AMDM

The speaker talks about different modes of sampling and their advantages and disadvantages. They then shift focus to biases that can occur in statistics, specifically confirmation bias and sampling bias. Confirmation bias is when preconceived notions influence data analysis and collection. Sampling bias is when certain groups are disproportionately represented in a sample. The speaker highlights the need to consider biases when choosing a sampling model and emphasizes the impact biases can have on research. The audience is asked to choose one bias and provide a scenario illustrating its effects. Hey, everyone. Welcome back. Last week, we talked about different modes of sampling, such as stratified sampling. We talked about cluster sampling, convenience sampling, and also volunteer sampling. And we also talked about the advantages and disadvantages of each of those different models. What today I want to focus on is different types of bias that can emerge when we're doing statistics. The first major bias that can come up is confirmation bias. Confirmation bias in an experimental study is when you have a preconceived notion of what the outcome of the study should or should not be, and you let that preconceived notion dictate the way that you analyze data, the way that you might collect that data, the way that you might interview people or create your questionnaires, things like that. Confirmation bias can affect all five parts of the research cycle. And so can another type of bias, which is sampling bias. You have to keep in mind when you're sampling different groups of people using whichever model you prefer, you have to weigh the advantages and disadvantages of each one of those and which one is going to be most appropriate in your context. And you may also even want to consider paying participants versus just asking for free volunteers. Things like that will also affect your data. So keeping in mind how different sampling models can bias your data, and also if you have any biases going into the study, how those can affect your data. So what I want you guys to do now is, given those two different types of biases, is pick one, either confirmation bias or sampling bias, and I want you to create a scenario that illustrates one of those and explain why that's the case, where that bias is emerging, and why and how it would affect our research.