# Help with college research paper

465

I am enrolled in a statistics class and we were assigned a semester-long quantitative research paper. The study requires the use of a survey as the instrument to obtain data. I chose a correlational design so that cause and effect do not need to be established, but I do need some help on the types of correlation. I was also hoping to gain some insight on what I should consider in creating a survey and any other useful ideas that you may have.

Researchers use correlational studies to determine if a relationship exists between two variables. This means that an increase or decrease in the value of one variable coincides with an increase or a decrease on the other variable. As you pointed out, correlation does not establish if the changes in one variable influenced the changes on the other value. In other words, correlation is not causation, which is a common misconception among students.

When determining correlation, three possible outcomes exist, including positive correlation, negative correlation, and no correlation. A positive correlation exists between variables when an increase in the value of the first variable coincides with an increase in the second variable. An example could be a rise in the number of hours that a student studies correlated in a positive manner with higher test scores. A negative correlation exists when an increase in one variable coincides with a decrease in the other, or vice versa. An example of a negative correlation can be an increase in class absences can correlate with grades in a negative way.  You should keep in mind that this does not mean that class absences cause lower grades. Last, no correlation means that no relationship exists between two variables and a change in a value of one does not coincide with a change in a value of the other. Researchers often use a correlation coefficient that ranges from +1 to -1. Strong positive relationships have values closer to +1 while strong negative relationships have values closer to -1. Values close to zero suggest that the variables do not have a relationship.

A survey is a systematic way to obtain information where researchers ask subjects questions and obtain responses. In research, these surveys would be used to obtain data from a sample population and make a generalization.  The design of the survey is critical to ensure that the instrument obtains the right information and measures what the researcher intends on measuring. In research, this concept is called validity. The survey must translate concrete questions that elicit valid, trustworthy, and quantifiable results to empirical variables that will contain the measurable values.

To obtain measurable results, you need to design your survey so that the questionnaire obtains one or more types of data, which include nominal, ordinal, interval, and ratio. Nominal is a measurement scale that categorizes or labels attributes and such attributes may not be numerical. For example, researchers can classify sex as male, female, or undefined. Ordinal data provides a hierarchy that can generate an order. This type of data can be numerical or non-numerical. For example, small, medium, and large. Neither nominal or ordinal types have measurable distances between the data. Interval data has the characteristics of nominal and ordinal data, but the values must be numerical. Also, this type has a defined measurement and distance between values can be determined. Last, ratio data has all the characteristics of the previous types and has an absolute zero. For example, height, distance, weight, and time are ratio-type data.

As you can see, the type of questions in your questionnaire will determine the type of data that you will collect for your research. However, you need to ensure that what you are collecting and what you are trying to measure represents your study’s intent. Also, when conducting the interview, researchers have multiple strategies such as mail, phone, face to face, or web surveys. While each tactic may have its merits, I recommend using an online service to make a survey, because these not only increase the responses, but the websites themselves help guide the creation of the questions and the necessary formatting.