Research and Evaluation in Education and Psychology

Mertens, D. M. (2015). Research and evaluation in education and psychology: Integrating diversity with quantitative, qualitative, and mixed methods (4th ed.). Los Angeles, CA; Sage Publications.
Read Chapter 13
Week 6
Print
Statistical Techniques
This week, you will focus on statistical techniques and how statistical techniques are related to research questions.
As you read various journal articles, you will see that researchers use a range of statistical tests. But have you wondered how they decide which test to use? This week, you will work through some of the most common tests. In particular, these include T-tests, Correlations, Regression, and Chi-Square.
But you can begin by examining why statistics are used. When you conduct research, you are always working with a population sample. It is important to ensure any findings you discover are generalizable to the whole population. Therefore, this is the reason statistics are used. Statistics allow you to conclude with a degree of probability that the findings are generalized to the population. For example, if you tested 500 people, 250 men, and 250 women, and asked them to complete a questionnaire that measured their self-esteem, and you found that men had significantly lower self-esteem than women, then you could reasonably generalize that conclusion to the entire population (again, within limits of confidence).
Statistics come in two forms, descriptive and inferential. Descriptive statistics describe the data. These could include means or frequency counts. But just knowing a mean or a frequency does not tell you anything about whether different means or frequency counts are significantly different from one another. For example, you could ask 100 people if they prefer cats or dogs. You might find that 56 people prefer cats and 44 prefer dogs. But is that difference significant? Clearly, 56 is higher than 44; but is that a statistically significant difference?