For researchers, obtaining statistically significant results is the desired outcome. In a research proposal, researchers write the background and research problem. Futhermore, based on the research problem, the research objectives are formulated.
Based on these research objectives, researchers will create hypotheses for the study to be conducted. These hypotheses must be tested through research activities following scientific principles. Therefore, research methodology and data analysis methods are important aspects for researchers.
Researchers hope that the data analysis will yield estimates that align with the hypothesized outcomes. These estimated results are expected to be consistent with both existing theories and empirical findings from previous research.
What can be done if the statistical tests yield no significant results? Below, I outline six steps to be taken when facing this situation or when significant results do not align with the hypothesized outcomes.
Please review the equation specifications
The first step you need to take is to re-examine the specifications of the equation that has been constructed. Ensure that the equation specifications are grounded in existing theories and empirical findings from previous research.
The message conveyed here is that when developing an equation, it is crucial not to select variables based on the availability of data or data accessibility. Instead, it should be based on the relevant theories and empirical findings from previous research.
If it turns out that the equation specifications are not based on theories and empirical experiences, then you need to perform a re-specification of the equation. If there is an opportunity for re-specification, conduct data analysis again. Hopefully, there will be some changes.
Ensure that the required assumptions are fulfilled
In selecting a statistical test, researchers have made thoughtful considerations. By thoughtful considerations, I mean that all the assumptions required for the chosen test are indeed met.
For instance, if a researcher chooses to use the ordinary least squares regression method, certain assumptions must be fulfilled to obtain consistent and unbiased estimation results.
For example, the researcher needs to ensure that residuals are normally distributed, the residual variance is constant, there is no correlation among independent variables, and the scatter plot of independent and dependent variables forms a linear relationship.
Check the data collection method
The next step that researchers can take when all test results are not statistically significant is to examine the data collection method. It is essential to verify whether the data collection process has been carried out correctly. For example, in survey research using enumerators, have we ensured that proper briefings were conducted before the enumerators went into the field?
If conducting the survey personally, have we made sure that the respondents’ answers align with the actual conditions? In the case of experimental research, have we confirmed that each research stage was conducted according to the prescribed procedures? The data collection method significantly impacts the obtained data. If we have ensured the data collection process was accurate, we will be more confident regardless of the results of the analysis.
Ensure the sampling technique is appropriate
The sampling technique also plays a crucial role, as it determines which samples will be selected for data collection. In theory, if we are sampling from a population, it must be a representative sample.
Representative means that the selected samples accurately represent the observed population. Take a moment to reevaluate the sampling technique and then carefully consider the conditions in the field. Reflect on whether the sampling technique was appropriate or not.
Double-check the data tabulation results
Another important step to take when no significant results are found in the tests is to reexamine the data tabulation results. Whether the tabulation was performed by the researcher personally or in collaboration with a team, it is crucial to review this process.
Data tabulation is of utmost importance, as any errors in data entry can significantly impact the analysis results. Therefore, it is necessary to carefully review the data tabulation that has been conducted. Go through each entry meticulously. Once you are confident that there are no errors in the data tabulation, you can proceed to the final point.
Ensure the data analysis steps are correct
The next step involves verifying that the data analysis steps performed are in line with the established procedures. When conducting data analysis independently, understanding of statistics is essential.
During the data analysis process, from data transformation to interpretation, ensure that each step adheres to statistical principles.
Follow-up steps
So, are you prepared to check points 1 through 6? If you have completed all the checks and identified areas that need improvement, please address them promptly. Once you’ve made the necessary corrections, reanalyze the data. Hopefully, you will achieve better results.
If, after conducting the checks mentioned in the previous paragraphs, everything is in accordance with the procedures, that’s fine. Let’s move on to the final step.
The final step is to gather previous studies, both supporting and contradicting the findings of our research. Then, compare them with the conditions observed in the field.
After conducting a thorough examination, create a justification for why none of the tested variables showed significance. This justification should be based on theories and references from previous research.
Conclusion
If, after analyzing the data, none of the variables show significance, don’t worry, you are not alone. If there are areas that need improvement based on the examination, make the necessary corrections and conduct a reanalysis. If you are confident that everything is in order after the checks, proceed to construct a justification based on theories and empirical findings from previous research.
Well, this is the article I can write on this occasion. I hope it proves beneficial and provides new insights for all of us. Thank you for your attention, and see you in the next article next week.
