Research Paper 2: How to Determine Appropriate Levels of Analysis

As you work on this assignment, think about how far you have come in this course and use what you have learned to inform your approach to the assignment. As a reminder, the learning outcomes associated with this assignment are:

•Identify and define public problems to determine the type of analysis required.

•Organize data, construct analysis, and analyze data to generate reports for decision making using Excel.

•Create and analyze basic statistical hypotheses, tests, and tools.

•Engage in critical thinking, analysis, and forecasting to analyze and interpret data to formulate and influence policy options.

In a 3-5 page paper, complete the following:

1. Building upon the previous assignment, use state or local websites to obtain data relevant to your identified problem. Calculate the central tendencies.

2. Present the descriptive statistics in an APA-formatted chart. Discuss why your selected data is analytically useful.

3. Generate a hypothesis about your problem. Explain why your hypothesis would reduce or alleviate the presence of this problem.

4. Include at least two variables (one independent and one dependent) that may impact the problem. Evaluate the relationship between the variables.

5. Using your hypothesis, propose a statistical tool to analyze your data.

6. Construct a hypothetical scatterplot chart of how your variables may interact based on your proposed analysis.

7. Use proper APA formatting and citations to support your position. A minimum of four peer-reviewed sources, one of which must be a government site, is required.

Module 4: Module Notes: Surveying Citizens or Customers?

◾Atomic Learning. (n.d.). Excel 2013 – Protecting your spreadsheets (Links to an external site.)Links to an external site.. Retrieved from http://www.atomiclearning.com/highed/excel-2013-protect-spreadsheets-training

Cassia, F. (2008). Citizen surveys impact on decisions in local government (Links to an external site.)Links to an external site.. Journal of Services Research, 8(1), pp. 109-122. Retrieved from http://vlib.excelsior.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=32732748&site=eds-live&scope=site
direct=true&db=bth&AN=32732748&site=eds-live&scope=site

Recently, Americans have seen a number of issues that were once considered controversial introduced in the form of legislation or added to local voting ballots. In order for those items to be added to the ballots, grassroots organizations and nonprofits collect a number of signatures. This is known as petitioning. Petitions come in all forms, including electronic, face to face, phone, or mail. Petitions are used to suggest changes to a wide range of public (and perhaps not so public) concerns.

Petition
Petitions are similar to surveys in that they are used to obtain an overview of how a group of people view an issue. Similar to cluster sampling, petitions work because they require a percentage of the population to respond affirmatively. However, petitions are easier to execute than a survey because they focus on a single subject and require only a signature to show whether a person agrees or disagrees with the issue presented.

Survey
As you can see, it appears that it is easier for the nonprofit, or even the private sector, to make decisions using surveys. Typically, nonprofits have already identified their opinion and simply look for support, while the private sector is likely to execute any action the customer finds favorable, in order to enhance profits. However, in the public sector, political will or bureaucratic necessity may outweigh public opinion. Administrators must be very careful to balance the needs of the citizenry with the desires of the citizens as customers. In both cases, it is essential that public administrators are capable of designing and executing an effective survey.

Module Notes: Designing Surveys & Questionnaires

Once you have identified your sample, you are ready to design your survey. Your survey design is, again, dependent upon your sample. For example, it might not be useful to design an electronic survey for citizens who are unlikely to respond due to lack of Internet connectivity or familiarity. In fact, reaching your sample, or getting your population to respond to the survey, is perhaps one of the most difficult aspects of executing a survey. Therefore, it is important to select the type that best suits your sample

Types of Surveys
There are generally four types of surveys to choose from:

Designing Survey Questions
No matter what survey type you select, you must ask good questions. The questions should be neutral and unbiased, so as not to influence the response in any way. Have multiple people, who have different perspectives, read each question so you can delete as much bias as possible from the questions. In addition, you should avoid double-barreled questions, which ask two questions in one sentence and both can be answered simply with a yes or no. Finally, you should save personal or demographic questions for the end of the survey. Respondents may become uncomfortable if too much personal information is collected at the beginning of the survey and feel hesitant to complete it.

Expert Help in Conducting a Survey
In some cases, it may be necessary to hire an expert to write and conduct your survey. If your agency lacks an expert who is capable of executing the survey and generating a good response rate, you may consider a consultant. Some agencies, especially smaller state and local agencies, simply lack the human capital to execute surveys. In addition, the agency may want to be sure that the survey remains objective, and therefore seek out a consultant to ensure the survey is not biased.

Focus Groups

If it is not possible to conduct a survey, a focus group is a good alternative. A small target group of citizens can provide a great deal of information about a public problem or service. Focus groups work best when facilitated by a moderator who is able to create a comfortable environment that encourages self-expression. Focus groups may prove to be more effective than surveys because the process invites a variety of responses rather than the fixed set of responses offered in surveys.

Common Types of Analysis
Once the data is collected, you must determine the type of analysis that is best suited for the data you have obtained. Most likely, you already have some idea about the type of analysis you would like to use because you have already determined whether your data will be qualitative or quantitative, and determined the scale of measurement (nominal, ordinal, ratio, or interval). Your analysis will also be influenced by the type of study you have conducted (observational, correlational, cross-sectional, or longitudinal). Whether qualitative or quantitative, you will use this data to calculate the descriptive statistics. You may even place the descriptive statistics in a graph or some other visual aid as you begin to consider additional analytical procedures. There are a variety of statistical tools, each used for specific data types and scenarios. The Chi-Square, T-Test, and ANOVA are common statistical tools that can be performed using Excel.

Finding Relationships
Once you have selected and run your statistical test(s), you will begin to look at relationships between the variables that were extracted from the survey questions. When both variables increase, there is a positive relationship between the two variables. This is known as statistical significance. A relationship is considered to be statistically significant if the analysis shows that the test could be repeated and generate the same results 95% of the time.

Recently, there has been some debate about whether statistical significance is useful in public sector research. It has been suggested that public administrators give more consideration to confidence intervals, which provide a percentage of a sample as a response. Using the confidence interval, a 55%-65% range would be a significant finding and potentially impact the decision-making process.

To Sum Up
Determining whether findings are significant is contextual. It may vary according to each scenario. In any case, it is most important to remember that people are not numbers. No matter what statistical analysis is used, administrators must consider more than the numeric data when making decisions. This is where subjective qualitative data such as interviews, focus groups, and case studies may also serve as useful supplements.