The teaching of statistics can be approached in two ways: statistical reasoning and statistical methods. Statistical reasoning means “the way people reason with statistical ideas and make sense of statistical information†(Garfield & Ben-Zvi, 2007). Approached in this way, teaching statistics encourages the students to understand the concepts, link one concept to another, and make correct interpretations about the results of computation. On the other hand, statistical methods refer to the techniques of crunching a large number of data to make them meaningful. Bradstreet (1993) argued that for better learning outcomes, statistical reasoning should be taught first, and statistical methods—usually by means of software for more convenient computation—follow after the students have mastered the statistical concepts. This principle triggers important questions: Do the students know that they should acquire the concepts well before learning to compute numeric data? Do they think they learn statistics better by reasoning the concepts, or are they content by using software to perform computations? To get the answer to this question, the present study stated the research problems as follows: (1) Which did the students prefer—manual or software computation—in learning statistics? (2) What was the underlying reason for their preference?

The design of this study was descriptive qualitative methodology. The participants consisted of twenty undergraduates taking the course “Educational Research Methodology†at the English Department in 2018. In the next semester they would take the course “Quantitative and Qualitative Analysis†where they were going to learn statistics, so this study attempted to find out whether they knew that there was more to statistics than merely computation. These students were taught how to compute central tendency and variability by means of Excel^® for about 20 minutes. Afterwards, in the same session they learned the same materials by listening to a lecture about the concepts of central tendency and variability, then compute the mean, range and standard deviation manually. At the end of the session, they completed an unstructured questionnaire, in which they had to state whether they preferred the first method of learning statistics (machine computation) or the second one (manual computation) when they learned the two topics for the first time. They also had to explain all of the reasons for their preference. The data were analyzed by using this model: familiarizing & categorizing, coding & reducing, interpreting & representing (Ary et al, 2010).

The results showed that most of the students preferred manual computation, indicating they are aware of the importance of statistical reasoning. For them learning statistics was more effective if they computed manually because (1) they know the process of the computation in detail, and (2) they get full grasp of the concepts during the process. Some other students, however, preferred machine computation. It means for these students teaching statistics should be performed by using statistical method, in which the computation became the emphasis of the instruction rather than understanding the concepts. The reasons for this preference were threefold: accuracy, efficiency and convenience. The students claimed that the results of the machine computation were more accurate, eliminating the chance of human errors. They could also save much time because they got the results with only one click on their computer. For them, it was very convenient as they did not need to go through complicated steps of manual computation.

These findings have practical implications for the teaching of statistics. Teachers need to make the students aware of the two approaches—statistical reasoning and statistical methods—in the introductory session of the statistics class. Further, the teachers should explicitly emphasize the importance of statistical reasoning when the students learn new concepts in statistics. After communicating these verbally to the students, the teachers could begin the instructions by applying statistical reasoning in facilitating the students to get full grasp of the concepts, using only manual computation to introduce the concepts. When the mastery of concepts and manual computation has been achieved, the teachers may introduce various types of statistical software to ease computation.