Statistical analysis—the backbone of scientific research studies—uses statistical methods to obtain usable information from raw data. It gives meaning to data and the study, determines if the proposed method is more efficient than the existing or conventional methods, and indicates the relationship between variables. Additionally, statistical analysis results can verify the conclusions of a study, thereby affecting the publication status of that study and the researchers’ reputation.
Basic methods of statistical analysis in research
Statistical analysis tends to become quite complex depending on the data collected, information sought, and technique used. Different analysis methods can describe phenomena, estimate relationships, and predict future outcomes, among other tasks. Two types of analysis are commonly used in scientific research studies, and each includes basic statistical tools that all reviewers are and all researchers should become familiar with.
- Descriptive statistics
As the name suggests, descriptive statistics is used to describe datasets. Such an analysis is not intended for creating new knowledge but to understand the data better. The primary statistical tools employed for the descriptive analysis are well known and are described below. Since the calculation methods can be found in any basic statistics resource, they have not been discussed here.
- Mean: The arithmetic average of a group of numerical values. The mean is a measure of central tendency, meaning that it provides a single-point estimation for the value of an entire dataset. In research studies, for example, it can be the average temperature during a specific month or the average reaction time in wastewater treatment. Other measures of central tendency include the median and mode. Researchers must be cautious of outliers in the data as they can significantly affect the mean value.
- Standard deviation: It indicates the distance of data points from the mean. The standard deviation is a measure of variability. A small standard deviation indicates that the data are grouped close to the mean, and a large standard deviation indicates the data are more dispersed. Other measures of central tendency include the variance and box plots.
Data visualizations are often used when describing datasets in scientific research. For example, graphs, charts, and frequency distributions can provide valuable information on the data gathered for a study and can guide further statistical analyses.
- Inferential statistics
Inferential statistical analyses are used to infer knowledge about a population based on the sample data. A population can be an actual population, such as that of a city, or something more nebulous, such as the average monthly temperatures—past and future—of that city or all possible reaction times of an oxidation reaction. Some common statistical tools utilized in inferential statistical analyses include the following.
- Regression: It is used to investigate the association between variables of interest. Simple linear regression is the most frequently used regression analysis method; it produces a linear equation relating two variables. Other regression models, such as multiple regression and non-linear regressions, are also used widely. However, researchers must take great care while conducting a regression analysis as it may offer various pitfalls.
- Hypothesis testing: It determines the differences between groups. The research design specifies the test conditions, often using mean or variance. After data collection, the probability for two groups to be different is calculated, and the results are interpreted.
In inferential statistics, the required assumptions must be carefully checked to ensure that no incorrect conclusions are reached. In addition, because these analyses involve samples, researchers must obtain proper sample sizes.
Tips for conducting a statistical analysis in research studies
- Plan the analysis prior to conducting the experiment. Do not let already collected data dictate the analysis method used; this practice is usually considered to be unethical.
- Use descriptive statistics to investigate your data prior to applying other analysis methods. This will reveal issues that may potentially affect the validity of your results, such as unexpected distributions or outliers.
- Be sure to check the required assumptions for every statistical tool you use. Some will require that the data be normally distributed, others that the data be continuous. If the assumptions are not met, the results will not be valid.
- Avoid overgeneralizing the results. Statistical analyses are based on probability, a science of estimations. It is important to word your conclusions correctly. For example, a hypothesis test indicating a difference between groups does not prove whether the difference exists.
Seek help for the statistical analysis, if required. Editage offers a Statistical Analysis and Review Service that will ensure your research is publication and reviewer ready.