Background and objectives
A retirement home has undergone extensive environmental renovations focusing on fall prevention. The next phase involves helping clients becoming more active to address balance and coordination, strength and aerobic capacity. The mobilityintervention_fall16.xlsx dataset represents data from two time points (baseline and 90 days later) of a fictitious study that explores the 90 day impact of two different interventions designed to help reduce clients’ think they have to totally rely on staff to help them get around. Clients completed several scales and then were randomly assigned to one of two exercise groups: strength exercises only or strength plus flexibility and balance. Higher scores on the fear of falling and anxiety scales implies more fear and anxiety symptoms. Higher scores on the confidence scale implies greater confidence in being able to walk unassisted.
The intent of this analysis is to have you use these data to demonstrate your ability to:
1) Identify the appropriate statistical tests to test differences in means;
2) Conduct Excel analyses to test for statistical differences; and
3) Summarize findings as you would in a journal article or report.
Assignment
Part 1: Describe the sample characteristics and baseline values, comparing the two groups’ characteristics.
With any analysis, the first step is to examine your data, assessing the degree of missing data, the potential miscodes, and conducting a descriptive analysis. Then, a table is created to describe the sample characteristics. Since this is an intervention study with two groups, the characteristics of the 2 groups should be described separately, rather than report on the entire sample. The example table shell below gives you an illustration of how this might look. Using Excel, calculate the descriptive statistics for sex, age, the baseline depression, fear of falling and confidence scores. Then describe the findings in the table.
Table 1
Characteristics of clients at baseline (n=)
Strength only
n= Strength plus
n=
Mean SD Mean SD
Age (years)
Baseline Values
Fear of falling
Confidence to walk unassisted
Anxiety
Gender Frequency % Frequency %
Female
Male
Note. Data source: mobilityintervention_fall16.xlsx.
Part 2: Comparing two independent group means
In an intervention it is always good to examine if the groups are comparable at the start, so we want to find statistical evidence to reject the following hypothesis:
The mean score for confidence to walk unassisted at time 1 will be different for the two intervention groups.
Follow the steps of hypothesis testing, stating the null hypothesis, review assumptions, and consider if the hypothesis is directional or not. Consider if the distribution of the confidence score is approximately normal. You do not need to statistically test for homogeneity of variance in Excel but you should consider (calculate and ‘eye ball compare’) if the standard deviations of the groups being compared are similar (hint: know how SD relates to variation) so you can select the proper t-test condition to run. Write a few sentences describing your steps and the results of the testing, making sure that you note what group means are being compared and what the results would mean to someone who does not understand statistics. This video may be useful http://www.youtube.com/watch?v=X14z9r8FUKY
See the end of the assignment for examples of how to write up the results.
Part 3: Comparing means of a single group, pretest-posttest
After 90 days do we see early results that overall regardless of any group assignment that our interventions appear to be helping (i.e, reducing the fear of falling). For the entire sample (regardless of the intervention), test the hypothesis:
The fear of falling score will change over the 90 day period (i.e., compare Time 1 to Time 2).
Again, follow the steps of hypothesis testing, stating the null hypothesis, review assumptions, and consider if the hypothesis is directional or not. Make sure that you evaluate the assumptions for the statistical test, particularly the level of the measurement and the normality of the distribution. Write a few sentences describing your steps and the results of the testing, making sure that you note what group means are being compared and what the results would mean to someone who does not understand statistics. See the end of the assignment for examples of how to write up the results.
Part 4: Comparing means of more than 2 groups
Feelings of anxiety may influence your success to reduce the fear of falling. A variable has been created for you reflecting low, medium and high anxiety at baseline. Test the following hypothesis:
The fear of falling score at 90 days is statistically different for the three levels of baseline anxiety groups (low, medium, and high).
As with the t-tests, assumptions of the statistical tests must be considered. You do not need to test for homogeneity of variance in Excel but you should consider if the standard deviations of the 3 groups being compared are similar. Write a few sentences describing your steps and the results of the testing, making sure that you note what group means are being compared and what the results would mean to someone who does not understand statistics. If the ANOVA F test is significant, further post hoc testing would be needed to compare which means are statistically different (e.g., Group 1 vs 2? Group 1 vs 3? Group 2 vs. 3?). This video shows how to do this but it is not required for this assignment.
This video may be helpful http://www.youtube.com/watch?v=tPGPV_XPw-o
See the end of the assignment for examples of how to write up the results.
Examples of how results may be written for …..
Independent samples t-test (from analysis of another database):
An independent samples t-test was conducted to compare self-esteem scores for males and female high school students. There was no significant difference in scores for males (M=34.1, SD=4.9) and females (M=33.2, SD=5.7); t=1.62, p=.11, two tailed.
Paired samples t-test (from analysis of another database):
A paired samples t-test was conducted to evaluate the impact of the weight loss intervention on women’s confidence in ability to lose weight. There was a statistically significant increase in confidence from baseline (M=55.2, SD=5.8) to three months after the intervention (M=68.5, SD= 5.0); t=5.39, P<.0005, one-tailed.
One-way ANOVA (from analysis of another database):
A one-way between-groups ANOVA was conducted to explore the impact of education (high school or less, some college, college degree) on levels of optimism, as measured by the Life Orientation Test. There was a statistically significant difference in at least two of the group means (F=4.6, p=.01). The mean scores of the “high school or less groups” was the lowest (M=21.4, SD=4.6); the means of the “some college group” was higher (M=23.1, SD=4.5). The means of the “college degree” were the highest of the three groups (M=23.4, SD=4.0). Further post hoc analyses would be needed to determine which groups were significantly different from each other.
