8.10 – Step 6: Share the Results

6. Share the Results

The final step of hypothesis testing is to share the results with the research community. The Scientific Method is ultimately a social endeavor because it is strengthened by researchers sharing and debating the results. It is through this social process that science helps us get closer to the truth.

For our purposes, we will focus on how the statistical results are typically reported in the research literature, usually in the Results section of original empirical research articles published in peer-reviewed research journals. Through learning how statistical results are written, we are able to learn how to better read and interpret the results.

Reporting our results in the research literature will differ for each of the different test statistics. In this textbook, each chapter that covers a test statistic will provide an example report, as well as some tips about how to write them for different research questions and hypotheses. Thankfully, though, there are some commonalities between the different test statistics.

In this chapter, we used the z-score as the test statistic. Let’s use our example result where our sample of n = 100 individuals ended up with a sample mean of M = 87.85 on the memory test after experiencing sleep deprivation. In Step 4, we determined that the z-score for that sample mean was z_observed = -2.25, and in Step 5, we determined that this z-score is in the critical region. As a result, we would “reject the null hypothesis” and conclude that “sleep deprivation probably impaired memory.” This would then get translated to the following report in the literature:

When we look at how the results get reported, it is worth noting that there is no mention of rejecting or retaining hypotheses. Instead, the results are described in terms of “significance.”

Unfortunately, this phrasing often results in a misinterpretation of the results. When a research study is described as having a “significant” result, it is shorthand for “statistically significant.” Statistical significance simply means that “the result is unlikely to have happened if the null hypothesis is true.” In other words, the research result probably did not happen by chance due to sampling error.

It is extremely important, though, to pay attention to the use of words like “likely” and “probably” in the interpretation of statistical significance. Technically, we are making an inference about the research results using the odds or probability of getting results similar to what the research found.  As a result, a claim of “statistical significance” is ultimately a “guess” about the reality of the world. However, it is really an “educated guess” because it is not just a blind guess but instead uses the information that our statistics can provide us.

Thus, with our above finding, we don’t know for sure that sleep deprivation truly impairs memory. While we found that the participants in our research study had lower memory scores after being sleep deprived, it is still possible that this result happened due to sampling error because we happened to sample a bunch of people who already had worse memories.

While this interpretation is still a possibility, by using inferential statistics, we can be reasonably confident that the results are not due to sampling error because the probability of that possibility is relatively low. That’s what the p-value depicts. The p-value is the probability of obtaining research results that are at least as extreme as the actual observed result, assuming that the null hypothesis is correct. When the statistical results have a p-value that is less than the alpha level (α), the research is considered to be statistically significant.

In our study, the researchers were using an alpha level of α = 0.05. Because the z-score of -2.25 is in the critical region, it has a p-value less than 0.05. In other words, the likelihood of getting a z = -2.25 is less than a 5% chance of happening if the null hypothesis is true. Because that p-value is less than the alpha level, the researchers claim that there is a statistically significant result.

You’ll notice that we reported the p-value in the report in the literature. In our case, it was reported as “p < 0.05.” The “p” means “probability” and is the depiction of the p-value. In our case, the probability of getting the research results that were observed by chance if there is no impact of sleep deprivation on memory (null hypothesis is true) is less than 5%. In other words, it’s not very likely that our results are due to chance because of sampling error.

Do we then know for a fact that sleep deprivation leads to memory impairment? No. There is still a non-zero chance that sleep deprivation has no effect on memory. However, we now have research evidence and data that support the idea that sleep deprivation leads to memory impairment. More research studies will need to be done. If these other research studies continue to find the same type of result, where sleep deprivation leads to memory impairment, then we can be more and more confident that sleep deprivation is bad for memory.