Statistical Analysis

Experienced data analysts know that statistical analysis and a meaningful report often require more work in acquiring, merging, editing, and coding data than in carrying out the analysis or report itself. Before the, transformations may be required (such as weighting sample data) to adjust the distribution of cases to more accurately reflect the larger population. Missing values can affect results. They are one of the most pervasive problems in data analysis. Research Design Associates conducts the appropriate assessments and adjustments for the pattern and amount of missing data. As a result the client can have confidence in results of the analyses.
Contact Research Design Associates to learn more about Statistical Analysis.
A crucial step in any data analysis understanding the statistical description of the data set. The descriptive analysis of each variable in the study which includes results such as mean, standard deviation, range and simple frequency tables. The frequency tables or bar charts help to bring order out of what appears to be chaos.

Example: Univariate & Bivariate Analysis

For test scores, for example, score values are ordered (let’s say 0 through 10) and the number of times or frequency that each score occurs is recorded. With a bar chart, the size of the bar indicates how many times, for instance, a score of 0 occurred. These univariate analyses provide a straightforward way to summarize and describe large amounts of data, and an additional method for uncovering data errors.

Univariate analysis contrasts with bivariate analysis (the analysis of two variables simultaneously, or multivariate analysis, the analysis of multiple variables simultaneously). Research Design Associates can help you explore your data, select the most appropriate analysis and execute the statistics necessary. Our experts understand the rigor of the science and the need to communicate complex results accurately and lucidly.

Example: Multivariate Research

A researcher might want to know if there is a relationship between SAT scores and college grade point averages and a bivariate correlation could be used to test that research question. In marketing research, market segmentation is a strategy that divides a larger market into subsets of consumers who have common needs and applications for goods and services. These subgroups of consumers can be identified by a number of different demographics, for example by age and income, and cross-tabular analysis of age group by income group can be conducted. Market segmentation strategies can be developed over a wide range of characteristics found among consumers and it takes experienced market researchers to uncover what demographic and psychographic criteria will best help a company plan a marketing campaign.

Multivariate analysis is used for analyzing complicated data sets.  Content analysis is a system for studying the meaning of the content of verbal communication. With the use of recently devised text analytic software for content analysis, unstructured survey data can be transformed into quantitative data. With such sophisticated tools, Research Design Associates’ analysts can provide clients with enhanced insights into survey respondents’ thoughts, attitudes and opinions.
Contact Research Design Associates to learn more about Statistical Analysis.
The meaningfulness of research results depends on the quality of the sample studied. Whether it is a placebo-controlled double-blind randomized clinical trial or a client survey, Research Design Associates develops meaningful samples. The ultimate objective of research design and statistics is to make valid inferences about a population from information contained in a sample. The method of inference most often used in business and the social sciences is estimation.

Example: Estimating Insurance Overpayments

You may wish to estimate the amount of insurance overpayments in a medical practice, the percentage of voters who favor a particular candidate, or the type of customers who visit a store. Research Design Associates’ scientists are experienced in selecting the appropriate sample design for the given research question(s): the method of sample selection, sample structure, sample size, and the plans for analyzing and interpreting the results. 

Research Design Associates delivers on any study. Large or small projects are well-designed with an appropriate sampling scheme (e.g., simple random, stratified random, two-stage cluster), and correct use of inferential statistics to estimate population values.