Statistical Analysis, Research Design, Data Management, and Scientific
Editing Services
Select a link on the left to see a detailed description of our services.
If possible, we prefer to be involved with researchers at the planning stage, because this greatly reduces the possibility of errors. Statistical analysis is effective when the data is well organized and comes from a sound research design. If you request an estimate from us, we will evaluate the project and recommend those services that yield high quality results.
Research Design Consultation
A weak or inadequate research design cannot be overcome by one choice of statistical method or another. This is not often considered by many researchers, but it is crucially important. Well regarded statistician John Tukey said
“The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data.”
If the body of data comes from poorly designed research, the results of the statistical analysis will not be reasonable. SAMC can provide research design recommendations that yield high quality results. Most methodology proposals contain the following:
- a well-defined population
- a method of sampling
- clear focused hypotheses
- a power analysis
- an appropriate choice of statistical method
Our research design service provides all of this as well as an extensive amount of general research design advice. SAMC, Inc. has helped our clients successfully apply for NIMH, NIH R01 grants, as well as grants from other government funding agencies. This service is also very useful for academic, business, and graduate student researchers.
Data Management Consultants
We have experience organizing and maintaining national biomedical databases. For researchers involved in multi-site clinical trials, we can create, maintain, and update confidential data.
For smaller research projects, data must be arranged in a highly specific numerical format before running a statistical analysis. If the data is not arranged properly or if the data is in a proprietary format, there may be no way to effectively analyze it. This is one of the most frequent problems that we encounter. Unless you have experience organizing data for statistical analysis, we strongly suggest that you seek advice before doing so. There is one right way to proceed and an almost infinite number of wrong ways. We know the right way.
Our data management service involves designing a data shell and giving project specific instructions for data entry. We then create a numerical data file with appropriate variable and value labels. We may have to transform and create new variables depending on the type of analysis we run.
We offer data entry as an optional service. We will need to know the sample size and number of variables in the research to give an accurate estimate. Obviously, the larger the sample size, the longer it will take for us to enter the data.
Statistical Analysis
We screen the data for errors, run the statistical analysis, and summarize the conclusions. We transfer the results from a statistics program to a word processed summary document. This may include text, tables, and high resolution graphs depending on the analysis. There is a detailed teaching explanation of the statistical method. We highlight the central points of the statistical output and explain if the central hypotheses/research questions have been supported. Many clients use this summary document as a starting point for writing up their results.
Scientific Editing
We edit your results section as if it were being submitted for scholarly publication. Our input usually involves comments to improve tables, text, and the overall presentation of the statistical output. Most research projects need only one edit. Having more than two edits is rare.
Be advised that this is an editing service, not a writing service. If you are a graduate student looking for someone to write your dissertation for you, look elsewhere.
All edits are done on a hardcopy printout, photocopied, and then mailed to the client. We are familiar with the popular APA format for scientific writing.
Test/Survey Construction
We can guide you through the process of developing a survey or test. This
can be a lengthy multi-stage process. This may include developing a table of
specifications (a.k.a., test blueprint), writing items, factor analyzing the
items, and assessing the psychometric properties (reliability, etc.) of the
measure. This is usually a multi-stage process.
Survey and Test Construction Services - Top
What methods have you used?
Many - and this is not an exhaustive list. We are not experts with a hammer who think everything is a nail.
Our approach is to get a clear understanding of your research
questions and select an appropriate statistical method to answer your
hypotheses. Following are some of the statistical methods we have
applied to such fields as business productivity, grant funded research,
marketing research, and academic research.
We are statisticians who have helped clients interpret the results from independent
and dependent samples t-tests, analysis of variance
(ANOVA), analysis of covariance (ANCOVA), multivariate analysis of
variance (MANOVA), repeated measures analysis of variance, and linear mixed
models analysis.
We have assisted researchers with Pearson correlation, point biserial
correlation, partial correlation, Spearman rank correlation, Kendall's
tau, simple linear regression, multiple regression (hierarchical,
stepwise, with or without the use of blocking variables), ordinal
regression, binomial logistic regression, multinomial logistic
regression.
S.A.M.C., Inc. has provided consultation with advanced multivariate
methods including canonical correlation, structural equation modeling,
path analysis, confirmatory factor analysis, common factor analysis,
principal components analysis, discriminant function analysis,
hierarchical cluster analysis, model selection log linear analysis,
general log linear analysis, and logit log linear analysis.
Statistical consulting for relatively simple output (means, standard
deviations, odds, odds ratios) as well as more advanced tests (Wilk's
Lambda, F tests, Greenhouse-Geisser, Mauchly, Bartlett, etc.) depending
on the task.
We have helped clients understand Phi, Cramer's V, Kruskal Wallis,
analysis of contingency tables,
Kolmogorov-Smirnov, Wilcoxon, binomial, crosstabs, chi-square,
Mann-Whitney U, Moses extreme reactions, Wald-Wolfowitz runs, sign test,
McNemars test, Friedmans test, Kendall's W, Cochran's Q, reliability
analysis (including Cronbach's alpha, split half, etc.).
S.A.M.C., Inc. can use our extensive graphing capability to help clients
present results using bar, line, and pie charts, ordinary frequency
distributions, boxplots, error bar charts, scatterplots, histograms
(with or without the normal curve), normal probability plots (P-P or
Q-Q), and detrended normal probability plots.
In the field of medical research, we have experience with ROC curves to
estimate sensitivity and specificity. In addition, we can run survival
type analyses including Life tables, Kaplan-Meier curves, and cox
regression.
We can't list them all. Regardless of the field of research, we
take pride in our ability to explain the output to researchers who may
not have a strong quantitative background, all for a reasonable fee.
Some of the Statistical Methods Used - Top