Can an Online Tutoring Program Help in Algebra?

I was browsing the internet last night to find sites that might help my son improve his reading comprehension. One thing I noticed with various sites was that, although it might seem at first a child would be able to navigate easily, there were in fact numerous stumbling blocks. Learning or working on anything online obviously has prerequisites which adults may easily overlook. The effectiveness of an online tutoring program still hinges significantly on guidance. Thus, there is clearly a strong argument in favor of blended learning, one that mixes online learning with classroom instruction. However, the question remains on how much an online program really contributes to learning. This question has been addressed by numerous studies and results are unfortunately murky. One example is a paper recently published in Educational Evaluation and Policy Analysis:

"Improving the median student's performance by approximately eight percentile points" may indeed sound impressive. This size effect is in fact comparable to the amount of mathematical growth typically seen over one full year of instruction in middle school or high school. The software, Cognitive Tutor Algebra I (CTAI), has likewise been reviewed in and screenshots of the program have been provided (An example is shown below): describes CTAI with the following paragraph:
"It is a "tutor": for all problems, the software includes on-demand hints that advance from general hints to more specific ones. It is cognitive: the software adapts the problems to the student's performance, and sometimes provides hints for the student ("just-in-time help").
I especially liked the grapher problems. The student is presented with a problem situation from the real world, such as work hours & pay, time & distance, or number of packages & price. The students need to choose the variables, build an algebraic equation to model the situation, answer questions about specific values of the variables, and make a graph using the Grapher."
One can indeed appreciate the above description with the screenshot provided. The program does provide a lot of tools to the student. And those who understand algebra can easily appreciate this. The question is whether this is really effective for someone who is just learning algebra for the first time. The study by Pane and coworkers unfortunately does not provide a clear answer. The abstract must be read carefully. The effects are absent in the first year of implementation. Actually, the effects were negative in the first cohort of students for both middle school and high school (except for the lowest performing quintiles in middle school):

Above copied from
 Reading the paper closely also brings us to the following section:

Above copied from
The above does show that we may need to reflect a lot more on what blended learning really entails....