The evaluation should asses/reflect on the METHOD used to collect data. Was the method good enough to lead to a valid conclusion? What where the strengths of the method? This is an opportunity to think in what was well designed and what should be improved if the experiment is done again. Review the variables and EACH of the steps in method and mention ONLY the steps that you believe are good and those that are not good. Is there any indirect measurements?. Can you trust your results measuring indirect variables? Discuss the CV and even more important, in the variables that SHOULD HAVE been kept constant but were not: UNCONTROLLED VARIABLES. The evaluation is NOT the place to describe human mistakes, but to reflect on how the experiment was designed and the data collection conducted. Human mistakes may be presented as weakness only if they affected the lab results directly.
The evaluation can be presented in the form of a data tables, in which the method is carefully evaluated. First the strengths of the method should be outlined and explaining why they are identified as strengths. The all the limitations should be listed explaining the effect on the data and the suggestions of improvement.
The evaluation should present a reflection on the method used to collect data. Was the method good enough to lead to a valid conclusion? This is an opportunity to think in what was well designed and what should be improved if the experiment is done again. Review EACH of the variables and the method used to measure them (indirect measurement?). Focus on the CV and even more important, in the variables that SHOULD HAVE been kept constant but were not: UNCONTROLLED VARIABLES. The evaluation is NOT the place to describe human mistakes, but to reflect on how the lab was designed and the data collection conducted. Human mistakes may be presented as weakness only if they affected the lab results directly.
A measurement of the difference between the results obtained and the theoretical results (accepted or published results). Data is accurate if it follows what is expected, according to previous scientific context.
Is a measurement of validity of data. When identical measurements are done (trials) identical results should be obtained. If the standard deviation of trials is large, the data is NOT reliable: trials should produce the same results but they didn't. If the method is well designed, theoretically, an increase in trials lead to a decrease in standard deviation, and an increase in reliability.
An experiment has sufficient data if the data collected in enough to reach a conclusion. If the experiment had multiple outliers the data collected may not be enough to state a valid conclusion. If there are outliers in the raw data and only 5 trials are collected, the standard deviation can't be correctly calculated.
In English the word significant is often used a synonymous of important. In science, the word significant has a different meaning: it means "having significance". This word is used as an adjective for something that has been statistically tested and that is supported by a statistical analysis. A difference is only SIGNIFICANT if it has been tested with an accepted statistical analysis, such as an ANOVA TEST, a PEARSON TEST, a CHI-SQUARED TEST, a T-STUDENT TEST, or any other accepted test. You can read more about this here.
The appropriate test used to test for significance depends on a number of factors, being the most important the nature of the data: discrete or continuous data?
The coefficient of determination is ofter represented as R-squared.
Is the proportion of the variance in the dependent variable that is predictable from the independent variable(s).
Suggest other experiments that could be done to obtain more information and to gather more evidence to study your RQ. What other experiments can be done to test if your conclusion is correct or not? Suggest other investigations to improve your conclusion and your analysis. One experiment is not enough to state a definite conclusion... what else could be done? Show your PERSONAL ENGAGEMENT and you genuine interest in the topic by suggesting other experiments that you could do to further investigate your RQ.
Students often struggle to evaluate scientific methods beyond simple mistakes. Here is an example of a great scientific evaluation that analyzes the limitations of a method in detail. Thanks to Dr. Amit Khanna for providing this amazing example!
In 1949 Daniel L Arnon described a simple method to estimate chlorophyll in plants. The method was widely accepted and used. In 2017, Raquel Esteban published an extended analysis of the limitations of the method.
Arnon's method described in page 307
Controversy: loads of limitations in the method
Research question is clearly restated.
Research question is clearly answered.
Research question is supported by raw and processed data.
The impact of measurement uncertainty on analysis is addressed (if applicable).