SPSS 1 Assignment Instructions
This assignment is designed to help teach you to describe a single variable – its central location, its dispersion, to create an appropriate graphic to illustrate the variable, and to discuss the way in which you variables distribute.
For each of the variables:
- Identify the level of measurement for each variable.
- Build a table that shows the cumulative % and frequencies.
- The table must be in APA format
- Some variables will have to be recoded to effectively display in a table. RULE OF THUMB – no more than 10 categories should appear in ANY table.
- Report the summary statistics that describes the variable in terms of all the appropriate measures of central location and dispersion.
- Create 1 appropriate graphic to display the distribution.
- Write a paragraph that describes the distribution in terms of central location, dispersion, outliers, and skew (if any).
These are your variables:
- From the WORLD 2012dataset use the variable named “polity” with the label “Higher scores more democratic (Polity)”.
- From the NES 2012 dataset use the variable named “dem_marital” with the label “Marital Status”.
- From the NES 2012 dataset use the variable named “relig_attend” with the label “Attendance: Religious Services”.
- From the GSS 2012 dataset use the variable named “wordsum” with the label “Number of words correct in vocabulary test”.
- From the GSS 2012 dataset use the variable named “educ_4” with the label “Education in 4 Categories”.
- Descriptive paragraph requirements:
- What does your variable measure?
- Describe this variable in terms of all appropriate measurements of central
- Describe this variable in terms of the appropriate measurements of dispersion.
- If appropriate, is this distribution skewed negative or positive?
- Discuss any other interesting and relevant details about this distribution.
SPSS1 Frequently Asked Questions and Point Breakdown
5 questions:
- Level of measurement:
- You must have the correct level of measurement
- You should not put interval/ratio. You must clearly identify if the data is interval or ratio for full
- You need to report the level of measurement on the original data, not the recoded data.
- Cumulative Frequency Table:
- You must have a title, a source and appropriate columns.
- Your table must be formatted according to APA guidelines. Week 6 under course content on our Blackboard Course website has both the template and an instructional video if you would like to refresh your memory from lab.
- If you have more than 10 rows you should recode the data. Make sure that the valid total of your recoded cumulative frequency table matches the valid total of the cumulative frequency table on the original data.
- Descriptive Statistics:
- Do not directly copy and paste from SPSS. You must format the tables according to the APA Guidelines.
- Measures of central tendency must be correct for the level of measurement of the data.
- Measures of dispersion must be appropriate for the level of measurement of the data.
- Always report the category labels for categorical data.
- You will need to perform some calculations. Remember, SPSS does not give you the IQR or the V ratio. You will need to calculate these out correctly for full points.
- Remember, if you have ordinal data the numbers associated with the data are just the coding Therefore, you need to simply the IQR and the Range as much as possible. (For example, if you were using a Likert Scale, you could end up with Hate-Love for the range and Somewhat dislike-Somewhat love for the IQR). Remember for categorical data you must work with the category labels.
- You must run statistics on the original data, not the recoded data. As such, your descriptive statistics should be appropriate to the level of measurement of your original data.
- Graphics:
- These can be copied and pasted from SPSS. You need to make sure that your graphs have titles and sources. You may use Excel but we strongly encourage you to use SPSS
- Your graph must be appropriate to the level of measurement of your data.
- (nominal = pie chart; ordinal = bar graph; interval/ratio = histogram)
- If using recoded variable, should use the level of measurement AFTER the recoding (interval recoded to ordinal should use bar graph)
- However, if you use a histogram with the interval/ratio original data, you will not be penalized
- Have value labels. Remember to use the crosshairs in the Graph Editor on SPSS. d. Bar chart should use % of cases (“Percent”) not count.
- Paragraph:
- You need to describe both central tendency and dispersion. Do not just laundry list all of the statistics. Focus on the most salient univariate statistics for your level of measurements (i.e. if you have ratio data, you should discuss the mean for central tendency and the standard deviation and the variance for measures of dispersion). You need to interpret theses as well. What do they tell you about the variable being measured? (For example, if you have a variable on age and the average age was 25, what could you infer about the population?). Report interesting trends from the statistics.
- Discuss the shape of the distribution curve (positive or negative skew) if it is appropriate for the level of measurement.
- Make sure that you report outliers if it is appropriate to the level of measurement for your data.