Sometimes a few key terms can be the difference between having a grasp of the research in an article or not. These are common terms that are used repeatedly.
Quantitative research: Research that focuses on numerical analysis. The most common method for this type of study is multiple choice surveys, but can also include other numerical information (such as SAT scores or years of college attendance).
Qualitative research: Research focusing on creating a comprehensive picture of a situation or problem through words. Methods for this type of study include interviews, open ended questions, observation, and document analysis.
Research question (or “Purpose of the Study”): The central question that should be answered by the research. A good research article has this labeled clearly, either at the beginning or after reviewing some literature.
n: The number of people responding.
Population: An entire collection of people. (For example, the entire University of Michigan student body)
Sample: A subset of a population. (For example, a random sample of University of Michigan students)
Random: Chosen in a way to allow each member of the population an equal chance of being selected.
Response rate: The proportion of people who respond to a survey. If a researcher send surveys to 100 people, and 89 respond, the response rate is 89%.
Average: There are three different types of average:
- Mean: What we often think of as “average” – we add all the values, and divide by the number of values added.
- Median (or midpoint): The value in the exact center of the list of responses.
- Mode: The most selected value.
For example, if nine graduate students and Madonna are in a room, how we determine the “average” annual earnings can present vastly differing results. Let’s say each graduate student earns $10k, and Madonna earns $10 million. The mean would be $1,009,000 (adding together everyone’s earnings and dividing by 10). The mean makes the graduate students appear to be doing rather well. On the other hand, the median (the value in the middle of the list) is $10,000. The mode is also $10,000, since nine people have that response. An example like this makes clear the need to be careful about which value we choose.
Skew: When one response (or a few) is vastly different from the rest. In the example of the graduate students and Madonna, Madonna’s earnings skewed the mean by moving it far from where the bulk of the responses actually were.
ANOVA: An acronym for “Analysis of variance.” This is a test of the hypothesis that all means are equal. This test helps determine if differences, should they exist, are more than we might expect to happen by chance.
Standard deviation: A description of the variability of responses, using the same measurement as the responses. The larger the standard deviation, the more varied (or potentially skewed) the data.
F: “Frequency ratio” or “Fisher ratio” – The difference in variances between variables. A larger number is better.
Significance: – The probability that the results presented are a result of chance and not an actual difference in variables. A smaller number is better. (For example, 0.01 is better than 0.05. As well, 0.001 is even better than 0.01.) Also known as “probability value” or “p”.
p: “Probability value” – the probability that the results presented are a result of chance and not an actual difference in variables. A smaller number is better. (For example, 0.01 is better than 0.05. As well, 0.001 is even better than 0.01.) Also known as “significance.”
Reliability: Do people respond to survey questions accurately? (Are the questions clear? Is there a tendency for people to deliberately answer a particular question incorrectly?)
Validity: Do the questions on your survey instrument (or interview protocol, or focus group protocol) measure the concepts you think they are measuring?
Coding: A term in qualitative research to refer to the systematic categorization of interview responses, focus group conversations, or observed documents.
By Malinda Matney, Ph.D., Senior Research Associate