Report

By Wendiann Sethi Spring 2011 The second stages of using SPSS is data analysis. We will review descriptive statistics and then move onto other methods of data exploration using crosstabulations, inferences on the mean, regression and ANOVA. Students are encouraged to bring data that they are analyzing in class or projects to discuss what methods would be best to use. Descriptive statistics Identifying outliers Missing data – by variable, by respondent Manipulating data – ◦ reversing the scale ◦ Collapsing a continuous variable into groups Choosing the right statistic Data Analysis: Multiple Responses ◦ ◦ ◦ ◦ Cross tabulations Correlations and regression Tests about mean and proportions ANOVA Measures of central tendency Measures of spread Analyze> Descriptive Statistics What do you use for each type of variable and why? Look at the mean, median, st.dev and skewness to determine if there might be outliers Could also use a box plot to see Two potential problems with missing data: 1. Large amount of missing data – number of valid cases decreases – drops the statistical power 2. Nonrandom missing data – either related to respondent characteristics and/or to respondent attitudes – may create a bias Examine missing data By variable By respondent By analysis If no problem found, go directly to your analysis If a problem is found: Delete the cases with missing data Try to estimate the value of the missing data Use Analyze > Descriptive Statistics > Frequencies Look at the frequency tables to see how much missing If the amount is more than 5%, there is too much. Need analyze further. 1. 2. 3. 4. 5. 6. 7. 8. Use transform>count Create NMISS in the target variable Pick a set of variables that have more missing data Click on define values Click on system- or user-missing Click add Click continue and then ok Use the frequency table to show you the results of NMISS Use Analyze>descriptive statistics>crosstabs Look to see if there is a correlation between NMISS (row) and another variable (column) Use column percents to view the % of missing for the value of the variable Proceed anyway Estimate (impute) the missing data with substituting the mean or median value Recoding Calculating When to create a new variable versus creating a new one. What do you want to explore? What do you need in your data to do that exploration? Analyze>Descriptive Statistics>Crosstab Good for categorical data to see the relationship between two or more variables Statistics: correlation, Chi Square, association Cells: Percentages – row or column Cluster bar charts Finding the relationship between two scale or ordinal variables. Analyze > Correlate > bivariate Analyze > regression > linear Aim: find out whether a relationship exists and determining its magnitude and direction Two correlation coefficients: Assumptions: ◦ Pearson product moment correlation coefficient –rinterval or ratio scale variables ◦ Spearman rank order correlation coefficient –rhoordered or ranked data ◦ ◦ ◦ ◦ Related pairs of scores Relationship of the variables is linear Variables are measured at least at the ordinal level Homoscedasticity – variability of y variable should remain constant at all values of x variable Aim: Use after finding there is a correlation to find an appropriate Linear model to predict the results of the DV based on one or more IV’s Assumptions: ◦ ◦ ◦ ◦ Related pairs of scores Relationship of the variables is linear Variables are measured at least at the ordinal level Homoscedasticity – variability of y variable should remain constant at all values of x variable Procedure: Linear Regression ◦ ◦ ◦ ◦ One IV to one DV ANALYZE>REGRESSION>LINEAR After placing the appropriate DV and IV, click STATISTICS Click CONTINUE and then OK to run the analysis Comparing the means of a scale (or ordinal) when grouped by a category Analyze > Compare means ◦ Means – simplest form DV – scale to be compared given the IV – categories ◦ One-sample t-test : test the mean of the variable against a set value. ◦ Independent samples t-test: looking at the difference of two means of the variable given a grouping variable (twogroups only) ◦ Paired-samples t-test: looking at the difference of the means when there is paired data (pre-test vs post-test) ◦ One-way ANOVA: comparing the means of dependent variables (scale or ordinal) given a factor (one IV-category) Aim: Testing the differences between the means of two independent samples or groups Requirements: Assumptions: Procedure: ◦ Only one independent (grouping) variable IV (ex. Gender) ◦ Only two levels for that IV (ex. Male or Female) ◦ Only one dependent variable (DV) ◦ Sampling distribution of the difference between the means is normally distributed ◦ Homogeneity of variances – Tested by Levene’s Test for Equality of Variances ◦ ◦ ◦ ◦ ◦ ANALYZE>COMPARE MEANS>INDEPENDENT SAMPLES T-TEST Test variable – DV Grouping variable – IV DEFINE GROUPS (need to remember your coding of the IV) Can also divide a range by using a cut point Aim:used in repeated measures or correlated groups designs, each subject is tested twice on the same variable, also matched pairs Requirements: ◦ Looking at two sets of data – (ex. pre-test vs. post-test) ◦ Two sets of data must be obtained from the same subjects or from two matched groups of subjects Assumptions: ◦ Sampling distribution of the means is normally distributed ◦ Sampling distribution of the difference scores should be normally distributed Procedure: ◦ ANALYZE>COMPARE MEANS>PAIRED SAMPLES T-TEST Aim: looks at the means from several independent groups, extension of the independent sample t-test Requirements: Assumptions: Procedure: ◦ Only one IV ◦ More than two levels for that IV ◦ Only one DV ◦ The populations that the sample are drawn are normally distributed ◦ Homogeneity of variances ◦ Observations are all independent of one another ANALYZE>COMPARE MEANS>One-Way ANOVA Dependent List – DV Factor – IV How to deal with questions were the participant can choose several choices. ANALYZE>MULTIPLE RESPONSE ◦ Define sets ◦ Frequencies ◦ Crosstabs Example data: survey_sample.sav ◦ Eth1, 2, 3 – multiple response method ◦ News 1, 2, 3 – multiple dichotomy method Wendiann Sethi Wendiann.sethi@shu.edu AS 202C or SC 128