Print - Back

Depression Scores in Older Adults
Generated Sep 20, 2019 by kelseywhitford


Depression is a diagnosable condition that is classified as a mood disorder. Depression can foster long-term symptoms such as overwhelming sadness, low energy, loss of appetite and lack of interest in things that once brought someone happiness. Depression affects nearly 16.2 million adults in the United States, however depression is less prevalent among older adults than it is younger adults. (Amy Fiske, 2009) Depression in older adults is associated with increased risk of death and disability. Cognitive impairment, increased risk of dementia and increased risk of suicide are all symptoms of depression that are suffered more by older adults than younger adults. (Joanne Rodda, 2011) Depression can be caused by a vast number of things, some of those being seasonal depression, if a person has any comorbidities, or where they live. A long-term study was done on individuals 65 years of age or older to investigate the relationship between geographical location, health status, and depression.


Researchers at Wentworth Medical Center in upstate New York gathered information from 20 healthy individuals in three different geographical locations: Florida, New York and North Carolina. Each person was then given a standardized test to measure depression where higher scores indicate higher levels of depression. Similarly, information was gathered from 20 individuals with one or more comorbidities, arthritis, hypertension, and/or heart ailment, for each of the three geographical locations. Each of these individuals were also given the standardized test to measure depression.


A QQ-Plot for residuals was used to check normality and a scatter plot of residuals vs. fits was used to check constant variability. Nearly all points in the QQ-Plot are generally close to the diagonal line, which indicates no evidence of deviation from normality. The scatter plot shows some grouping between 5 to 8 and 14 to 16, however the vertical variability doesn’t show that any are substantially larger or smaller compared to the rest, so we are comfortable assuming constant variability. <result1><result2>

A two-way ANOVA design with interaction was used to compare the geographical location and comorbidities on depression scores. We find that with a (p-value = 0.262), there is no evidence of an interaction in this case. Analyzing our main effects separately, we see that Location has a (p-value = 0.0516) while Health has a (p-value <0.0001), which shows us that there is not a difference in group means among location, but there is a difference in group means among health. We will look at Tukey comparisons to examine our main effects for differences in more detail. <result3>

Since the p-value for location is close to our significance level of 0.05, I’d still like to add it in our discussion of differences, for the sake of understanding.

With 95% confidence we estimate that average scores for depression are 0.018 to 3.18 higher for individuals living in New York than for individuals living in Florida. However, the average scores for individuals living in New York or North Carolina may be up to 2.71 points different but the study results did not show evidence of differences.

With 95% confidence we estimate that average scores for depression are 6.62 to 8.78 higher for individuals with comorbidities than for individuals that are healthy.


While our p-value showed no evidence of interaction between comorbidities and geographical location on depression scores in older adults, we were able to find differences in location and health status separately based on Tukey pairwise comparisons. Individuals with comorbidities are on average more depressed than those individuals who are healthy, while individuals living in Florida are on average less depressed than those individuals living in New York or North Carolina.


The purpose of this study was to investigate the relationship between geographical location, health status and depression. We’ve found no evidence that depression relates to location. However, we might suspect that individuals living in Florida are less depressed than individuals living elsewhere because of the sunshine and beaches within Florida, but geographical location is ultimately a personal choice that people most likely have the ability to change if they want. We did find that individuals with comorbidities are more likely to have more severe depression than individuals that are healthy, which is reasonable to think that older adults tend to be more depressed when they have one or multiple severe health conditions that could potentially shorten their life span.


Amy Fiske, J. L. (2009). Depression in Older Adults. Annual Review of Clinical Psychology, 363-389.


Joanne Rodda, Z. W. (2011). Depression in Older Adults. The BMJ, 343.

Result 1: QQ Plot   [Info]
Right click to copy

Result 2: Scatter Plot   [Info]
Right click to copy

Result 3: Two Way ANOVA   [Info]

Two Way Analysis of Variance results:

Responses: Score
Row factor: Location
Column factor: Health

ANOVA table


Means table

New York15.25811.625
North Carolina13.957.0510.5

Tukey HSD results (95% level) for Location:

Florida subtracted from
New York1.60.0182111423.18178890.0468
North Carolina0.475-1.10678892.05678890.7563
New York subtracted from
North Carolina-1.125-2.70678890.456788860.2138

Tukey HSD results (95% level) for Health:

CMB subtracted from

Tukey HSD results (95% level) for Location*Health:

Florida,CMB subtracted from
New York,CMB0.75-1.98065453.48065450.9676
New York,Healthy-6.5-9.2306545-3.7693455<0.0001
North Carolina,CMB-0.55-3.28065452.18065450.9919
North Carolina,Healthy-7.45-10.180654-4.7193455<0.0001
Florida,Healthy subtracted from
New York,CMB9.76.969345512.430654<0.0001
New York,Healthy