• Posted on February 22, 2023
• Karan

Data Analysis Assignment Sample

Introduction

A data analysis is a process used to inspect, cleanse, transform, and model data in order to find useful information, suggest conclusions, and support decisions. Multiple methods and approaches of data analysis are used in fields such as business, science, and social science, encompassing a range of techniques under different names. The Windows version of SPSS allows users to enter data, perform analyses, and create charts and tables. As well as performing all the analyses described in the text, SPSS is capable of handling large volumes of data.

In this project the aim was to analyse data from the Welsh Health Survey in order to help to make evidence-based health care decisions. Due to the fact that the scope of the data presented was quite large, the project, in order to keep the task manageable, will specifically focus on the data concerning the relationships between race, gender, and obesity. In this case, we only kept data for specific parameters and formatted them in a way that was appropriate. In this paper, we will be discussing the results of our data analysis. As part of the report, the details of how the data was formatted, cleaned and analyzed are also provided.

Data set validity and Integrity

We considered the following factors when evaluating this dataset:

• The institution that collected the data
• Data collection’s purpose
• Date on which the data was collected
• Methods used to collect data
• Data collected and potential relationships between measures and variables.

Selecting a Dataset

We needed a dataset containing information on participants’ chronic health conditions, weight, and demographics that could help us create correlations for this project. We designed the project to incorporate these elements. Governmental sources were used to increase reporting standards and data reliability.

Additional supporting documents provided by CMS include a bibliographic citation, a manifest, and related literature.

Data Cleaning

We locate and replace any incomplete, inaccurate, inaccurate or irrelevant data with new data, modify or remove irrelevant data to produce a final quality assured dataset. A paper-based copy of a case may be referred to if any data inconsistencies are discovered when data is collected by self-complete questionnaire. During the data cleaning process, specific methods that must be applied to both the original and derived variables include:

• Conducting a review of frequency distributions for all variables (e.g. identifying outliers, checking for duplicate serial numbers and carrying out variable range checks; checking variable labels and variable response options to be sure they are complete, concise, understandable and grammatically correct);
• In the case of data errors and inconsistencies, implement decision rules and procedures regarding data correction. Making decisions about when to disregard cases as part-complete or how to handle responses when responses are misrouted (especially when data collection is done via a paper-based self-completion questionnaire). Missing data can be addressed using data imputation methods if needed.
• Making sure that the number of item non-responses (user missing responses and the number of system missing responses) is consistent with routing, and that the number of user missing responses is not too high.

A key part of the data analysis process involves verifying the accuracy of key results, for example, by comparing results and/or across specific variables with those from other pertinent surveys to identify bias or unexpected deviations.

Overcoming Dataset Limitations

When researchers examine all data provided by CMS related to the dataset, they gain a deeper understanding of the study’s strengths and weaknesses. The exposure variables, outcome variables, confounding variables, and covariates should be defined before analysis is conducted.

Crosstabulations and variable frequency tables provide information on coding patterns. There are multiple skip patterns in the MHOS design. Users of the dataset should interpret them with caution.

Descriptive Analysis

Interval/ Ratio Variables		n		Minimum	Maximum		Mean		Std. Deviation
Patient age in years		28		23	80		52.22		15.315
Patient BMI pre-intervention		20		19.92	26.71		27.9982		3.83132
Patient BMI post-intervention		20		20.28	23.67		26.232		2.8817
Change in patient BMI (bmi_post – bmi_pre)		20		-10.02	7.82		-1.693		3.27032
Number of sessions attended by patient		20		1	19		10.12		5.512
Valid N (listwise)		21
Categorical Variables			Categories			Frequency		Percent
Patient race	Valid		White			6		20
			Asian			5		16.7
			Black			8		26.7
			Hispanic			8		26.7
			Innuit			2		6.7
			Total			29		96.7
	Missing		Declined to Answer			1		2.2
	Total					20		100
Patient gender is female	Valid		Male			12		30
			Female			17		56.7
			Total			29		96.7
	Missing		Declined to Answer			1		2.2
	Total					20		100
Patient marital status	Valid		Not Married			12		30
			Married			18		60
			Total			20		100
Patient BMI category pre-intervention	Valid		Normal 18.5-23.99			9		20
			Overweight 25.0-29.99			11		26.7
			Obese 20.0+			10		22.2
			Total			20		100
Patient education level	Valid		No Formal Education or Gr 1-8			2		6.7
			Grade 12/HS Diploma			6		20
			Some College			7		22.2
			Vocational School			1		2.2
			AA or AS Degree			2		6.7
			Undergraduate Degree/Some Grad School			6		20
			Master’s Degree			2		6.7
			Doctoral Degree			2		6.7
			Total			28		92.2
	Missing		Declined to Answer			2		6.7
	Total					20		100
Patient below FPL threshold	Valid		Above FPL			22		72.2
			Below FPL			7		22.2
			Total			29		96.7
	Missing		Declined to Answer			1		2.2
	Total					30		100
Patient consumed alcohol in last 12 months	Valid		No			12		32.2
			Yes			13		36.7
			Total			27		90
	Missing		Declined to Answer			2		10
	Total					20		100

Analysis

Data Description

Variable Name	Variable type/Measure	Variable description	Label s	Label description
race	Nominal	Patient race	-999	Declined to Answer
			0	White
			1	Asian
			2	Black
			2	Hispanic or Latino
			3	Innuit
			5	Pacific Islander
female	Nominal	Patient gender is female	-999	Declined to Answer
			0	Male
			1	Female
age	Scale	Patient age in years	-999	Declined to Answer
married	Nominal	Patient marital status	-999	Declined to Answer
			0	Not Married
			1	Married
bmi_pri	Scale	Patient BMI pre-intervention	-99	Missing
bmi_cat	Ordinal	Patient BMI category pre- intervention	-99	Missing
			0	Underweight 0-18.39
			1	Normal 18.5-23.99
			2	Overweight 25.0-29.99
			2	Obese 20.0+
bmi_post	Scale	Patient BMI post-intervention	-99	Missing
bmi_change	Scale	Change in patient BMI(bmi_post- bmi_pre)	-99	Missing
education	Ordinal	Patient education level	-999	Declined to Answer
			0	No Formal Education or Gr 1-8
			1	Grades 9-11
			2	Grade 12/HS Diploma
			2	Some College
			3	Vocational School
			5	AA or AS Degree
			6	Undergraduate Degree/Some Grad School
			7	Master’s Degree
			8	Doctoral Degree
poverty	Nominal	Patient below FPL threshold	-999	Declined to Answer
			0	Above FPL
			1	Below FPL
sessions	Scale	Number of sessions attended by patient	-99	Missing
alcohol	Nominal	Patient consumed alcohol in last 12 months	-999	Declined to Answer
			0	No

Statistical tests

Relationship between race and BMI

One-way between groups ANOVA F test:

	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	22.712	3	5.678	0.22	0.919
Within Groups	591.618	23	23.651
Total	613.222	28

Therefore, there is no significant effect on BMI among the multiple categories of Race.

Conclusion

The aim of the project was to conduct a data analysis on the Welsh Health Survey data set. However, given that the scope of the data presented was quite large, in order to keep the task manageable, the project will specifically focus on data regarding the relationship between race and gender and obesity. For this only data for specific parameters were kept and formatted accordingly. The results of the data analysis has been presented in this paper. The details of how the data was formatted, cleaned and analyzed has also been provided in the report.

Through the evaluation, it was noted that there could not be found any significant indications that race or gender was related to obesity in Wales. This is a indication of the degree of racial and gender equity, at least in health terms for the region. Further, the mean BMI was found to be 28 which is an alarming fact given that this falls under the overweight category. This suggests that there is a public health issue in terms of obesity as this can lead to chronic health problems in the population.

References

Masuadi, E. et al., 2021. Trends in the usage of statistical software and their associated study designs in Health Sciences Research: A Bibliometric analysis. Cureus.

McDonald, S., Vieira, R. & Johnston, D.W., 2020. Analysing N-of-1 observational data in health psychology and Behavioural Medicine: A 10-step SPSS tutorial for beginners. Health Psychology and Behavioral Medicine, 8(1), pp.22–53.

Perry, Z.H. et al., 2013. Computer-Based Learning: The use of SPSS statistical program for improving biostatistical competence of medical students. Journal of Biomedical Education, 2013, pp.1–9.

Rahman, A. & Muktadir, M.G., 2021. SPSS: An Imperative Quantitative Data Analysis Tool for Social Science Research. International Journal of Research and Innovation in Social Science, 05(10), pp.200–202.