representing 38.9%. This is followed by students from the faculty of health sciences (11,729) who represented 30.3%. The least number of students came from the faculty of theology and philosophy. They were 588 representing 1.5%.

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

Table 9

The table above shows distribution of participants by the type of their degrees. It can be observed that 89.5% were pursuing single degrees while 10.5% were pursuing double degrees.

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

Table 10

The table above shows the location of origin of the students. It can be observed that majority of them came from metropolitan areas (70.4%) while 13% came from non-metropolitan areas.

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

Table 11

From the table above, it can be observed that 89.9% of the students pursued full time studies while 10.1% pursued part time studies.

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

Table 12

It can be observed that 83.3% (32,238) of the students are domestic students while 16.7% (6,443) are international students.

1. Test for the difference in mean for aggression, thrill seeking and risk acceptance scores by gender

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% 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	18712.803	.934	.004	.050	-.093	.102
thrill	Equal variances assumed	.847	.357	-.370	38679	.711	-.005	.014	-.033	.022
	Equal variances not assumed			-.371	18783.250	.710	-.005	.014	-.033	.022
risk_accep	Equal variances assumed	.054	.817	1.571	38679	.116	.078	.050	-.019	.176
	Equal variances not assumed			1.571	18663.180	.116	.078	.050	-.019	.176

Table 13

From the t-test table results above, it can be observed that the p-values computed are large compared to the level of significance (0.05). This means that the mean aggression, thrill seeking and risk acceptance scores do not differ by gender.

1. Test for the difference in mean for aggression, thrill seeking and risk acceptance scores by metropolitan background status

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% 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	7029.087	.472	.048	.066	-.082	.177
thrill	Equal variances assumed	1.845	.174	.686	32236	.493	.013	.019	-.024	.050
	Equal variances not assumed			.692	7048.178	.489	.013	.019	-.024	.049
risk_accep	Equal variances assumed	3.228	.072	-.866	32236	.386	-.058	.067	-.189	.073
	Equal variances not assumed			-.874	7040.476	.382	-.058	.066	-.188	.072

Table 14

From the t-test table results above, it can be observed that the p-values computed are large (0.3, 0.17 and 0.07) compared to the level of significance (0.05). This means that the mean aggression, thrill seeking and risk acceptance scores do not differ by metropolitan background status.

1. Test for the difference in mean for aggression, thrill seeking and risk acceptance scores by study mode.

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% 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	4834.453	.757	-.023	.073	-.166	.121
thrill	Equal variances assumed	.222	.637	.132	38679	.895	.003	.021	-.038	.043
	Equal variances not assumed			.132	4829.635	.895	.003	.021	-.038	.043
risk_accep	Equal variances assumed	.045	.832	-2.269	38679	.023	-.167	.073	-.311	-.023
	Equal variances not assumed			-2.261	4823.706	.024	-.167	.074	-.311	-.022

Table 15

From the t-test table results above, it can be observed that the p-values computed are large (0.57, 0.63 and 0.83) compared to the level of significance (0.05). This means that the mean aggression, thrill seeking and risk acceptance scores do not differ by study mode.

1. Test for the difference in mean for aggression, thrill seeking and risk acceptance scores by RTA (follow up survey)

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% Confidence Interval of the Difference
									Lower	Upper
driver_agg	Equal variances assumed	3179.609	.000	-93.863	38679	.000	-5.552	.059	-5.668	-5.436
	Equal variances not assumed			-144.454	11183.466	.000	-5.552	.038	-5.627	-5.476
thrill	Equal variances assumed	1715.363	.000	-92.063	38679	.000	-1.539	.017	-1.572	-1.507
	Equal variances not assumed			-133.493	10036.697	.000	-1.539	.012	-1.562	-1.517
risk_accep	Equal variances assumed	1951.956	.000	-78.154	38679	.000	-4.775	.061	-4.895	-4.655
	Equal variances not assumed			-106.209	9076.181	.000	-4.775	.045	-4.863	-4.687

Table 16

From the t-test table results above, it can be observed that the p-values computed are less (0.00) compared to the level of significance (0.05). This means that the mean aggression, thrill seeking and risk acceptance scores differ significantly by RTA.

Question four

1. Depression by gender

ANOVA
depression
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	.026	1	.026	.280	.597
Within Groups	3555.361	38679	.092
Total	3555.387	38680

Table 17

The anova results show that the computed p-value (0.57) is greater compared to the level of significance (0.05). This means that the null hypothesis is accepted. It is concluded therefore that null hypothesis is significant at 95% level of confidence.

1. Depression by metropolitan background status

ANOVA
depression
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	.010	1	.010	.114	.736
Within Groups	2962.189	32236	.092
Total	2962.200	32237

Table 18

The anova results show that the computed p-value (0.736) is greater compared to the level of significance (0.05). This means that the null hypothesis is accepted. It is concluded therefore that null hypothesis is significant at 95% level of confidence.

1. Depression by study mode

Results table

ANOVA
depression
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	.282	1	.282	3.072	.080
Within Groups	3555.105	38679	.092
Total	3555.387	38680

Table 19

The anova results show that the computed p-value (0.08) is greater compared to the level of significance (0.05). This means that the null hypothesis is accepted. It is concluded therefore that null hypothesis is significant at 95% level of confidence.

1. Depression by fee status

Results table

ANOVA
depression
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	.000	1	.000	.003	.956
Within Groups	3555.387	38679	.092
Total	3555.387	38680

Table 20

The anova results show that the computed p-value (0.956) is greater compared to the level of significance (0.05). This means that the null hypothesis is accepted. It is concluded therefore that null hypothesis is significant at 95% level of confidence.

Question five

• Binary logistic regression (RTA and Demographics).

Table of results

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)
Step 1a	Age_category	-.003	.000	42.072	1	.000	.997
	GENDER	-.262	.033	64.299	1	.000	.769
	LIVING_ARRANGE	-.049	.019	6.801	1	.009	.952
	FEE_STATUS	.248	.041	36.290	1	.000	1.282
	Constant	-1.673	.031	2885.476	1	.000	.188
a. Variable(s) entered on step 1: Age_category, GENDER, LIVING_ARRANGE, FEE_STATUS. Table 21

From the results table above, it can be observed that the value of the coefficient for the living arrangement is -0.049. This value is close to zero. It is an indication that there is no association between RTA and living arrangement. To add on, the odds of the predictor variables are tending towards 1, this is an indication that they cause a great variation in RTA if they are increased.

• Binary logistic regression (RTA and driving distance).

Results table

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)
Step 1a	dist_driving	-.016	.031	.268	1	.605	.984
	Constant	-1.885	.024	5937.971	1	.000	.152
a. Variable(s) entered on step 1: dist_driving.

Table 22

From the results table above, it can be observed that the value of the coefficient for the driving distance is -0.016. This value is close to zero. It is an indication that there is no association between RTA and driving distance.  The odd of the predictor variable is 0.94 indicating a strong influence on RTA.  

• Binary logistic regression (RTA with aggression, thrill seeking and risk acceptance).

Table of results

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)
Step 1a	Driver aggression	.612	.024	661.998	1	.000	1.844
	thrill	.516	.078	43.584	1	.000	1.675
	Risk acceptance	.596	.009	4017.531	1	.000	1.815
	Constant	-17.579	.327	2887.675	1	.000	.000
a. Variable(s) entered on step 1: driver_agg, thrill, risk_accep.

Table 23

From the results table above, it can be observed that the values of the coefficients for the predictor variables are 0.61, 0.52 and 0.596.  It is an indication that there are significant associations between RTA and predictor variables. To add on, the odds of the predictor variables are tending towards 1, this is an indication that they cause a great variation in RTA if they are increased.

Question six

• Binary logistic regression (OB and Demographics).

Results table

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)
Step 1a	Age_category	-.003	.000	40.576	1	.000	.997
	GENDER	-.267	.033	66.548	1	.000	.766
	LIVING_ARRANGE	-.012	.017	.447	1	.504	.988
	Constant	-1.654	.031	2863.657	1	.000	.191
a. Variable(s) entered on step 1: Age_category, GENDER, LIVING_ARRANGE.

Table 24

From the results table above, it can be observed that the value of the coefficient for the living arrangement is -0.012. This value is close to zero. It is an indication that there is no association between obesity at third year follow up and living arrangement.

• Binary logistic regression (OB and overweight and depression).

Table of results

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)
Step 1a	depression	1.787	.037	2325.596	1	.000	5.970
	BL_owob	-.017	.032	.270	1	.603	.983
	Constant	-2.186	.027	6644.566	1	.000	.112
a. Variable(s) entered on step 1: depression, BL_owob.

Table 25

• Binary logistic regression (OB and edu_par and presence or absence of obese).

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)
Step 1a	owob_par	1.881	.169	124.423	1	.000	6.561
	edu_par	-2.210	.061	1331.774	1	.000	.110
	Constant	-2.826	.170	277.272	1	.000	.059
a. Variable(s) entered on step 1: owob_par, edu_par.

Table 26

The odds of the predictor variable (parents university education) is low (0.11), this is an indication that it causes minimal variation in obesity if they are increased.