As you can see from the distribution graph above, most students in this class either did really well, or very poorly on the exam. The average score was 69%, which is much lower than I expect, but perhaps not too surprising for the first exam. Especially when the exam addresses many concepts with which most of you are unfamiliar.
Based on an item analysis, there are really only two items missed consistently by most, and I present them below with an explanation of why the correct answers are what they are:
First, remember that variables are anything capable of being changed. In many research designs (especially experimental types), the independent variable is the variable that the researcher manipulates or changes in order to test a hypothesis, and the dependent variables is the effect of the change...its change depends on the independent variable. So in this scenario, the independent variable is the type of manipulative (physical, 3D, none). EXTRANEOUS variable are all those variables in the scenario that could affect the dependent variable, but you don't want them to because they are not the variable being tested. And the best way to minimize their effects is to control for them by either eliminating them or making sure that each group of subjects have an equal chance of having the same (or similar) distributions of variables among the subjects. Certainly "amount of experience using computers" could impact whether or not a 3D manipulative might aid in the learning process, so you would want to make sure this variables is controlled in some way. The same could be said for grade level and amount of time working in the assignments. Even gender is a potential issue because a gender effect has been observed at certain grade levels (though 2nd grade might be too young to be an issue). And you definitely wouldn't want to ask different questions between groups, because if the students performed differently it might simply be the result of different questions.
The other problem many students in this class struggled with is presented below. This does seem a bit tricky since it is Choose all that apply, and the research design is very specific. But when you think about it, a counterbalanced quasi-experimental study in an education study ensures that all groups in the study receive the same multiple interventions (or treatments), but the groups will get these treatments in a different order so that any factors related to the order can be accounted for (and so that all groups get everything, which is an ethical consideration if you are confident the experiences will ultimately be good for the students). But in this design, there is no attempt to control for location of the groups, and the implementations are not controlled either since the groups remain intact throughout the study (and presumably have their own unique parameters associated with how things will be implemented...such as teacher, classroom, etc.).
It can be confusing to keep track of all the different threats to validity that might be addressed by different designs, which is why the chart on Chapter 13 of the text is so valuable:
Hopefully these explanations help you better understand the items you may have struggled with. I am happy to review any others with you individually via email or office hours. I can tell you that all the students who did reach out to me earlier (before the exam) with questions did, in fact, perform better than average on the exam. But as mentioned earlier, you have an opportunity to drop your lowest exam score, so hopefully you will find more success on the next one.
Speaking of which...the material addressed by Exam #2 includes sampling, instrumentation, and related validity/reliability. These are all natural "next steps" in the study of research design because the material from the first part of the course addressed research designs in general...all of which require subjects and some type of instrumentation to collect data. The concepts of validity and reliability speak to the quality of instruments yo might design and/or use to collect research data.