1. Report your 1) independent variable and 2) dependent variable. In a few sentences, explain why you think your independent variable might influence the dependent variable. For this project I decided to assess the relationship between number of children, the independent variable, and the respondent's perception of government support for daycare services, the dependent variable. I believe that the more children a person has, the more inclined they will be to support help with daycare because it's terribly expensive! *For the purposes of illustration, I re-coded number of children, usually an interval-ratio variable, into an ordinal variable. The categories are 0, 1, 2, and 3 or more children. 2. State the null and research hypothesis using symbols and words. H0: There is no association between number of children and daycare support in the population. H1: Number of children and daycare support are statistically dependent. 3. Perform a cross-tabulation of your variables in SPSS. Also have SPSS produce column percentages. In Word or Excel, produce your own cross-tabulation table with column percentages AND row/column totals table using the information gathered from SPSS. HINT: Your independent variable should be reported as the column variable and your dependent variable should be the row variable. Table 1: Perception of Childcare Support by Number of Children Number of Children Support 0 1 138 72 Too Little 47.4% 51.8% 124 60 About Right 42.6% 43.2% 29 7 Too Much 10.0% 5.0% 291 139 Total 100.0% 100.0% 2 125 45.5% 118 42.9% 32 11.6% 275 100.0% 3 or More 106 49.5% 88 41.1% 20 9.3% 214 100.0% Total 441 48.0% 390 42.4% 88 9.6% 919 100.0% 4. Report (just a list) the obtained Pearson chi-square statistic, the degrees of freedom, the P-value, and the chi-square critical value. Pearson chi-square statistic: 5.305 Degrees of freedom: 6 P-value: 0.505 Chi-square critical value: 12.592 (Alpha = .05; 6 df) 5. Using appropriate statistical language, make a decision regarding your null hypotheses (one sentence). I failed to reject the null hypothesis: The number of children a person has and their perception of childcare support by the government isn't related. 6. Select and report the most appropriate measure of association for your variables. You can report lambda, Cramer's V, gamma, Kendall's tau-b or Kendall's tau-c, depending on the nature of your variables. Here you need to choose the CORRECT measure of Association and Statistical Test. Please refer to the How-to instructions and the memo for bivariate relationships for instructions on how to assess crosstabs with ordinal and/or nominal data. For this test, I used Gamma and Kendall's Tau-C. Gamma: 0.000 There is no relationship between number of children a person has and their perception of childcare support by the government ( = 0.00). Kendall's Tau-C: 0.000 There is no association between number of children a person has and their perception of childcare support by the government (b = 0.00). Using information about respondents' number of children and their perception of childcare support by the government, we did not reduce our prediction error. 7. In three to seven sentences, explain what the findings from the cross-tabulation, chi-square test, and the measures of association mean in words. Make sure you use appropriate statistical language. Here, I expect a full paragraph describing the relationship you thought you would find, the information in the table, (i.e., the relationship you thought you had and the one you observed in the data. I also want a somewhat detailed discussion of significance and measures of association. Mine will be brief so that you don't copy it in your answers. So, I expected that groups with greater numbers of kids would be associated with perceptions of "not enough" childcare support. For example, I thought that folks with 0 or 1 kid would perceive that there is just enough or too much support compared with people that had 2 or 3 kids -- I was wrong. Nearly across the board, citizens perceived that the government provides either too little or just enough support for childcare. I think the categories of children might have been coded with greater distance, there might have been a difference -- that is, would you think support would differ that much for people with 1 versus 2 kids? Unlikely. For more information on childcare resources and spending, please visit: https://www.usa.gov/child-care. HINT: It's not much. ------------------------------------------------------------------------------------------------------------------PART 2: REGRESSION AND CORRELATION 8. Conduct the regression and write the complete regression equation. y-hat = 9.711 =+ 0.31 (x) 9. Using appropriate statistical language, explain what the regression equation means in two to five sentences. This equation means that a one-hour increase in time spent on the Internet (independent variable) results in a 0.31 hour increase in time spend on household work. Essentially, the more time spent on the internet, the more time spent on household labor. I did not expect this relationship at all -- I thought that more time on the Internet would result in less time performing household labor. 10. Report and label the correlation coefficient. The correlation coefficient is .045. NOTE: You will have to report the CORRECT NAME OF THIS COEFFICIENT. 11. Using appropriate statistical language, interpret the correlation coefficient in one to two sentences. This _________ coefficient indicates that there is a weak, yet positive association between the hours spent on the internet and hours spent performing housework. 12. Report and label the coefficient of determination. The coefficient of determination is .002. NOTE: You will have to report the CORRECT NAME OF THIS COEFFICIENT. 13. Using appropriate statistical language, interpret the coefficient of determination. Nearly NONE of the variation in hours spent performing housework is explained by hours spent on the internet. NOTE: You will need to use NUMBERS for your interpretation as well as a line or two on what that means for your analysis generally. ------------------------------------------------------------------------------------------------------------------PART 3: ANOVA 14. State the null and research hypothesis using symbols and words. H0: 1 = 2 = ... k: There is no difference in the mean hours spent on the Internet by marital status. H1: There is at least one mean difference in the hours spent on the internet by marital status. 15. In a few sentences, explain why you think your independent variable might influence the dependent variable. You would need to elaborate more here, but overall, I think that if people are married, they will likely spend less time on the internet compared with those that are widowed, divorced, separated, or never married. List the SSB, SSW, dfb, dfw, MSb, MSw, the obtained F-statistic, the P-value, and the F-critical value. SSB: 4119.326 SSW: 2000916.674 dfb: 4 dfw: 920 MSb: 1029.831 MSw: 218.388 Obtained F-statistic: 4.716 P-value: .001 F-critical value: 2.37 16. In a sentence, and using appropriate statistical language, make a decision regarding your null hypotheses. My F-obtained exceeds my F-critical, which means I can reject my null hypothesis -- there is at least one mean difference in the amount of hours spent on the internet by marital status. 17. Examine the table created from Tukey's HSD. In a list, report all statistically significant differences (i.e., p<.05). Report the mean difference and the P-value for each statistically significant difference. HINT: If there are none, you must clearly state there are no statistically significant differences between groups in order to receive credit for this portion of the assignment. Difference between Married and Never Married: -4.080 hours; p value: 0.003 18. In three to five sentences, explain what the findings from the ANOVA and Tukey's HSD mean, using your own logic and statistically appropriate language. You would need to elaborate more here, but I found out there were mean differences in the hours spent on the internet among marital status groups; however the relationship was not as clear as I suspected. Only one group significantly differed: Married and non-married. Married people spend a little over 4 fewer hours on the internet compared with never married respondents. Other groups were not significantly different from each other; however, some were close (e.g., widowed and never married and divorced and never married). What this finding also means is that married, divorced, separated, and widowed people spend about the same amount of time on the internet! PART 2: Regression and Correlation (25 points) In PART 2, you will perform a regression and correlation. You will need to select two appropriate variables for this portion of the assignment. NOTE: Be sure to select an independent and dependent variable. Also remember which levels of measurement are appropriate for regression. If you select variables with inappropriate levels of measurement, you may lose credit on this portion of the assignment. Your is set at 0.05. independant = age, dependant = general happiness 8. Conduct the regression and write the complete regression equation. 9. 10. 11. Using appropriate statistical language, explain what the regression equation means in two to five sentences. Report and label the correlation coefficient. The correlation coefficient is -.037 Using appropriate statistical language, interpret the correlation coefficient in one to two sentences. This _________ coefficient indicates that there is a weak, yet positive association between the age for general happiness. 12. 13. Report and label the coefficient of determination. The coefficient of determination Using appropriate statistical language, interpret the coefficient of determination. Coefficientsa Standardized Unstandardized Coefficients Model 1 B (Constant) Coefficients Std. Error Beta 1.706 .070 .002 .002 AGE OF RESPONDENT t 24.499 .000 1.003 .316 .037 a. Dependent Variable: GENERAL HAPPINESS ANOVAa Model 1 Sum of Squares Regression df Mean Square .339 1 .339 Residual 250.272 742 .337 Total 250.612 743 a. Dependent Variable: GENERAL HAPPINESS b. Predictors: (Constant), AGE OF RESPONDENT Model Summary Model 1 R .037a R Square .001 Adjusted R Std. Error of the Square Estimate .000 a. Predictors: (Constant), AGE OF RESPONDENT .581 F 1.007 Sig. Sig. .316b