Statistical Research Design

Statistical Research Design

Summary
SurveySilo.com gathers responses to online surveys and polls for the sole purpose of modeling social trends. Its mission is
to advance research in the social sciences. Responses to our surveys and polls are made with scaling, such as strongly
agree . . . agree . . . disagree, dropdown list choices, yes, no, maybe answers, and comments. Responses are analyzed in
the agregate using standard statistical methods.

Design
SurveySilo.com collects data on opinions, attitudes, values, beliefs, and perceptions about items under study that can be
used to identify group and social trends for statistical modeling. Modeling group and social trends can guide decisions that
can affect our future. The mission of SurveySilo.com is to use discovered trends for better decision-making and ultimately
make our human existence more enjoyable, egalitarian, and sustainable. SurveySilo.com policy and research rigorously
analyzes the many aspects of human behavior in the effort to succeed in its mission so that all stakeholders can benefit.

Our surveys and polls collect quantitative information about items under study. In addition, factual information is sometimes
gathered to construct demographic profiles of the aggregate of respondents. Surveys and polls designed and used by
SurveySilo.com are common in social science research. Questions are structured and ordered in such a way as not to
influence responses to subsequent questions.

SurveySilo.com uses a cross-sectional research design for gathering responses. A cross-sectional research design is more
commonly referred to as a survey or poll in that it collects data about backgrounds, past experiences, and attitudes from a
cross-section of respondents, a sample that is assumed to represent the general population under study. Cross-sectional
research design in conjunction with additional statistical analyses allows approximations of causality and relationships
among variables. SurveySilo.com utilizes cross-tabulations, filters, and bivariate analyses to allow these approximations
where possible without compromising validity and reliability.

SurveySilo.com utilizes surveys and polls designed to collect data that can be measured on several levels. Likert Scales
and Guttman Scaling are commonly used to collect nominal, ordinal, interval, or ratio measured data. A respondent's
answer to an open-ended question is coded into a response scale afterwards. Every attempt is made to standardize
SurveySilo.com surveys and polls in an effort to make them relatively free from several types of threats to internal validity.

One important factor of internal validity is the way the sample of respondents is selected. Online use of surveys and polls
certainly brings an aspect of risk to the generalizability of the results gained from an online sample to that of the wider
population. In other words, SurveySilo.com cannot perform a probability random sampling of users for its surveys and polls.
Participation is voluntary and done online. People who choose to respond to online surveys and polls may be different from
those who do not respond, thus biasing the estimates. Careful use of biased sampling can be used if it is justified and as
long as it is noted that the resulting sample may not be a true representation of the population under study. Respondents
are allowed to participate in a survey or a poll a maximum of one time. Very large sample sizes are possible because of the
online nature of survey or poll delivery, thus making generalizations to the wider population more reliable.

Raw data is analyzed using advanced statistical analysis procedures. The following types of analyses can be performed on
most data derived from SurveySilo.com instruments:

Descriptive Statistics
t-tests
Analysis of Variance (ANOVA)
Pearson's r correlation
Regression Analysis
Nonparametric Chi-square
Multivariate Analysis (MANOVA)

Survey and poll results are available to users of SurveySilo.com. We use charts, tables, indexes, and/or Excel
spreadsheets to report results on a periodic basis. Please see our Results page to find out more.