Summary
Highlights
The class begins by outlining the topics to be covered: hypothesis, data collection methods, and sampling methods. A hypothesis is defined as an assumption or tentative explanation for an observed phenomenon, which can be proven or disproven through research. Examples like the relationship between social media usage and academic performance are used to illustrate how some hypotheses are globally proven, while others might be location-specific.
A good hypothesis should be simple, clear, precise, self-explanatory, include variables, and allow for further investigation. The two main types are alternative (positive, denoted as H1) and null (negative, denoted as H0) hypotheses. Alternative hypotheses suggest a relationship or influence (e.g., 'physiotherapy improves performance'), while null hypotheses state no relationship ('physiotherapy does not affect performance').
A simple hypothesis reflects a relationship between one dependent and one independent variable (e.g., 'smoking causes lung cancer'). A complex hypothesis describes the relationship between multiple dependent and independent variables (e.g., 'eating more fruits leads to higher immunity, less cholesterol, and high metabolism').
The instructor guides a student, Da, in developing hypotheses for her research topic: 'The impact of ODL (Online Distance Learning) on student academic performance in Malaysian private universities.' Da identifies six independent variables (student engagement, technological accessibility, self-discipline, academic support, assessment and evaluation, communication and interaction) that are hypothesized to have a positive relationship with academic performance.
Sampling is explained as the process of selecting a subset from a larger group (population) to make inferences about the whole group. Advantages include saving time and resources, while a disadvantage is the potential for error. The choice of sampling method significantly impacts the accuracy of findings. The initial step is to determine the research population size, which then helps in deciding the sample size (e.g., 384 for a population over a million).
In quantitative research, samples are selected to be unbiased and representative, allowing for generalization of results. In contrast, qualitative research focuses on in-depth understanding and does not necessarily aim for generalization, often sampling until data saturation is reached. The predetermined sample size is a characteristic of quantitative research, while qualitative research is more flexible.
Probability sampling ensures every element in the population has an equal and independent chance of selection, minimizing bias and allowing for generalization. Techniques include simple random sampling (like a 'fishbowl draw' or using computer programs where every individual has an equal chance), stratified random sampling (dividing the population into subgroups and sampling from each), and cluster sampling (used for large populations, where entire clusters are randomly selected).
Non-probability sampling is used when the population size is unknown or elements cannot be individually identified. These methods do not rely on random selection and are often guided by specific considerations. Techniques include quota sampling (based on convenience and visible characteristics), accidental sampling (stopping when the required number of respondents is met), judgmental or purposive sampling (selecting experts who can provide the best information), expert sampling (similar to purposive, but specifically targeting known experts), and snowball sampling (using networks to find participants, starting with a few and expanding through referrals).
The strengths of probability sampling include high representativeness, generalizability, reduced bias, and suitability for statistical analysis. Its weaknesses are that it can be time-consuming, costly, and requires a complete population list. Non-probability sampling is faster, cheaper, useful for exploratory research, and effective when no population list is available. However, its weaknesses include a high risk of bias, lack of generalizability, and unsuitability for statistical inference.
The instructor reiterates that the choice of sampling method depends entirely on the research problem statement. A clear problem statement guides the selection of the appropriate research method, sampling technique, and sample size. Changing the research focus (e.g., from ODL impact to ODL challenges) would necessitate a different sampling approach or research methodology.