Reading one research article can be a relaxing experience. You are gaining knowledge and learning ways to implement the research outcomes in your occupational therapy practice. But, when you have multiple articles to read or if you had a long day at work, the thought of reading a 20-page research study is not particularly appealing. However, the information in that article has to be there to ensure the scientific method is strong to allow for generalization to implement in one’s practice. So, how do you efficiently (and effectively) read an article to gleam the important information to complete a paper or to implement into your practice?
The most tempting response is to read the abstract. The abstract provides small portions of all the important components: background, methods, results, and conclusion. It is expected that one would feel the implementation of the results can occur once you have a brief description of the research study. While the abstract is a good start, understanding all the components of that particular study is important to comprehend the research study and whether you can implement the results in practice or if the research presents strong data for your own research.
When reading an article, knowing what parts of the article to focus on can help you understand the research for implementation, despite a lack of time.
- Purpose. Start with the purpose of the study to understand what the researcher wants to accomplish with this study. Knowing the purpose, will guide you with the remaining nuts and bolts of the article: research questions, objectives, and/or hypotheses. Once you understand the purpose and then the research questions of that particular research study, you will have a clearer understanding of what the researcher hopes to specifically address in this study.
The purpose also provides the variables that are being analyzed in the research study. This information can also set the stage for determining what possible statistical methods that will be used to answer the research questions (for example, comparing two variable could result in a correlation, comparing two groups may result in a t-test, etc.). While the literature review sets up the research article very well and gives you context for the study, the literature review is really for the researcher to set up the purpose of the study. So, scan the literature review and focus on the purpose, research questions, and objectives.
- Sample. Next understand the sample. Understanding the sample sets you up to interpret the generalizability of the results. You may read in the abstract that this study was statistically significant, and you think, great! I got a study I can implement in my practice! However, you may learn that the sample size is heavily one race, gender, disability, etc., which limits generalizability. When looking at sample size, understand these key components:
- How the sample was recruited (random sample or convenience sample)?
- What were the inclusion and exclusion criteria?
- Number in the sample.
- Return rate after recruitment; attrition rate and why was those participants not included in the data results.
- Measures. Once you visualize the makeup of the sample, recognize the measures that were used to collect data from that sample. Understanding the measures used will also set you up for understanding the statistical data. The assessment types provide context for the numbers. Some measures that are used to collect data: surveys, standardized assessments, or open-ended questions. The point of these measures is to objectively quantify the performance of the sample to reach a conclusion to answer the research questions posed in the beginning of the study or to address the purpose of the study.
Measures can also be characteristics of the participants if the research questions are looking at attributes of the sample (example: income) to determine a correlation between two variables. For example, income and education performance. The measures would be the income level of the participants and grades to see the correlation between the two variables.
- Results. When you understand the numbers that come from the measures, you will comprehend the results of the study with ease. Researchers will most likely provide two types of statistics: (1) Descriptive and (2) Inferential. Some studies will provide one or the other. However, many researchers like to prep you for the inferential statistics with the ‘softer’ descriptive statistics.
Descriptive statistics are the basic statistical information of the sample. The most common ones you will see in a research study are mean, median, and standard deviation. Researchers will then provide the inferential statistical data to provide more refined information on the differences between the groups or the impact of the intervention on participants in the sample.
Inferential statistics can be test statistics such as t-test, correlation, or ANOVA. Overall, these numbers are the data that allows you to understand the differences between the groups that are being compared or the impact of the intervention.
There are key factors when reviewing the results:
a. Standard deviations: Standard deviations are probably the most underrated statistic out there. If you think about it in terms of exams, if you have an exam taken by 100 students with a mean of 70 and a standard deviation of 20, that gives you a lot of information. This lets you know that there was a lot of variability between the highest and lowest score. This means that some got 90 and others got 50 and everything in between. Taking the same test taken by 100 students, a mean of 70 and a standard deviation of 5, lets you know that there was less variability in the scores. So, when you look at the standard deviation, taking into account the sample size, you will get a starting picture of the score variability and an idea of lowest and highest scores.
You can apply this concept to the standard deviation provided as part of the preliminary results of a study. Look for the average and the standard deviation to determine the level of variability between the groups on a measure. Greater variability sets you up to realize that there is a difference between the groups and the inferential statistics may show a statistically significant difference between the groups or an intervention. Small variability will most likely result in less likelihood of a statistically significant difference between the groups or an intervention.
The issue with the standard deviation, is it provides general information of the variability of the scores in the total sample. So, when comparing sample or group scores, it does not give you a true picture of the performance of the participants in a sample or differences between groups. Inferential statistics provides more sound data on whether or not there are statistically significant differences in the sample post intervention or between the groups.
Most studies like to provide some descriptive statistics to setup the reader for some general information. Review the descriptive data to get a feel for the results. Then move on the inferential statistics to determine if there is really a statistically significance between the groups.
b. Determine your test statistics. Reading the results part is always more nerve racking than it actually is. Once you understand the test statistic that is being used, you can contextualize what is going on and figure out the strength of the results. With each statistical result, the researchers will walk you through the meaning of the test statistic and whether or not the outcome shows a difference between the groups or with the intervention. The most popular inferential statistical measures:
i. Pearson Correlation
vi. Chi Square Statistic
vii. Bi-variate Regression
viii. Multi-variate Regression
While these test statistics appear intimidating, at the end of the day, the focus is whether or not the researchers found that the results using the test-statistic were significant or not. If it was not found to be significant, the researchers will always provide the silver lining in the results to help you understand what impacted the outcome. Therefore, from a clinical perspective you will learn what affected the efficacy of the intervention and as a researcher you will learn how to make a better study.
c. Organization of statistical results. Each study organizes information statistical information differently. The most popular method is to return to the research questions, state the measures that were used to answer the research question, and present the statistical differences in the outcomes of the sample with the measures used in the study.
d. Qualitative data results. With qualitative studies, themes are indicated to reach a conclusion on the research questions asked. The researchers will provide the process of organizing the data into themes and the conclusions that were determined from the themes.
- Discussion. Each study does the discussion section differently. Making a research study is hard work! So, the researcher will present the results and provide some input to support the results. For example: The correlation was weak, but the sample size was small. Another example: no statistically significant differences were found between the groups, but the demographics of all the groups was heterogenous in one area, which may have affected the outcome.
Ultimately, you can create your own discussion in your head as you read the study and you scan the discussion section. Scanning allows you to see if there are points that the researchers saw that you did not. Note: some people say read the discussion first. I disagree. Reading the discussion clouds your understanding of the article because, as I mentioned, the researcher will provide support for the results. The interpretation of the article will not be your own, which will affect your understanding of the research, which will limit your ability to determine how you can implement the measure in your practice.
- Limitations. Along the lines of the discussion, the limitations can also be determined prior to reading what the researchers concluded from reading the article as I stated above. This is also an area that can be scanned.
- Takeaways. At the end of the article, always think of takeaways for implementation. Some research articles provide this information. But, honestly, your practice area is not the research setting. Implementation in your practice should be part of your thought process as you learn about the sample and the measures that were analyzed.
For my next post, I will provide tips for implementing research into your practice. The advice provided will help you hone your skills to be able to focus on the takeaways from research outcomes.
Thanks for reading!