I will do single group pre post test in evaluating training. In that case, what strategy i should follow between experimental or survey research?
If you have an intervention (i.e., training), then you are doing an experiment, but you may well be using a survey as the means for collecting the data. So, this is not a matter of either experiment or survey, but instead it is an experimental design that uses a survey instrument to gather the data.
Also, I disagree with Daniel Wright that doing a pre- post- design without a control group involves a regression discontinuity design. Instead, all it requires is the equivalent of a paired comparisons t-Test between the earlier and later scores on your outcome variable(s).
Do you have a control or comparison group? If you have a control group and have used random allocation, it is an experiment. If random allocation wasn't used, there are more considerations (e.g., https://www.degruyter.com/view/journals/edu/2/1/article-p45.xml). If you don't have a comparison group, then it is a regression discontinuity design, and I assume you have lots of measurements at different times pre and post.
If you have an intervention (i.e., training), then you are doing an experiment, but you may well be using a survey as the means for collecting the data. So, this is not a matter of either experiment or survey, but instead it is an experimental design that uses a survey instrument to gather the data.
Also, I disagree with Daniel Wright that doing a pre- post- design without a control group involves a regression discontinuity design. Instead, all it requires is the equivalent of a paired comparisons t-Test between the earlier and later scores on your outcome variable(s).
Correct David L Morgan , I assumed that if there was no comparison group that he has longitudinal data. Technically of course with just two time points the more complex design would become just the paired test, and depending on a few thing the t-test might be an appropriate choice.
A t test can be used to compare two groups in pairs if the group is large, but if it is small, you can use the Wilkexon Rank Signal Test.
Respected Brother,
Salam from Pakistani Brother, In fact there are only two ways for data collection either interview or observation. Experiment is data creation method. Once you conduct an experiment, there is change which is either observed or asked about. So experiment will always use observation or interview for data collection.
So your options are:
either Experiment along with interview
or Experiment along with observation
- Badges
- Science method
- Similar topics
- Methodology