Intro to SPSS
General tips and hints about using SPSS and opening up files into SPSS
https://www.youtube.com/watch?v=ADDR3_Ng5CA
SPSS Tutorial on Frequencies
Descriptive statistics for continuous variables (age, income, BMI)
https://www.youtube.com/watch?v=zI8tE81IeSk&list=UUXLbK1bH-w1oklGm4dLYrHw
SPSS Tutorial on Descriptive Statistics
Descriptive statistics for other variables like dichotomous and categorical variables (education, gender, marital status, race, etc.)
https://www.youtube.com/watch?v=c4mGKguUnvc&list=UUXLbK1bH-w1oklGm4dLYrHw
One-Sample t-Test
Comparing one continuous variable with a population mean (etc. is the study population’s BMI different than the population BMI; is the study populations mean age different than the mean age of the general population, etc.)
https://www.youtube.com/watch?v=jTJdj7ZYmmU&list=UUXLbK1bH-w1oklGm4dLYrHw
Paired-Sample t-Tests
Comparing one continuous variable in either two matched samples (blood pressure of participants on new drug versus placebo, matched on age, gender, and race), or in the same sample at two time points (e.g. SAT scores before and after an SAT prep class; BMI before and after a weight-loss intervention, blood pressure before and after meditation)
https://www.youtube.com/watch?v=eanXmHlW5qE&list=UUXLbK1bH-w1oklGm4dLYrHw
Two Independent Sample t-Tests
Comparing a continuous variable in two different populations, not matched on any variables: mean age of breast cancer diagnosis in Hispanic versus African-American population; # colds in a year in prechool aged children versus school-aged children; height of children on steroids versus those never had steroids
https://www.youtube.com/watch?v=qOH46VVm1Uo&list=UUXLbK1bH-w1oklGm4dLYrHw
Correlations
How do two continuous variables relate to one another? As one variable increases, does the other variable decrease, stay the same, or increase? E.g correlation of age and income is generally relatively high and linear – as one gets older, one increases in income
https://www.youtube.com/watch?v=cNrnSEWKJgg&list=UUXLbK1bH-w1oklGm4dLYrHw
Chi-Square
Comparison of a dichotomous or categorical outcome across two or more independent groups: e.g. are there more women with cervical cancer among Caucasian, African-American, Hispanic, Asian-American, Native American, or other racial groups? Comparison of cancer status across racial groups, or across education groups, or across income levels; comparison of BMI across different job types or levels of employment (salaried, hourly, contractor, commission-based, etc.) or different shifts (day-shift, night-shift, varied).
https://www.youtube.com/watch?v=Ahs8jS5mJKk
Analysis of Variance (ANOVA) and F-Test
Comparison of a continuous outcome across more than two groups: e.g. BMI comparison of low fat, low calories, low carb, and control groups; blood pressure comparison across those on clinical trial drug A, clinical trial drug B, and a control group.
https://www.youtube.com/watch?v=C3-a5jrCjhk
Linear Regression.
Intro to SPSSGeneral tips and hints about using SPSS and opening u.docx
1. Intro to SPSS
General tips and hints about using SPSS and opening up files
into SPSS
https://www.youtube.com/watch?v=ADDR3_Ng5CA
SPSS Tutorial on Frequencies
Descriptive statistics for continuous variables (age, income,
BMI)
https://www.youtube.com/watch?v=zI8tE81IeSk&list=UUXLbK
1bH-w1oklGm4dLYrHw
SPSS Tutorial on Descriptive Statistics
Descriptive statistics for other variables like dichotomous and
categorical variables (education, gender, marital status, race,
etc.)
https://www.youtube.com/watch?v=c4mGKguUnvc&list=UUXL
bK1bH-w1oklGm4dLYrHw
One-Sample t-Test
Comparing one continuous variable with a population mean
(etc. is the study population’s BMI different than the population
BMI; is the study populations mean age different than the mean
age of the general population, etc.)
https://www.youtube.com/watch?v=jTJdj7ZYmmU&list=UUXL
bK1bH-w1oklGm4dLYrHw
Paired-Sample t-Tests
Comparing one continuous variable in either two matched
samples (blood pressure of participants on new drug versus
placebo, matched on age, gender, and race), or in the same
sample at two time points (e.g. SAT scores before and after an
SAT prep class; BMI before and after a weight-loss
intervention, blood pressure before and after meditation)
https://www.youtube.com/watch?v=eanXmHlW5qE&list=UUXL
bK1bH-w1oklGm4dLYrHw
Two Independent Sample t-Tests
Comparing a continuous variable in two different populations,
not matched on any variables: mean age of breast cancer
diagnosis in Hispanic versus African-American population; #
2. colds in a year in prechool aged children versus school-aged
children; height of children on steroids versus those never had
steroids
https://www.youtube.com/watch?v=qOH46VVm1Uo&list=UUX
LbK1bH-w1oklGm4dLYrHw
Correlations
How do two continuous variables relate to one another? As one
variable increases, does the other variable decrease, stay the
same, or increase? E.g correlation of age and income is
generally relatively high and linear – as one gets older, one
increases in income
https://www.youtube.com/watch?v=cNrnSEWKJgg&list=UUXL
bK1bH-w1oklGm4dLYrHw
Chi-Square
Comparison of a dichotomous or categorical outcome across two
or more independent groups: e.g. are there more women with
cervical cancer among Caucasian, African-American, Hispanic,
Asian-American, Native American, or other racial groups?
Comparison of cancer status across racial groups, or across
education groups, or across income levels; comparison of BMI
across different job types or levels of employment (salaried,
hourly, contractor, commission-based, etc.) or different shifts
(day-shift, night-shift, varied).
https://www.youtube.com/watch?v=Ahs8jS5mJKk
Analysis of Variance (ANOVA) and F-Test
Comparison of a continuous outcome across more than two
groups: e.g. BMI comparison of low fat, low calories, low carb,
and control groups; blood pressure comparison across those on
clinical trial drug A, clinical trial drug B, and a control group.
https://www.youtube.com/watch?v=C3-a5jrCjhk
Linear Regression
Continuous variable outcome with any number of risk factors in
a regression model; using risk factors, can we model a
prediction of the continuous variable
E.g. knowing the risk factors of age, gender, race, income,
education, BMI, family history, and other health history, can we
3. predict fasting blood glucose levels?
Fasting blood glucose level here is the continuous outcome.
https://www.youtube.com/watch?v=JVwEdhEiGJg
Logistic Regression
Dichotomous variable outcome with any number of risk factors
in a regression model; using risk factors, can we model a
prediction of the continuous variable
E..g. knowing the risk factors of age, gender, race, income,
education, BMI, family history, and other health history, can we
predict diabetes status?
Diabetes status here is the dichotomous variable (diabetes or no
diabetes)
https://www.youtube.com/watch?v=ODyRncSMMo4
Logistic Regression, Part II
https://www.youtube.com/watch?v=ILEdg0UTTXQ&list=UU3j4
fRCt2VGsttiPs9DXWMQ
Code book for Dataset – MPH 606 Research Methods
Variable
Categories/Label
Coding
Gender
Male
Female
0
1
Race
White
Hispanic
African American
Asian American
Native American
Other
0
1
4. 2
3
4
5
Age
# of years
18 - 91
Salary
Annual Income in Dollars
$10,123 - $117,878
Education
Highschool
College
Masters
Professional
1
2
3
4
Height
inches
60 - 74
BMI
Body Mass Index
18 – 35
Diabetes
No
Yes
0
1
Allergies
No
Yes
0
1
Family History of Diabetes
5. No
Yes
0
1
Family History of Allergies
No
Yes
0
1
Family History of Asthma
No
Yes
0
1
Asthma
No
Yes
0
1
Location
Urban
Suburban
Rural
1
2
3
To feel depressed during the winter
Strongly Disagree
Disagree
Neutral
Agree
Strongly Agree
1
2
3
4
6. 5
To exercise during the summer
Strongly Disagree
Disagree
Neutral
Agree
Strongly Agree
1
2
3
4
5
To overeat when depressed
Strongly Disagree
Disagree
Neutral
Agree
Strongly Agree
1
2
3
4
5
Variable
Categories/Label
Coding
Diet
Low Cal
Low Carb
Low Fat
Vegan
1
2
3
4
7. Hypertensive
No
Yes
0
1
Fasting Blood Glucose
mg/dL
60 - 180
Post Challenge Glucose
mg/dL
100 - 220
Sleep Apnea
No
Yes
0
1
Hours of Sleep
# of hours
5 – 10
Glasses of Water Daily
# of glasses
1 – 9
Servings of Fruit/Veg Weekly
# of servings
5 – 30
Physical Activity
None
Low
Moderate
High
0
1
2
3
Type of Exercise
None
8. Jogging
Swimming
Yoga
Weight Training
0
1
2
3
4
Blood Lead Levels
1 – 14 ug/dL
House Built Before 1978
No
Yes
0
1
Plays with Imported Toys
No
Yes
0
1
PICA
No
Yes
0
1
TB Test
Negative
Positive
0
1
TB Infection
Negative
Acute
Reinfection
9. 0
1
2
Foreign Travel
No
Yes
0
1
Childhood Ear Infection
No
Yes
0
1
Low Vitamin D Levels
No
Yes
0
1
Calcium Levels
Blood Calcium Levels
89 – 101 ug/dL
Bone Mineral Density T-Scores
Negative = less than “normal” bone mineral density levels
0 – -2.5 (all negative)
Variable
Categories/Label
Coding
HPV
No
Yes
0
1
Cervical Cancer
No
10. Yes
0
1
Penile Cancer
No
Yes
0
1
# of Sex Partners
Number of Partners
1 – 4
Unprotected Sex
No
Yes
0
1
Hepatitis B Infection
No
Yes
0
1
Injection Drug Users
No
Yes
0
1
Hamilton Rating Depression Scale
0-7 (normal)
20+ depressive
1 – 30
Iron Deficiency Pre-TB Dx
No
Yes
0
1
Iron Supplement for TB
11. No
Yes
0
1
Blood Transfusion for TB
No
Yes
0
1
Dialysis for TB
No
Yes
0
1
Hemochromatosis
No
Yes
0
1
IQ
IQ Score
75 - 140
Cardiovascular Disease (CVD)
No
Yes
0
1
Dental Carries
No
Yes
0
1
Gingivitis
No
Yes
0
12. 1
Periodontal Disease
No
Yes
0
1
Research Topics:
Diabetes and Obesity
Diet and Exercise and Hypertension
Allergies and Asthma
Depression and Physical Activity/Diet
Sleep Apnea and Physical Activity/Diet
Lead Exposure
Tuberculosis
Osteoporosis and Vitamin D/Calcium Levels
HPV and Cancer
Hepatitis B Infection and Risk Factors
SPSS - Health Study Research Project Topic: Link Between The
Low Carbohydrate Diet and Cardiovascular Disease
**(Use data set for calculations)** **APA Format***
Null Hypothesis: The null hypothesis will be: Adults (age >18)
with self-reported low carbohydrate diets will not have
statistically significant differences in cardio-vascular disease
scores than adults who self-reported other diets. You can use a
single tail test or two tail test, depending on the data chosen.
1. Summarize the 9 studies shown below (or other studies about
the topic) and draft the Background section of your paper. The
background section of a paper sets up the rationale for the
research study. Generally, it communicates three important
things: a description of the problem that will be addressed, a
synthesis of the previous research on this particular topic that
13. highlights the “gap” in the literature, and the research aim. It is
common to see these ideas described in three distinct
paragraphs. The first one describes the public health problem
and generally includes statistics on the prevalence or incidence
of your outcome and potential adverse effects associated with
this topic (e.g., mortality, morbidity, costs). The next paragraph
summarizes or synthesizes the previous work on the topic. The
key here is to justify the research aim by highlighting a
shortcoming or "gap" in the literature. For instance, the "gap"
might be that previous work has not looked at your particular
exposure of interest and its association with the outcome.
Alternatively, it might be that the literature is inconsistent on
your topic and thus more research is needed or it might be that
the data examining this question is too old and needs to be
updated. It might also be that no one has focused on a particular
age group or examined the influence of a third variable (e.g.,
BMI) in the way you think it should be dealt with. These are
just examples; there are lots of ways to justify your research
study, but it requires you to be familiar with previous findings
and sometimes requires a bit of creativity! The last paragraph of
the background describes the research aim and the implications
of this work. Consider why the findings will be important and
how they will be used. Please work on strong scientific writing,
keeping your language objective, without judgement or opinion.
Please also work on using appropriate topic sentences to give
the reader the main point of the paragraph so they know what to
expect in the body of the paragraph. Please use double-spaced
formatting throughout
2. First, identify which variables you will analyze for your
study. Specify your exposure and your outcome of interest and
determine potential confounders of this relationship. Remember
that confounders are those “third variables” that are associated
with both the exposure and the outcome. Then, indicate the
variables’ type – either continuous or categorical.
· Next, in SPSS, run frequency distributions for each of the
categorical variables and descriptivestatistics for each of the
14. continuous variables. Consider also creating a graph or chart in
SPSS to describe these distributions. For variables that have
very small frequencies in at least one category (e.g., less than 7
people in a category), think about combining that category with
another if it makes conceptual sense (i.e., there are very few
Native Americans, does it make sense to combine them with the
Asian group?). See the instructions for recoding variables. Try
the recode, give the recoded variable a new name, run a
frequency of the new variable (it will be on the bottom of the
list of variables), and SAVE THE DATA FILE. IF YOU
RECODE OR CREATE NEW VARIABLES, YOU MUST SAVE
YOUR DATA FILE TO KEEP THE NEW VARIABLES.
· Cut and paste the SPSS output to a word document. Include
the frequency distributions or descriptive statistics for each
variable of interest. If you have recoded a variable, please
include a distribution of just the recoded data (i.e., the new
variable). Then, for each distribution, graph, or statistic
presented, summarize in words what the data means (i.e.,
“Approximately half (54%) the sample is female.”) Type the
summary below the corresponding SPSS result.
3. In SPSS, run the appropriate test of the unadjusted
relationship between your exposure and outcome variables. Use
your recoded or computed variables if created them last week
(i.e. cross-tab tables, chi-squares, correlation, and coefficient)
in order to answer your research question.
Cut and paste your cross-tab tables and test statistics to a word
document. Below each SPSS result, state in words the meaning
of the result (e.g., Females are more likely than males to wash
their hands after every visit to the washroom [Chi-square =
3.84, p=0.05]).
4. Organize the results you will be presenting (no more than 5
tables or graphs). Create new tables or graphs in Word so you
can modify the formatting if needed and add descriptive titles
and labels (refer to Chapter 11 of your textbook). Then, add
your text from the last two weeks in paragraph form.
15. After organizing your findings, consider the following
questions:
How do your results compare to and/or expand the results
reported by studies you reviewed for your introduction?
What are the possible reasons for finding the results that you
did? (Make clear that statements you make that are not backed
up by your results are speculation.)
What do the results suggest for the goals of future studies on
this topic?
What do the results suggest in terms of interventions,
preventions, outreach, policy etc. to improve the health outcome
of interest to your study?
Using the answers to these questions as your guide, draft your
Discussion section. Begin by briefly describing your overall
findings with respect to your research question (i.e., did you
find an association between your exposure and outcome?). Next,
discuss how these findings align or are consistent with the
previous literature. If they are inconsistent, describe the
inconsistency and add possible explanations for it. Next, discuss
the strengths and limitations of your study, with respect to both
internal and external validity. Consider potential biases here.
Last, describe future directions for your readers, which may
include policy recommendations, changes in clinical practice,
and/or the need for additional research. It is also common to
include a final concluding paragraph that summarizes your
findings and the implications of this work.
5. Complete The Draft of the Paper
To do so, first draft the Methods section of your paper.
Generally, the Methods section should include the following
subsections, usually denoted with subheadings:
1. Study overview (this section is often brief - maybe a sentence
or two - that includes the study design)
2. Participants and procedures (here is where you'll need the
detailed information about how the study was carried out - a
paragraph of approximately 4-5 sentences is usually sufficient
3. Measures (here, clearly describe your outcome, exposure, and
16. potential confounders and how each one was measured)
4. Statistical analysis (here, include statements describing your
approach to generating descriptive and inferential statistics, the
statistical analysis package used, and how you treated missing
data)
Please also review the articles you chose for your literature
review table to see examples of how these Methods sections are
written.
Second, finalize your Background, Results, and Discussion
sections. Use the feedback the instructor has provided
throughout the course to revise these sections accordingly.
Then, assemble the draft by including your Background,
Methods, Results, and Discussion in a single document. Please
also include a title page and a structured 250-word abstract on a
separate page at the beginning of the document (prior to the
Background section). Also include the list of your references on
a separate page after your Discussion section and then insert
your tables and/or graphs at the end of the document, each on a
separate page. The entire paper should be in 11- or 12-point
font and double-spaced using APA Format.
9 Studies:
An Epic Debunking of The Saturated Fat Myth.
(2018). Healthline. Retrieved 14 May 2018, from
https://www.healthline.com/nutrition/it-aint-the-fat-people
Brehm, B., Seeley, R., Daniels, S., & D’Alessio, D. (2003). A
Randomized Trial Comparing a Very Low Carbohydrate Diet
and a Calorie-Restricted Low Fat Diet on Body Weight and
Cardiovascular Risk Factors in Healthy Women. The Journal Of
Clinical Endocrinology & Metabolism, 88(4), 1617-1623.
17. doi:10.1210/jc.2002-021480
Lee, T., & Pickard, A. (2013). Exposure Definition and
Measurement. Agency For Healthcare Research And Quality
(US). Retrieved from
https://www.ncbi.nlm.nih.gov/books/NBK126191/
Low-Carbohydrate-Diet Score and the Risk of Coronary Heart
Disease in Women | NEJM. (2018). New England Journal of
Medicine. Retrieved 14 May 2018, from
https://www.nejm.org/doi/full/10.1056/nejmoa055317
Beulens, J., de Bruijne, L., Stolk, R., Peeters, P., Bots, M.,
Grobbee, D., & van der Schouw, Y. (2007). High Dietary
Glycemic Load and Glycemic Index Increase Risk of
Cardiovascular Disease Among Middle-Aged Women. Journal
Of The American College Of Cardiology, 50(1), 14-21.
doi:10.1016/j.jacc.2007.02.068
Liu, S., Willett, W., Stampfer, M., Hu, F., Franz, M., &
Sampson, L. et al. (2000). A prospective study of dietary
glycemic load, carbohydrate intake, and risk of coronary heart
disease in US women. The American Journal Of Clinical
Nutrition, 71(6), 1455-1461. doi:10.1093/ajcn/71.6.1455
Lasker, D., Evans, E., & Layman, D. (2008). Moderate
carbohydrate, moderate protein weight loss diet reduces
cardiovascular disease risk compared to high carbohydrate, low
protein diet in obese adults: A randomized clinical
trial. Nutrition & Metabolism, 5(1), 30. doi:10.1186/1743-7075-
5-30
Keogh, J., Brinkworth, G., Noakes, M., Belobrajdic, D.,
Buckley, J., & Clifton, P. (2008). Effects of weight loss from a
very-low-carbohydrate diet on endothelial function and markers
of cardiovascular disease risk in subjects with abdominal
obesity. The American Journal Of Clinical Nutrition, 87(3),
567-576. doi:10.1093/ajcn/87.3.567
Sacks, F., Carey, V., Anderson, C., Miller, E., Copeland, T., &
Charleston, J. et al. (2014). Effects of High vs Low Glycemic
Index of Dietary Carbohydrate on Cardiovascular Disease Risk
Factors and Insulin Sensitivity. JAMA, 312(23), 2531.
18. doi:10.1001/jama.2014.16658
Foster, G. (2010). Weight and Metabolic Outcomes After 2
Years on a Low-Carbohydrate Versus Low-Fat Diet. Annals Of
Internal Medicine, 153(3), 147. doi:10.7326/0003-4819-153-3-
201008030-00005