Social Research Methods/Quantitative Research
Quantitative Research and Data Analysis
After a researcher had conducted experiments and/or surveys, the information he/she is left with is known as quantitative data. This type of information is measurable and focuses on numerical values, unlike qualitative data which is more descriptive. Once the quantitative data is collected, the researcher performs an analysis of the findings.
- Quantitative analysis
- The numerical representation and manipulation of observations for the purpose of describing and explaining the phenomena that those observations reflect
Quantification of Data
Researchers use the process of coding to analyze their findings. When conducting a survey some of the data is numerical while other data must be converted from qualitative to quantitative.
Developing Code Categories
- the process in which raw data is transformed into a standardized form suitable for machine processing and analysis
Coding is the act of assigning numerical values to a set of data in order to make analysis simpler and can be used to quantify both manifest and latent content. The difference between manifest content and latent content is very important when it comes to survey research.
Manifest content is the tangible or concrete surface content (data), as distinguished from latent content, the underlying meaning behind this information.
Advantages with manifest content are ease of testing and reliability and a disadvantage is its validity. Latent content is the underlying meaning of communications, as distinguished from manifest content. An advantage with latent content is that it is designed perfect for tapping the underlying meaning of communication and its disadvantages are its reliability and specificity.
To make the coded data understandable and manageable a codebook is created. This books explains the coding process and acts as a guide for locating variables in the data set. Codebooks also describe the meanings for each code used. There are two purposes for these codebooks, first they are a guide in the coding process. Secondly, codebooks act as guides in locating variables within a study.
Examples of quantitative coding:
- a survey that ranks responses on a 1-9 scale and has the respondent choose one of the nine listed.
- Other variables, such as gender or political affiliation, must be assigned a numerical value in order to conduct quantitative analysis
- i.e.: male= 1; female=2 OR Democrat=1; Republican=2; Independent=3
- Since ages are already represented numerically, the researcher may choose not to develop a coding system for this data.
Coding is necessary in analyzing data because one must be able to transform raw data into meaningful information.
Once data is properly coded it is then ready to be analyzed. One type of analysis is univariate data analysis, in which one variable (such as gender, race, or socioeconomic status) is singled out for analysis to allow for better description.
There are many different ways this data can be analyzed including:
- Frequency Distribution: Counting the number of times data was collected for the sample.
- Average: A term expressing a general trend of data.
- Mean: The sum of total data divided by the number of data points.
- Median: The "middle number" in the data if it is arranged numerically in descending or ascending order.
- Mode : The most frequently occurring data point.
- Mean: The sum of total data divided by the number of data points.
- Dispersion or variance: measures the range of data around a central value, such as the mean
- Standard deviation: measures the dispersion around the mean as well; however in a way that 68% of the sample will lie within plus or minus one standard deviation of the mean.
Technologically advanced programs, such as Microsoft Excel, have the ability to calculate the mean, median, mode, variance, and standard deviation from a set of data. This is a very convenient and easy way to analyze a set of data.
Understanding Distributions and Dispersion
- A normal distribution graph is usually used as diagram to display standard deviation and variance.
If a high proportion of values are tightly clustered around the mean, this represents a low standard deviation; whereas if values are widely spread out among the range of possibilities, this corresponds to a high standard deviation. In simple terms, a low dispersion and standard deviation indicate that the values are rather close to each other with a relatively small amount of variation.
Continuous and Discrete Variables
- Continuous variable a variable whose attributes form a steady progression.
- Discrete variable a variable whose attributes are separate from one another.
Therefore, a continuous variable (such as height) can have an infinite number of possible values, while a discrete variable (such as year) can only have certain values (2010 or 2011 but not 2010.5). Microsoft Excel can also compute these two values easily.
Univariate analysis can be described pictorially with a chart or graph. It is best to maintain simplicity when constructing a chart of graph for better comprehension.
- Bivariate Analysis
- the analysis of two variables simultaneously, for the purpose of determining the empirical relationship between them.
Bivariate analysis focuses on relationships between variables rather than on comparisons of groups. Bivariate analysis explores the statistical association between the independent variable and dependent variable. Its purpose is usually explanatory rather than merely descriptive.
Constructing Contingency Tables
The results of bivariate analyses often are presented in the form of contingency tables, which are constructed to reveal the effects of the independent variable on the dependent variable.
- Contingency Table
- a format for presenting the relationships among variables as percentage distributions
How to construct and read a bivariate table:
- Divide cases into groups according to attributes of the independent variable
- Describe subgroups in terms of attributes of the dependent variable
- Read the table by comparing the independent variable subgroups with another in terms of a given attribute of the dependent variable
Introduction to Multivariate Analysis
- Multivariate Analysis
- the analysis of the simultaneous relationships among several variables. It examines and explains variance in the dependent variable using independent, intervening, and antecedent variables.
- Intervening variables come between the independent variables and the dependent variable in a time order or as a causal mechanism.
- Antecedent variables precede the independent variable.
Multivariate tables can be be created on the basis of a the more sophisticated subgroup descriptions following a similar outline to bivariate tables. This is due to the usage of more than one independent variable and having the dependent variable rely on these independent variables to show the any relationship.
- An example would be attendance at a bingo competition. Assuming prior that older people usually attend this game, we could divide up the participants according to age. Age differences would be our independent variables when broken into various subgroups and will show a relationship to the dependable variable, attendance at bingo.
Dissemination of Information & Other Issues
To display the data, it is usually beneficial to use tables opposed to bar graphs and such due to the complications of portraying more than one independent variable. These tables are also very useful to rule out opinions and find true facts and data.
An issue with quantitative data analysis is the possibility of bias. It is common for researchers to favor one finding over another. To help eliminate this, a detailed hypothesis prior to the research is useful. By recording conclusions that may not prove your hypothesis correct, it is beneficial to other researches (who are working on similar topics) to know your results so it can aid them in their studies. Data can be collected and displayed without any bias when researching controversial topics if the research is conducted correctly and portrayed appropriately in the table.
Some useful research tips to keep in mind when carrying out quantitative analysis:
- It is a good idea to use percentages to make comparisons, and to create these percentages for each category of your dependent variable
- Variables should be recoded in order to make the comparisons you want to make, and this recoding can be done in different ways.
- Choose an independent variable that has enough explanatory power to make sense.