The research and evidence-based practice use statistics to determine if their prediction about a population is true. The first step of this process is to generate a hypothesis; a hypothesis is an inference made about a population. A hypothesis can either be null or alternative. A null hypothesis suggests that your prediction is caused by chance, but an alternative hypothesis suggests that your prediction is significantly true. Inferential statistics allows us to reach conclusions or tell if a conclusion is true by comparing it to the probability that the conclusion is due to chance. How does this work? Inferential tests are used to calculate a P-value, which is then compared to the probability that the results were caused by chance (Price, 2017). Some examples of inferential tests include chi-square tests, T-tests, ANOVA, Pearson Correlation, bi and multi-variate regression, and a one-sample test.
Inferential tests check for correlation or causation, and an independent T-test and ANOVA measure causation. An independent T-test tests if there is a statistical mean difference between two samples or variables. Analysis of Variance (ANOVA), on the other hand, tests if there is a mean difference between two or more samples or variables. The difference between the two tests is that ANOVA can compare two or more treatments. However, ANOVA is more complicated than an independent t-test (Surbhi, 2017).
The main advantage of inferential tests is that they can make inferences or draw conclusions about a population. The results of inferential tests can be generalized to the whole population when estimated to a specific confidence level. Therefore, a lot of time and effort is saved, which would have been used to test inferences on a population instead of a sample. On the other hand, a disadvantage of inferential tests is that a sample represents a population. Still, it is not the population itself, so we cannot be really sure about the results; so long as the results are to a certain level of confidence, a certain degree of uncertainty still exists. Furthermore, certain assumptions have to be met with inferential tests, and sometimes they are not, and that is a big problem. Lastly, the data is always vulnerable to biasness because the results always require people to make informed conclusions based on the theory they are testing.
From the table on descriptive statistics, we can see an actual difference between the three variables. The study tests for mean life satisfaction score difference between people with no housing problems (MN= 12.71, SD=2.353), people with one housing problem (MN= 11.97, SD=2.588), and people with two or more housing problems (MN= 10.57, SD=2.594). According to the descriptive statistics, people with two or more housing problems have the lowest life satisfaction scores, and those with no housing problems have the highest satisfaction scores. To find out if this difference is significant ANOVA test was done.
Our test seemed to have no problem with Homogeneity of variance (an assumption) since our p-value was more than our alpha value, which means our test was not significant p=0.122. However, the analysis of variance test showed that there was a significant difference in mean life satisfaction scores between our levels of material well-being F (2,934) =61.674, p=0.000. No housing problems seemed to significantly differ in life satisfaction scores with the other two variables (One Housing Problem p=0.001 and Two or More Housing p=0.000). However, its life satisfaction scores were greater than both variables. One housing problem had significantly lower life satisfaction scores than no housing problem p=0.01 but a significantly higher score than two or more housing problems p=0.00. Lastly, two or more housing problems had significantly lower life satisfaction scores than the other two variables (One Housing Problem p= 0.00 and No Housing Problem p= 0.00).
References
Price, J. C. (2017, August 21). 2.6 Analyzing the Data – Research Methods in Psychology. Open Text. https://opentext.wsu.edu/carriecuttler/chapter/analyzing-the-data/
Surbhi, S. (2017, October 11). Difference Between T-test and ANOVA (with Comparison Chart). Key Differences. https://keydifferences.com/difference-between-t-test-and-anova.html
Leave a Reply