Carlton Statistics in Psychology Final Exam NAME: INSTRUCTIONS: For this exam, you will design and test one of each of the listed hypothesis test types, using the dataset “FinalExamData.sav,” This...

final exam for statistics in psychology


Carlton Statistics in Psychology Final Exam NAME: INSTRUCTIONS: For this exam, you will design and test one of each of the listed hypothesis test types, using the dataset “FinalExamData.sav,” This time, you are in the driver’s seat and will have to decide what research questions YOU want to answer, using the data that is provided for you! This dataset contains randomly generated data from 20 different demographic and college-test related variables from 100 seniors attending different state universities in NJ. For each test, you should do the following (you must show each numbered and lettered step below). 1. Create a research question that you can answer appropriately using the test and using the available data. Each variable contains labels that describe what was measured, and there are a variety of nominal and scale variables from which you can choose. 2. Identify the appropriate variables for your research questions, their measurement type, and their label (as this information appears in SPSS) 3. Compute descriptive statistics using Analyze Descriptives Explore command and copy your output table. You should also write a few sentence description summarizing your variables. 4. Complete the following three steps of hypothesis testing a. Step 1: Null/Alternative hypotheses b. Step 2: Alpha level, degrees of freedom (if appropriate), one/two tailed test type (if appropriate), other information as appropriate c. Step 3: Appropriate test statistic/table, as a copied and pasted/screenshot of SPSS results i. Post hoc tests if appropriate d. Step 4: A full write up for step four, including: i. A decision about the null hypothesis ii. A written explanation about the results of the test, including means and standard deviations, and a final answer to your research question iii. Proper APA style string, including effect size and confidence intervals, if appropriate. Please conduct one of each of the following hypotheses tests: 1. Paired-Samples t-test (10 points) 2. Independent Samples t-test (20 points) 3. One-Way ANOVA (20 points) 4. Correlation (10 points) 5. Regression (20 points) 6. Two-way Chi Square (20 points) BONUS: generate an appropriate chart for each test using the Graphs Chart Builder or the Graphs Legacy Dialogue (or, as in ANVOVA, the chart option built into the dialogue box) in SPSS. EXAMPLE ANSWER for ANOVA 1. Is there a significant difference in freshman GPA between students at the three different schools (Rowan, Rutgers, or TCNJ)? 2. IV: “College.” Label: “College Student is Attending.” Measurement Type: Nominal DV: “GPA_Freshman.” Label: “GPA at end of Freshman Year.” Measurement Type: Scale 3. 4. Step 1. H0: Ha: Not all are equal. Step 2. = .05 Step 3. Test of Homogeneity of Variances     Levene Statistic df1 df2 Sig. GPA at end of Freshman Year Based on Mean 1.166 2 97 0.316 Based on Median 0.878 2 97 0.419 Based on Median and with adjusted df 0.878 2 96.9 0.419   Based on trimmed mean 1.157 2 97 0.319 ANOVA: GPA at end of Freshman Year   SS df MS F Sig. Between Groups 0.305 2 0.152 0.252 0.778 Within Groups 58.568 97 0.604 Total 58.873 99       Step 4. Retain H0. Levene’s test showed that the variance was equal between the three groups, F(2, 97) = 1.166, p = .316. There was not a significant difference in GPA at the end of freshman year between students at Rowan (M = 2.34, SD = .82), Rutgers (M = 2.32, SD = .73), or TCNJ (M = 2.16, SD = .77). Students from all three schools had GPAs that were essentially equal at the end of school year, F(2, 97) = 0.252, p = 0.778, = .01. NOTES on this ANOVA EXAMPLE: 1. You cannot use this example as one of your tests! 2. “My God, man! How did Carlton get those tables so nice?! Curse him and all his descendants!” Here is the trick: I copied the table right from SPSS into Excel. That (usually) keeps everything in the right place, and allows me to add the fancy APA style horizontal line breaks which you’ll learn about in research methods. 3. I’ve not included the descriptive statistics here, but YOU NEED TO. You may use these variables for other analyses, so I didn’t want to provide you the answers. 4. A significant Levene’s test means a different course of action for ANOVA and Independent Samples t-test. (We haven’t used it for ANOVA, so I’m not concerned if you don’t include it. However, you do need to include this information for your independent samples t-test, so I wanted to give you an example). I’ve used APA style to report it here, feel free to copy my example using your numbers. If your Levene’s test is significant, make sure to include that and whatever correction you had to make in your write-up! 5. Because this test was not significant, I did not copy and paste or report the post-hoc tests. If your ANOVA is significant, that table should be included. LIST of SYMBOLS to COPY: H0 Ha ± 1 1 rpb2 Microsoft Word - Review of Tests and SPSS Instructions.docx Carlton Statistics in Psychology 07301 1 Overview of Hypothesis Test Types in SPSS Heiman, G. (2011). Behavior Sciences STAT (1st Ed). Cengage Learning. Carlton Statistics in Psychology 07301 2 One-Sample t-test Description: The one-sample t-test tests the probability that the difference between the sample mean and a test value is due to chance. Example: You are a report author for a bank. Your research tells you that the mean national household income is $50,740. The mean household income of the bank's customers is above the national average. You want to know if the difference in mean household incomes is genuinely higher than average, or if this the result of chance. Variables in SPSS The variable should be a scale (interval or ratio) measurement type. Probabilities (significance value) of .05 or less are typically considered significant. Statistical Hypotheses: Null Hypothesis: The underlying population mean of the sample is not different from the actual population mean (apparent differences due to chance) Alternative Hypothesis: The underlying population mean of the sample is significantly different from the population mean (comes from a different population with a different mean) H0: ?samplevariable = Value of Population Mean Ha: ?samplevariable ≠ Value of Population Mean SPSS Instructions: 1. Click AnalyzeàCompare MeansàOne-Sample t Test 2. In the dialog box, move over your variable to the “Test Variable(s)” box a. Make sure to change your Test Value to whatever your population mean is. (Don’t leave this value as 0!) 3. Click OK. Interpretation: Probabilities (P value) less than .05 (or less than specified ?) are considered significant and the null hypothesis is rejected. P value can be found in the One-Sample Test table under column Sig. (two-tailed). Effect Size: rpb2 = ("!"#) $ ("!"#)$$%& APA Style: t(df) = tobt, p = sig, rpb2 = value Carlton Statistics in Psychology 07301 3 Related Samples (Paired Samples) t-test Description: All T-tests compares sample means by calculating Student’s t and displays the two-tailed probability of the difference between the means. Paired-samples t-test compares the means between two groups of scores that are connected in a meaningful way, such as 1) scores on two variables from the same participants, but at different time points; 2) scores between two different individual that have been previously matched together. Example: You are a school psychologist. You want to compare the number of behavioral outbursts in the month of September to number of behavioral outbursts for the same class in December. Variables in SPSS: The two connected variables should be a scale (interval or ratio) measurement type and should be measured as separate variables in SPSS (eg, two separate columns of scores) Statistical Hypotheses: Null Hypothesis: The population mean of the difference scores from time 1 to time 2 is not different from 0, so no differences exist. Alternative Hypothesis: The population mean of the difference scores from time 1 to time 2 is different from 0, so no differences exist. H0: ?D = 0 Ha: ?D ≠ 0 SPSS Instructions: 1. Click AnalyzeàCompare MeansàPaired samples t Test 2. In the dialog box, move over your two variables to the “Paired Variables” box next to pair 1 (SPSS will automatically pair them). 3. Click OK. Interpretation: Probabilities (P value) less than .05 (or less than specified ?) are considered significant and the null hypothesis is rejected. P value can be found in the Paired Samples Test table under column Sig. (two-tailed). Effect Size: rpb2 = ("!"#) $ ("!"#)$$%& APA Style: t(df) = tobt, p = sig, rpb2 = value Carlton Statistics in Psychology 07301 4 Independent Samples t-test Description Independent samples t-test compares scores from two groups of subjects that are not related to one another in any meaningful way. The single dependent variable should be a scale (interval or ratio) measurement type. (Independent samples t-test assumes that there is homogeneity of variance, that is, that the variance within each of the groups is equal. You can check for homogeneity of variance by using the Levene's test.) Example: You are a college advisor at a university comparing whether seniors who transferred from community college or seniors who spent all four years at university have comparable GPAs. Variables in SPSS: The grouping variable (the IV) should be a nominal level variable that designates participants (each row) as belonging to one of two groups, typically measured as “0” and “1” or as “1” and “2”. The dependent variable should be scale measurement type and should be