The quantitative article I have analysed was written by two university language educators in Malaysia, Niraihan Mat Daud and Zamnah Husin, titled Developing critical thinking skills in computer-aided extended reading classes. The authors describe an experimental study they conducted to determine whether the use of computer software in an extended reading class can help students develop critical thinking skill. To be specific, the skills are subcategorized into four abilities; apply inductive and deductive reasoning, judge the credibility of assertions, identify assumptions in arguments.

The participants of the study were undergraduate students who took lower intermediate English proficiency. There were 21 students in the experimental group and 19 in the control group. The experimental group analysed a literary text using a computer concordancer whilst the control group analysed the text manually. To ensure the similarity of the treatment to both groups, the researchers used the same curriculum, lesson plans, and teacher. Students’ level of proficiency was also similar based on the placement exam given by the centre to determine their grouping.

At the beginning and end of the class the Cornell Critical Thinking Test Level X was administered to test students’ level of critical thinking. Level X is a multiple-choice test with 71 items. Each item has three response choices and the test usually allows 50 minutes for completion. The results of the pretest and posttest were then compared to see whether there were any significant differences in the variables measured.

Cornbach Alpha Coefficient, pre- and posttest means, and ANOVA were used in the statistical analysis. The findings indicate that “the use of a concordancer was found to enhance students’ ability to think critically.” Although the percentage of contribution of the concordancer to the difference in scores was relatively small, it was still significant. “The percentage is higher in students’ ability to apply deductive reasoning and to judge credibility of assertions” (p. 485).

Internal Validity
The following are a list of the extraneous variables and I will offer my opinion of what variables affected the experimental outcome of the Daud and Husin study.

• Maturation. The variable doesn’t affect the experimental result because the treatment was conducted for only 8 hours (p. 485), and physical or psychological changes in the research participants are not likely to occur in such a relatively short time frame.
• Testing. The authors administered a pre- and posttest (p. 479) but they didn’t say whether the questions in both tests were similar. They only described what Level X looked like (p. 480). I think that the extraneous variable could influence the result because students might show an improvement, simply because of having had the experience of the pretest.
• Instrumentation. The variable doesn’t affect the outcome because the researchers used the same procedure in pretest and posttest.
• Differential selection. I don’t think the variable affects the result. The authors used intact groups, and it was noted that students in the class had a similar level of proficiency, verified by a placement exam given by their university to determine their grouping (p. 481).
• Experimental mortality. In table 3 (pre- and posttest means), it shows the total number of participants (N) of the experimental and control groups (p. 484). I note that no individuals dropped out during the experiment. Therefore, the experimental mortality doesn’t influence the result.
• Experimental treatment diffusion. The variable could affect the result because participants in the control group may have wanted to seek access to the treatment condition. There were four two-hour sessions (p. 482). However, the authors didn’t mention how many sessions in a week were conducted. It is possible that participants in the control groups had had access to the concordancer from participation in a different learning environment.
• Compensatory rivalry by control group. The authors didn’t say whether assignments had been announced to both groups, and whether the control group participants could perform better by perceiving that they were in competition with the experimental group. I think the extraneous variable could affect the result. Table 3 in the article displays that “the control group gained more in the participants’ ability to apply inductive reasoning”(p. 484).
• Compensatory equalization of treatments. Daud and Husin described their experimental design (pp. 478, 479). They didn’t use comparison groups. I don’t think the extraneous variable influences the result.
• Resentful demoralization of the control group. The authors didn’t discuss the possibility that control group participants could become resentful and demoralized, but I think the extraneous variable could affect the result. I notice that in table 3, between the pretest and posttest mean, the control group is shown to have a lower mean of credibility of assertions (p. 484).

External Validity
Two threats to external validity can be described as follows:

• Interaction between setting and treatment. This treatment can’t be generalized from the setting where the experiment occurred in another setting. This research was conducted at a university in Malaysia. The outcome could be different from one carried out at elementary or secondary schools, public or private schools in other cities or countries.
• Interaction of selection and treatment. The authors realized that personological variables, such as the student’s level of proficiency, might affect the generalizability of findings from experiments. The research participants were undergraduate students who were taking lower intermediate English proficiency courses. The outcome might have been different if the research had included participants from a lower or higher English proficiency level.