Interpreting Data

Using data requires some understanding of the nature and quality of the available data. The following outlines some basic data concepts and provides additional resources to help users best utilize the information available through this site.

Using Data with Your Strategic Prevention Framework Model

In keeping with the use of the Strategic Prevention Framework (SPF), we have provided some sample questions for each of the 5 steps of SPF model.

The data you will find on this site is particularly useful for the Assessment and Capacity Building phase as noted below. The SPF model, however, builds on the following:

What intervening variables contribute to the consequences and consumption patterns you wish to change?
What does the local data (county indicators, focus groups, key leader interviews, etc) say about retail availability or laws and policies or social norms in your community? State the issue as a brief problem statement within the logic model.
Capacity Building
Do you have expertise on your task force or coalition to compile, share, and analyze data?
Are there local training and technical assistance needs surrounding the importance, meaningfulness, and utility of data that need to be addressed?
Community Planning
What strategies, related to the contributing factors, can positively impact the intervening variables?
Who are the experts in your community to carry out the strategy (policy, practice, or program) recommendations?
How will you best measure change, and also include them in the development of an evaluation plan?

Tips for Analyzing Data

  • Be sure to analyze the data. Read the definitions, explanations, and footnotes because some of the data have limitations that may lead to inaccurate interpretations.
  • Generally, the fewer the limitations of an indicator, the more you may be confident in their meaningfulness. A way to determine this is to fully understand the source from which the data came. Seek out experts on the information, ask them questions or read their reports.
  • Understanding the difference between percentages and numbers of respondents/cases can lead to less error when interpreting the findings. Percentages are useful in very large sample sizes and/or weighted samples. But number of respondents/cases may be preferred to percentages when, for example, examining a change over a year because it may be only a small increase or decrease. A very small increase could lead to a very large percentage change.
  • After all this analysis you may need to use different sets of data or sources to get a better picture of the issue. Additional datasets may provide better ways to get closer to more meaningful results. You may also need to seek another person or groups perspective that is not involved with the analysis but has background on the issue, to make sure that your proposed resolution is focusing on the right outcome. For example, you provide recommendations, based on your analysis stating that there is are 4.5% of 12th grade students who use methamphetamines in the local high school, to a high school student focus group. Not that this isn’t a problem but, the focus group thinks it is a bigger issue that 64.2% of 12th grade students have used alcohol within the past 30 days and of those students 32.5% of them reported binge drinking within the past 2 weeks. After looking at these numbers it is pretty easy to see that there is a much larger alcohol problem that may need to be addressed first.

Confidence Intervals

What is a confidence interval?
The 95% confidence interval is a commonly used measure of the sampling error of a statistic. It means that in 95% of samples of the size and type used here, the estimated percentage would fall in that range. In popular accounts of surveys and polls, it is often called the ‘‘margin of error’’ of the survey. For example, if a statistic is 20.0% and the confidence interval is 17%–23%, 95% of samples of that size and type would produce estimates between 17% and 23%.
What does a confidence interval tell you?
The confidence interval tells you more than just the possible range around the estimate. It also tells you about how stable the estimate is. A stable estimate is one that would be close to the same value if the survey were repeated. An unstable estimate is one that would vary from one sample to another. Wider confidence intervals in relation to the estimate itself indicate instability. On the other hand, narrow confidence intervals in relation to the point estimate tell you that the estimated value is relatively stable; that repeated surveys would give approximately the same results.
Significance of subgroup differences and of trends between two time points can determined by examining confidence intervals for overlap or lack of overlap. When the confidence interval for a statistic associated with category ‘a’ is overlapped with that for category ‘b’, it is to say that ‘a’ was not significantly different from ‘b.’ In other words, you can identify significant differences themselves by noting nonoverlapping confidence intervals.
Additional Resources
New York State, Department of Health, Confidence Intervals – Statistics Teaching Tool.
Center for Disease Control and Prevention, National Center for Health Statistics. (2010). Teenagers in the United States: Sexual Activity, Contraceptive Use, and Childbearing, National Survey of Family Growth 2006–2008.