The following items should be carefully considered when conducting research. Although they will only be mentioned briefly here, a wealth of information is available on each item. Suffice it to say, that each is important to generate accurate and useful information:
Although it is true that anyone can construct a survey or interview, it is not necessarily true that anyone can construct one that will provide valuable or accurate information. Some of the issues that must be confronted in design include:- Order effects : That is, the order in which questions or statements are arranged can have much influence on how the person responds.
- Emotional Impact : Different statements or questions often elicit different emotional responses. The strength of these responses and the placement of these questions/statements in the design will have a distinctive impact on the information collected.
- Specificity of Statements/Questions : Concerns the placement of statements or questions in an information collection session based on how general or specific they are.
- Ambiguity : Concerns how ambiguous the statement/question is. That is, does it have more than one interpretation? If it does, it renders the information collected useless.
- Instructions : Concerns the placement, clarity, and impartial characteristics of the instructions.
Sampling is the technique used to determine who will be asked to respond to a survey or interview when it is unnecessary to obtain information from the total population. Some of the issues that should be considered when determining what will constitute an appropriate sample include:
- Sample Size : How large does the sample need to be?
- Sample Type: How will the sample be drawn from the total population? Will it be a random, proportional, or equivalent sample, etc.?
- Response Rate: How can response rates be increased without contaminating the data?
The type of scale used determines much in terms of the types of analyses that can be performed, and the degree to which you are maximizing the value of the information being taken. For example, although frequently used, a simple yes/no scale can yield much less valuable information than can be gained from a 7-point interval scale. Some considerations include:
- Type of Scale : e.g., ordinal, interval, ratio, forced choice, etc.
- Labeling : e.g., are the individual scale items labeled with individual descriptors that make the scale categorical in nature, or are only the extreme scores (e.g., 1 and 7) labeled which can make the scale interval? The answer to this question is important in determining the types of analyses that can be appropriately applied.
An important consideration in any research is the type of analyses that will be applied to the data collected. The type of analyses must be carefully planned to support the type of information sought, and must be based on appropriately crafted and applied scaling techniques. Some types of analyses performed by Insight Research, Inc. include:
- Measures of Central Tendency (e.g., Mean, Median, Mode and Standard Deviation)
- Comparative Statistics (e.g., T-test, ANOVA, MANOVA)
- Post Hoc Analyses to compare different levels of a variable in order to pinpoint the source of significant differences
- Correlation and Multiple Regression Analyses
- etc.