Tuesday, August 10 Chapter Ten Educational Research
1. Level of significance: (text pg. 254) probability of being wrong in rejecting the null hypothesis.
Due to the probability being .20, the level of significance proved to be only 20% that the independent variable achieved the result the researchers wanted.
2. Type 1 error: (text pg. 254) rejecting the null hypothesis when it is true.
The researchers attributed the lack of correlation to type 1 error after the researchers found no difference between the group that received a reading intervention and those that didn’t.
3. Confidence interval: (text pg. 255) interval in which the true value of a trait lies.
When comparing the two groups, the data showed only a small overlap thus providing a high confidence interval to the researchers.
4. Parametric: (text pg. 258) statistical procedures based on certain assumptions.
When a researcher used interval-level measures and has a population that is normally distributed, they are looking at results that are parametric statistics.
5. Analysis of variance (ANOVA): (text pg. 259) compares two or more means.
The ANOVA test compares two or more means to find the probability of being wrong in rejecting the null hypothesis.
6. Univariate: (text pg. 264) one dependent variable analyzed.
A univariate study analyzes only one dependent variable.
After comparing the various types of experimental and non-experimental designs from Chapter Nine during class, my group held an interesting discussion regarding our hypothesis. As the group studying the effects of fitness on academic achievement, we felt that our question and hypothesis led us to the simple design of single-group pretest-posttest. Yet after some discussion on our population, we realized this wouldn’t yield great results for a comparative study. We knew that we are selecting essentially “couch potatoes” or students not involved in organized sports. Thus, despite allowing any students to participate in our fitness program, we thought that single-group would be the best design. Yet most research shows that by comparing a control group and an experimental group, researchers can find more relevant and important data. So we switched to nonequivalent-groups pretest-posttest design. But there is more! After discussing this as a class, it turns out that even though our population is selected, our sample can still be random. Thus we finalized the design of randomized-to-groups pretest-posttest. With this we can give ourselves the best chance for a reliable and valid study.