DEPRESSION AND GEOGRAPHY
AN ANALYSIS OF FACTORS INFLUENCING
DEPRESSION AMONG THE ELDERLY
Samantha Boettcher speery@[Link]
Derrick Espadas daespadas@[Link]
December 13, 2016
2
Table of Contents
Executive Summary………………………...……... 3
Data Analysis…….……………………….............. 3
Presentation of Data………………………………... 5
Concluding Statement………………………............... 8
Reference List…………………………………….…… 9
3
Executive Summary
The illness of depression prevails. Most discussion of the effected revolves around the younger
population segments. However, the elderly comprises 20% of sufferers (Cole, Bellavance, &
Mansour, 1999). In addition to genetic factors that leave the elderly predisposed to depression,
this population faces contemplation of mortality, deterioration of health, and separation from
family as familial dynamics alter. Geriatric psychologists recognize the contributions of these
factors leading to a prevalence of undertreated depression (Cole, et. al, 1999). To better serve
this community, physicians strive to identify the specific generating factors of depressive
symptoms.
As part of a long-term study of individuals 65 years of age or older, sociologists and physicians
at the Wentworth Medical Center in upstate New York investigated the relationship between
geographic location and depression. A sample of 60 individuals, all in reasonably good health,
was selected; 20 individuals were residents of Florida, 20 were residents of New York, and 20
were residents of North Carolina. Each of the individuals sampled were given a standardized test
to measure depression.
A second part of the study considered the relationship between geographic location and
depression for individuals 65 years of age or older who had a chronic health condition such as
arthritis, hypertension, and/or heart ailment. A sample of 60 individuals with such conditions was
identified. Again, 20 were residents of Florida, 20 were residents of New York, and 20 were
residents of North Carolina.
This report will illustrate factors with relevance to initiation of depressive symptoms in elderly
populations. It will denote the connectivity of depression regarding geography and chronic
health conditions.
Data Analysis
Assumption of Data
Results for the standardized tests were compiled (Figure 1). Higher scores indicate a higher
level of depression. An initial comparison of test results reveal those currently in good health
received lower depression scores than those ailing with chronic health conditions. Direct
comparison of factors (geography) in group one versus the scores in group two construct the
analysis. Instinctually, the eyes want to compare Treatment One-Cell One with Treatment Two-
Cell One. However, to ensure proper correlation, utilization of a more sophisticated analysis tool
becomes essential.
4
Data Set: Medical One Data Set: Medical Two
New New
Florida North Carolina Florida North Carolina
York York
3 8 10 13 14 10
7 11 7 12 9 12
7 9 3 17 15 15
3 7 5 17 12 18
8 8 11 20 16 12
8 7 8 21 24 14
8 8 4 16 18 17
5 4 3 14 14 8
5 13 7 13 15 14
2 10 8 17 17 16
6 6 8 12 20 18
2 8 7 9 11 17
6 12 3 12 23 19
6 8 9 15 19 15
9 6 8 16 17 13
7 8 12 15 14 14
5 5 6 13 9 11
4 7 3 10 14 12
7 7 8 11 13 13
3 8 11 17 11 11
Figure 1. Initial test scores presented for all participants. Geographic location and presence of chronic illness classify scores.
ANOVA Testing
The analysis of variance (ANOVA) compares the ratios of multiple variances. An ANOVA
allows for the comparison of variances beyond two populations. It also allows for the
quantitative illustration of interaction between variances. The ANOVA test will allow us to
compare the means of the scores in the geographic locations and see if a difference exists among
them.
For example, juxtaposition of an individual tmean against the combined mean will illustrate if an
individual mean comes from a common population. Calculation for each treatment’s distribution
5
is made. All the data points from each distribution join to illustrate a larger distribution. From
this point an analysis of relative distance between treatment mean and overall mean manifests
itself. How far does a treatment mean deviate from the whole mean?
The goal moves to proving a null hypothesis that all the means equal each other and the overall
mean: Null hypothesis Ho: µ1 = µ2 = µ 3. Hence, they do not deviate. They likely arise from
the same overall population.
Preference for performing this method instead of a t-test exists. A t-test increases the probability
of Type I error, or the incorrect rejection of a true null hypothesis. Although each statistic may
only bear a 5% chance of error, progressive tests compound the probability of error. This error
type is known as an experimentwise Type I error rate and experiments as those conducted by
Wentworth Medical Center may increase the probability of error upwards of 15%.
Ultimately, an ANOVA analysis will look at the ratio (F statistic) of the variability between the
means in comparison to the overall mean. The variance of the individual mean is known as it’s
spread. This ratio yields the total variance of components while adding the two components
yields the total variance. This method illustrates partitioning.
As a rule, if the numerator is larger than the denominator then reject the null hypothesis. This
signifies that one mean does not come from the same population and each distribution presents
as narrow: Each distribution is distinct. If each component of the ratio relates similarly to each
other, then we fail to reject the null hypothesis. When means calculate in fair proximity of each
other or overlap, they present as indistinct and it can be inferred they originate from the same
population. When the numerator is smaller than the denominator, a similar result occurs.
Presentation of Data
The following data presents in accordance with the method previously described and in reference
to the initial data set (Figure 1).
ANOVA Test: Healthy Group
Data Set: Medical One
ANOVA: Single Factor
SUMMARY
Groups Count Sum Average Variance
Florida 20 11 5.55 4.57632
New York 20 160 8 4.84211
North Carolina 20 141 7.05 8.05
6
ANOVA
Source of Variation SS df MS F P-Value F crit
BETWEEN Groups 61.0333 2 30.5167 5.24 0.0081 3.16
WITHIN Groups 331.9 57 5.82281
Total 392.933 59
Data Set One (see figure 1)
Let µi= mean value of depression score
Null hypothesis Ho: µ1 = µ2 = µ 3
Alternative hypothesis H1: At least one µi is different
Where µ1, µ2, and µ 3 are the means of
Florida, New York and North Carolina
The level of significance is = 0.05
The P-value of 0.0081 is less than the significance level of 0.05.
There is a significant difference in the mean depression score of healthy people in the
three locations (at least one µi is different).
Therefore, we reject the null hypothesis and we can conclude that the means are not equal.
7
ANOVA Test: Group with Chronic Illness
Data Set: Medical Two
SUMMARY
Groups Count Sum Average Variance
Florida 20 290 14.5 10.0526
New York 20 305 15.25 17.0395
North Carolina 20 279 13.95 8.68158
ANOVA
Source of Variation SS df MS F P-Value F crit
BETWEEN Groups 17.0333 2 8.51667 0.71 0.4939 3.16
WITHIN Groups 679.733 57 11.9246
Total 696.733 59
Data Set Two (see figure 1)
Let µi= mean value of depression score
Null hypothesis Ho: µ1 = µ2 = µ 3
Alternative hypothesis H1: At least one µi is different
Where µ1, µ2, and µ 3 are the means of Florida, New York and North Carolina
The level of significance is = 0.05
The P-value of 0.4939 is greater than the significance level of 0.05. There is no significant
difference in the mean depression score of healthy people in the three locations.
Therefore, we do not reject the null hypothesis and we can conclude that the means are equal.
8
Concluding Statement
Initial reading of the primary data reveals individuals with chronic illness to manifest greater
levels of depression. The ANOVA test for those without a diagnosis of chronic illness present
rejected the null hypothesis that the means bear equivalence. This signifies location bears little
significance in the depressive levels of individuals with an absence of chronic illness. However,
in the data for individuals with chronic illness, calculations failed to reject the null hypothesis
indicating a strong influence of geography upon an individual’s depressive levels when chronic
illness is present.
The ANOVA testing provides a sophisticated and clear picture of the effects of depressive
factors in the elderly population. This data will assist physicians in increasing efficacy in the
treatment of depression, specifically in relation to geographic factors. If you have further
questions or concerns please contact Derrick Espadas or Samantha Boetcher at
Daespadas@[Link] or SPeery@[Link] respectively.
9
Reference List
Cole, Martin, Bellavance, Francois, Mansour, Asmaa. (1999). Prognosis of Depression in
Elderly Community and Primary Care Populations: A Systematic Review and Meta-
Analysis. American Journal of Psychiatry. Retrieved from
[Link]