TU-L0022 - Statistical Research Methods D, Lecture, 2.11.2021-6.4.2022

To get started with quantitative research we have to first understand what is quantitative research. Quantitative research is research done with numbers.

When we do research with numbers we have three important questions. The first question is how do we get the numbers. That is about data collection and research design.

Then once we have the numbers. The second question is what do we do with the numbers. We can apply different statistical analysis techniques. We can calculate means, standard deviations, correlations, regression analysis, multi-level modeling, all kind of things.

And to understand what we can do with the numbers, we have to answer the third question. What do we want to say with the numbers? So typically our analysis has some kind of objective. We could be trying to answer a research question about, for example, the relationship between company's profitability and the CEO-gender. Or we could be trying to predict something. For example, given some characteristics of a house, how much are we going to have to pay for the house.

Statistical analysis can be used for multiple different purposes. So when you see a research results such as the regression table here. The three questions here concern what do we want to say with this table. So we have to understand what we can say with the table before we can make any claims. Then what goes to the table with variables. That is about how do we get the numbers and also about what we want to say with the numbers. It's about research design.

And then where the actual numbers in this regression table come from is about, what do we do with the numbers. Typically when we have a final result like this we have already answered these three important questions.Where do we get the numbers, what do we do with the numbers, and what do we want to say with the numbers? And of course these questions are interrelated so you can't really answer one without another.

If you want to do quantitative research you need to have some basic skills and my personal opinion is that you need four different skills. First of all you need to understand the subject of study. So you need to understand what are the most important theories of the phenomena that you're studying. This is important because a big part in quantitative analysis and research design quantitative research is about ruling out alternative explanations. If you don't understand the phenomenon, then it's very difficult to rule out the different causes that could be behind your observed data and it's very difficult then to make any valid causal claims.

The second thing that you need to understand is basics of statistical analysis method. You have to understand what different techniques do and how the results are interpreted. That's important for two reasons. First of all just even if you just knew or how the point-and-click yourself through our software the interpretation part is actually quite challenging. Also the software sometimes doesn't calculate things in a way that you would like it to do and then you need to do some adjustments and to do those adjustments you have to understand what you're doing.

The third thing is research design an application of methods. That refers to skills such as data collection, what data to collect, how do you do, for example, surveys and also what is the standards of methods in your field. Because it varies between fields and you have to know what is expected in terms of quality.

Then the final thing is use of statistical software. You have to be able to use a computer to run your analyses.

Which one of these is the most important one, depends on what you want to specialize in. So no one is a master of all of these four and typically it pays off to be really good in one of these and just have basic understandings of others. For example, if you are really great in theory, and you're good in research design, you are good in collecting data. Then you can always find someone else to help you with the analysis with the statistical methods and use of the statistical software parts. Then you can write the paper together.

Or it also works the other way around. If you are really good at statistical analysis methods and use of the statistical software and perhaps research design but you are weak in theory. Then you can ask a colleague to help you with the theory part. And then you can design the study together. So normally it's a joint effort, where you have people with different kinds of skills. But generally the skill set that you need is of this area and you should understand the basics of each of these four.

The question that I often face is: 'Is quantitative research difficult?' And it can be difficult, at least it's difficult to get started. So if you take, for example, Green's econometrics book and you start studying quantitative research by reading that book. It's nearly impossible to do because you don't have the background knowledge.

So I like this metaphor of quantitative research is as writing a novel. So to write the novel you first have to go to school to learn how to write things. But just a half a year of first grade, which will teach you how to write doesn't make you a great novelist. You're not going to be able to come up with Game of Thrones novel based on just first year degree.

Then qualitative research is like doing a painting and anyone if you give them a pen and paper can doodle something and if it's a child that doodles the parents will then call it art but just being able to get something on a paper doesn't make you a Picasso.It takes years of training.

The reason why people think that quantitative the research is difficult is that they can't write. So you don't have the basic skills of running a correlation, running a regression analysis, then it can be hard. But on the other hand, learning these basics is not that difficult.

Qualitative research, on the other hand, people can think that they know how to do qualitative research if you can do some interviews and write some conclusions about those interviews even if you have no idea about what qualitative research is really about.

In practice, quantitative research is easier to publish for a couple of reasons. First of all, particularly, if you publish to trade magazines, is that people like numbers because we think numbers are more objective than just the interpretations from qualitative data.

The second reason is that there is typically one best practice that can be followed. For example, if you do a survey you apply a factor analysis according to a factor analysis book. Then you apply regression analysis according to the regression analysis book. That's quite a mechanistic exercise and it'll get you published.

On the other hand, in qualitative research you need lots more imagination and your own creativity to get a useful results. So that can be harder. So it's more like an artistic problem and this is more like an engineering problem.