DRAFT: This module has unpublished changes.

Darius Muniz

Professor Godfrey

Research Methods II

September  19th, 2011

 

 

                                                                           SPSS                                          TextBook

I. a) HeartRate-                                               Scale                                                 Ratio

   b) Nervous-                                                  Scale                                               Interval

   c) Sex-                                                        Nominal                                           Nominal

   d) YearsInCollege-                                      Ordinal                                             Ordinal

   e) ExperimentalGroup-                               Nominal                                            Nominal

II. Nominal and Ordinal Variables:

 

What year are you in college?

 

Frequency

Percent

Valid Percent

Cumulative Percent

Valid

Junior

17

44.7

51.5

51.5

Senior

16

42.1

48.5

100.0

Total

33

86.8

100.0

 

Missing

99

5

13.2

 

 

Total

38

100.0

 

 

 

 

 

 

 

 

A) For “what year in college” (an ordinal variable) out of 38 students surveyed, 51 percent reported being in their junior year of college, 49 percent reported being in their senior year, and no students identified as freshman or sophomore.

B) 33 students, which made up 87 percent of the population, reported their academic level with 17 self identifying as juniors (representing 45 percent of the class) and 16 self identifying as seniors (representing 42 percent of the class). 5 students did not answer the question (representing 13 percent of the class) for a total population of N=38 students.

 

What is your sex

 

Frequency

Percent

Valid Percent

Cumulative Percent

Valid

Male

7

18.4

21.2

21.2

Female

26

68.4

78.8

100.0

Total

33

86.8

100.0

 

Missing

99

5

13.2

 

 

Total

38

100.0

 

 

 

A) For “what is your sex” (a nominal variable) out of 38 students surveyed, 21percent self identified as male and 79 percent self identified as female.

B) 33 students, which made up 87 percent of the population, reported their sex with 7 students (representing 18 percent of the class) self identifying as male and 26 students (representing 69 percent of the class) self identifying as female. 5 students did not answer the question (representing 13 percent of the class) for a total population of N=38 students.

 

What experimental group were you in?

 

Frequency

Percent

Valid Percent

Cumulative Percent

Valid

A (something good that happened to you)

9

23.7

27.3

27.3

B(something good happened to someone you knew)

10

26.3

30.3

57.6

C(write letters)

14

36.8

42.4

100.0

Total

33

86.8

100.0

 

Missing

99

5

13.2

 

 

Total

38

100.0

 

 

A) For “what experimental group were you in” (a nominal variable) out of 38 students surveyed, 27 percent were assigned to condition A, 30 percent were assigned to condition B, and 42 percent were assigned to condition C.

B) 33 students, which made up 87 percent of the population, reported which group they were assigned to with 9 students in condition A (representing 24 percent of the population), 10 students in condition B (representing 26 percent of the population), and 14 students in condition C (representing 37 percent of the population). 5 students did not answer the question (representing 13 percent of the population) for a total population of N=38 students.

III. HeartRate and Cheerful:

Statistics

 

Heart Rate

Right now, I feel cheerful...

N

Valid

38

38

Missing

0

0

Mean

65.79

2.89

Std. Deviation

14.994

1.060

Minimum

10

1

Maximum

90

5

 

 

A) The interval variable, “right now I feel cheerful”, had a mean of 2.89 and a standard deviation of 1.060. The interval variable, “HeartRate”, had a mean of 65.79 and a standard deviation of 14.994.

B) Out of 38 students the average (mean) heart rate reported was around 66 beats per minute with some students falling either 15 beats below the mean or 15 beats above the mean. The average response from students when asked to rate how cheerful they felt right now was 3 (somewhat) with some students rating themselves below the mean as 2 (a little) or above the mean as 4 (a lot).

C) The histogram of the variable “right now I feel cheerful” displays no outliers and would be considered a normal distribution with most scores tightly clustered around the mean.

     The histogram of the variable “HeartRate” displays the presence of outliers and would be considered a negatively skewed distribution with the outliers pulling the tail towards the left of the graph away from the mean.

DRAFT: This module has unpublished changes.

Darius Muniz

Professor Godfrey

Research Methods II

September 26th, 2011

 

T-Test

A. The Null hypothesis is that there is no difference whether being in group A (writing something good that happened to you) or being in group C (writing the alphabet) and which neither group has an effect on heart Rate.

The Alternate/Experimental hypothesis is that being in group A or group C does make a difference and that being in a specific group did have an effect on Heart Rate.

B.

T-TEST GROUPS=ExperimentalGroup(1 3)

  /MISSING=ANALYSIS

  /VARIABLES=HeartRate

  /CRITERIA=CI(.95).

 

Group Statistics

 

What experimental group were you in?

N

Mean

Std. Deviation

Std. Error Mean

Heart Rate

A (something good that happened to you)

9

64.67

12.649

4.216

C(write letters)

14

64.71

11.809

3.156

 

 

Independent Samples Test

 

Levene's Test for Equality of Variances

t-test for Equality of Means

F

Sig.

t

df

Sig. (2-tailed)

Mean Difference

Std. Error Difference

95% Confidence Interval of the Difference

Lower

Upper

Heart Rate

Equal variances assumed

.089

.768

-.009

21

.993

-.048

5.185

-10.830

10.735

Equal variances not assumed

 

 

-.009

16.323

.993

-.048

5.267

-11.195

11.099

 

 

C. An independent-samples t-test was conducted to compare the heart rates for groups A (write 90 seconds about something good that happened to you) N=9 and group C (write the alphabet backwards for 90 seconds) N=14. There was no significant difference in scores of heart rates between group A, (M=64.67, SD=12.649) and group C, (M=64.71, SD=11.809);  t(21)= -0.01, p=.99.

D. For this analysis, we were interested in knowing if whether being in group A (write 90 seconds about something good that happened to you) or group C (write the alphabet backwards for 90 seconds) had an effect on Heart Rate. Since the t-test was not significant, we can conclude that there was no difference in heart rate contributed to being in either group A or group B other than what would be expected to occur by chance. This study demonstrates that there is no association between experimental group/writing condition and HeartRate.

ANOVA

A. The Null hypothesis is that experimental group does not affect “being in good spirits”.

The Alternate/Experimental hypothesis is that experimental group does have an effect on being “in good spirits”.

B.

ONEWAY InGoodSpirits BY ExperimentalGroup

  /MISSING ANALYSIS

  /POSTHOC=TUKEY ALPHA(0.05).

 

ANOVA

Right now, I feel in good spirits...

 

Sum of Squares

df

Mean Square

F

Sig.

Between Groups

15.598

2

7.799

11.747

.000

Within Groups

19.917

30

.664

 

 

Total

35.515

32

 

 

 

 

 

Multiple Comparisons

Right now, I feel in good spirits...

Tukey HSD

(I) What experimental group were you in?

(J) What experimental group were you in?

Mean Difference (I-J)

Std. Error

Sig.

95% Confidence Interval

Lower Bound

Upper Bound

A (something good that happened to you)

B(something good happened to someone you knew)

.811

.374

.094

-.11

1.73

C(write letters)

-.817

.348

.064

-1.68

.04

B(something good happened to someone you knew)

A (something good that happened to you)

-.811

.374

.094

-1.73

.11

C(write letters)

-1.629*

.337

.000

-2.46

-.80

C(write letters)

A (something good that happened to you)

.817

.348

.064

-.04

1.68

B(something good happened to someone you knew)

1.629*

.337

.000

.80

2.46

*. The mean difference is significant at the 0.05 level.

 

 

Right now, I feel in good spirits...

Tukey HSDa,b

What experimental group were you in?

N

Subset for alpha = 0.05

1

2

B(something good happened to someone you knew)

10

2.30

 

A (something good that happened to you)

9

3.11

3.11

C(write letters)

14

 

3.93

Sig.

 

.072

.069

Means for groups in homogeneous subsets are displayed.

a. Uses Harmonic Mean Sample Size = 10.618.

b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

 

C. A one-way analysis of variance (ANOVA) was performed to investigate if “being in good spirits” was affected by experimental group A (writing 90 seconds about something good that happened to you) or experimental group C (writing backwards for 90 seconds). The ANOVA was statistically significant, F(2,30)=11.75, p<.001. Follow-up tests were to evaluate differences in pairs between the means. Tukey’s HSD was used to conduct the post hoc comparisons. There was a significant difference between the means of writing condition B and writing condition C. The mean for experimental group being in writing condition C “writing letters” (M=3.93) is higher than the means for both writing condition B “write something good about someone else” (M= 3.11) and writing condition A “write something good about you” (M= 2.30).

D. We wanted to find out if students who felt in good spirits were affected by the writing condition they were assigned to. The results showed that there was a significant difference about being in good spirits between the group of students who were assigned to writing condition C and the group of students who were assigned to condition B. Students who were in writing condition C (wrote the alphabet) reported being in good spirits more than students who were in writing condition B(wrote about someone else), and even greater spirits than students who were in writing condition A (who wrote about themselves).

DRAFT: This module has unpublished changes.

Darius Muniz

Professor Godfrey

Research Methods II

October 23, 2011

 

Analysis 1

Run bivariate correlations between quality of sleep, stress over the last month and anxiety.

A. Copy and paste the correlation matrix from the output below.

DATASET ACTIVATE DataSet1.

CORRELATIONS

  /VARIABLES=stressmo qualslp anxiety

  /PRINT=TWOTAIL NOSIG

  /MISSING=PAIRWISE.

 

Correlations

 

how stressed over last month

quality of sleep

HADS Anxiety

how stressed over last month

Pearson Correlation

1

-.247**

.532**

Sig. (2-tailed)

 

.000

.000

N

271

268

268

quality of sleep

Pearson Correlation

-.247**

1

-.360**

Sig. (2-tailed)

.000

 

.000

N

268

268

265

HADS Anxiety

Pearson Correlation

.532**

-.360**

1

Sig. (2-tailed)

.000

.000

 

N

268

265

268

**. Correlation is significant at the 0.01 level (2-tailed).

 

 

B. Write one paragraph of a Results section to report the findings in APA style. Be sure to include the significant and non-significant findings and the statistics.

 

            In order to determine the correlation between quality of sleep, stress over last month and anxiety, a correlations analysis was performed. Anxiety was significantly positively correlated with stress over the last month (r = .53, p<.001) and negatively correlated with quality of sleep  (r = -.36, p<.001). Quality of sleep was significantly negatively correlated with stress over the last month (r =-.25, p<.001).

 

Analysis 2

First, run a linear regression using sex to predict anxiety.

A. Copy and paste the output from the linear regression below. (This should include the model summary, ANOVA and coefficients boxes).

 

Model Summary

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

dimension0

1

.247a

.061

.057

3.390

a. Predictors: (Constant), Sex

      

 

 

ANOVAb

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

197.998

1

197.998

17.224

.000a

Residual

3057.778

266

11.495

 

 

Total

3255.776

267

 

 

 

a. Predictors: (Constant), Sex

b. Dependent Variable: HADS Anxiety

 

 

Coefficientsa

Model

Unstandardized Coefficients

Standardized Coefficients

t

Sig.

B

Std. Error

Beta

1

(Constant)

5.388

.308

 

17.482

.000

Sex

1.727

.416

.247

4.150

.000

a. Dependent Variable: HADS Anxiety

 

B.  After seeing the output from the regression analyses, in your own words, describe what each of the following pieces of output are and what are they mean for this example: R square, regression F-test, the intercept and the coefficient (b) for sex.

1. R Square: 6.1% of the variance in anxiety can be attributed to sex.

2. Regression F-Test: Is it large enough than what we could expect by chance alone, yes because 17.22, p<.001, which is significant.

3. Intercept and Coefficient (b) for sex: The intercept is 5.39 which is the predicted score on anxiety when sex=0, or when sex is male, the positive sign shows that as sex increases by one unit or changes to female, anxiety increases by 1.73, and Beta tells us that for every standard deviation increase or change from male to female, anxiety increases by .25 standard deviations.

 

C. Write one paragraph of a Results section to report the findings in APA style. Be sure to include the significant and non-significant findings and the statistics from each of the analyses.

            Standard multiple regression was used to assess the ability of sex to predict anxiety. The total variance explained by the model was 6.1%, F (1,266) = 17.22, p < .001. For every unit of sex, anxiety increased by 1.73 (beta=.25, b = 1.73, p<.001), making sex a significant predictor for anxiety in this sample.

 

D. Suppose you want to explain the results of this analysis to a friend who is not in the class. Translate the results paragraph into a one-sentence take-home message.

            This analysis shows that when a person’s sex changes from male to female it is positively associated with the amount of anxiety that a person feels.

 

Analysis 3

First, recode quality of sleep such that it ranges from 0 to 5 and call the new variable “quality of sleep recoded”. We will do this together in lab.

Second, run a hierarchical linear regression using quality of sleep recoded and stress over last month to predict anxiety. In the first step, include quality of sleep recoded as the predictor. In the second step, add stress over last month as a second predictor. Make sure you check the “change in Rsquare” box under the statistics tab.

 

A. Copy and paste the output from the linear regression below. (This should include the model summary, ANOVA and coefficients boxes).

 

 

Model Summary

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

Change Statistics

R Square Change

F Change

df1

df2

Sig. F Change

dimension0

1

.360a

.130

.127

3.247

.130

39.261

1

263

.000

2

.579b

.335

.330

2.844

.205

80.844

1

262

.000

a. Predictors: (Constant), QualslpRecoded

b. Predictors: (Constant), QualslpRecoded, how stressed over last month

 

 

 

ANOVAc

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

413.844

1

413.844

39.261

.000a

Residual

2772.217

263

10.541

 

 

Total

3186.060

264

 

 

 

2

Regression

1067.543

2

533.771

66.012

.000b

Residual

2118.518

262

8.086

 

 

Total

3186.060

264

 

 

 

a. Predictors: (Constant), QualslpRecoded

b. Predictors: (Constant), QualslpRecoded, how stressed over last month

c. Dependent Variable: HADS Anxiety

 

 

Coefficientsa

Model

Unstandardized Coefficients

Standardized Coefficients

t

Sig.

B

Std. Error

Beta

1

(Constant)

9.307

.513

 

18.151

.000

QualslpRecoded

-1.097

.175

-.360

-6.266

.000

2

(Constant)

5.247

.637

 

8.239

.000

QualslpRecoded

-.751

.158

-.247

-4.748

.000

how stressed over last month

.664

.074

.467

8.991

.000

a. Dependent Variable: HADS Anxiety

 

B. Write one paragraph of a Results section to report the findings in APA style. Be sure to include the significant and non-significant findings and the statistics from each of the analyses.

            A multiple regression was used to assess the ability of two variables (quality of sleep recoded and how stressed) to predict anxiety. The total variance explained by the model was 33.5%, F (2,262)= 66.01, p<.001. The greatest significant predictor for this model was stress over the last month (beta= .47, b= .66, p<.001), followed by quality of sleep recoded (beta=-.25, b= -.75, p<.001). For every unit increase in stress, anxiety increased by 0.66 and for every unit of increased quality in sleep, anxiety decreased by 0.75.

           

C. Suppose you want to explain the results of this analysis to a friend who is not in the class. Translate the results paragraph into a one-sentence take-home message.

            The higher level of stress a person feels the more anxiety they feel and the more quality sleep a person gets the lower the anxiety they feel. So basically sleep more and reduce stress to feel less anxiety.

 

D. Compare and contrast the results from Model 1 and Model 2. (a) Describe and explain why there is a difference in Rsquare across the two models. (b) Describe and explain why there are differences in the intercept and the coefficient for quality of sleep recoded across the two models.

 

A) Model 1 accounts for the total variance of anxiety that is only attributed to quality of sleep recoded which is 12.7 percent of the total variance of anxiety. Whereas, model 2 accounts for the total variance of anxiety that is attributed to the combined effects of both quality of sleep recoded and how stressed over the last month, which accounts for 33.5 percent of the total variance in anxiety.

B) The intercept in model 1 is 9.31 which is the predicted score on anxiety when quality of sleep recoded alone =0, the negative sign shows that as quality of sleep increases by one unit, anxiety decreases by 1.10, and Beta tells us that for every standard deviation increase in quality of sleep, anxiety decreases by .36 standard deviations. Whereas the intercept in model 2 is 5.25 which is the predicted score on anxiety when both quality of sleep recoded and how stressed over the last month combined =0, the negative sign shows that as quality of sleep increases by one unit, anxiety decreases by .75, and Beta tells us that for every standard deviation increase in quality of sleep, anxiety decreases by .25 standard deviations. The positive sign shows us that as stress increases by one unit, anxiety increases by .66, and Beta tells us that for every standard deviation increase in stress, anxiety increases by .47 standard deviations.

 

DRAFT: This module has unpublished changes.

Darius Muniz

Professor Godfrey

Research Methods II

November 13th, 2011

 

 

Step 1:

 

Total Variance Explained

Factor

Initial Eigenvalues

Extraction Sums of Squared Loadings

Rotation Sums of Squared Loadingsa

Total

% of Variance

Cumulative %

Total

% of Variance

Cumulative %

Total

dimension0

1

4.214

35.116

35.116

2.576

21.467

21.467

2.310

2

2.268

18.897

54.013

2.849

23.739

45.206

3.422

3

1.169

9.742

63.756

1.093

9.111

54.317

2.045

4

.759

6.326

70.081

 

 

 

 

5

.688

5.730

75.812

 

 

 

 

6

.630

5.251

81.062

 

 

 

 

7

.556

4.635

85.698

 

 

 

 

8

.478

3.982

89.680

 

 

 

 

9

.404

3.363

93.043

 

 

 

 

10

.351

2.929

95.972

 

 

 

 

11

.302

2.513

98.485

 

 

 

 

12

.182

1.515

100.000

 

 

 

 

Extraction Method: Maximum Likelihood.

a. When factors are correlated, sums of squared loadings cannot be added to obtain a total variance.

 

 

A) I have decided to retain three Factors based on the eigen values chart in which these factors are greater than 1. Factors 1, 2, and 3 made up the majority of the percentage of variance explained totaling upwards of 63 percent of the total variance and factor three was the cutoff point for where the elbow of the scree plot begins.

 

Step 2:

Structure Matrix

 

Factor

1

2

3

1. interested

-.171

.663

-.207

2. upset

.485

-.240

.589

3. scared

.975

-.223

.390

6. determined

-.109

.498

-.008

10. hostile

.260

-.114

.716

7. active

-.278

.565

-.209

11. irritable

.362

-.215

.824

12. inspired

-.156

.770

-.136

13. attentive

-.183

.676

-.212

14. afraid

.833

-.221

.386

15. excited

-.099

.658

-.099

17. enthsiastic

-.214

.848

-.214

Extraction Method: Maximum Likelihood.

 Rotation Method: Oblimin with Kaiser Normalization.

 

Factor Correlation Matrix

Factor

1

2

3

dimension0

1

1.000

-.250

.425

2

-.250

1.000

-.215

3

.425

-.215

1.000

Extraction Method: Maximum Likelihood. 

 Rotation Method: Oblimin with Kaiser Normalization.

 

A) I labeled factor 1 as Fearfulness due to the high loadings of the variables scared and afraid and how these variables were representative of fearfulness. Factor 2 was labeled as Positive Motivation due to the high loadings of variables that were representative of the construct of positive motivation including interested, determined, active, inspired, attentive, excited, and enthusiastic. Finally I labeled factor 3 as Aggression due to the high loadings of hostile and irritable and how these variables were representative of aggression.

 

B) The variable upset double loaded onto factor 1 at .485 as well as factor 3 at .589. Though this variable had a higher loading for factor 3 (aggression) I chose not to include this variable in either factor 1 nor factor 3 due to its ability to be included in both factor 1 (fearfulness) and factor 3 (aggression). This fact made placing this variable in either fearfulness or aggression a difficult and ambiguous decision.

 

C) Factor 1 (Fearfulness) was negatively correlated with factor 2 (Positive Motivation) and positively correlated with factor 3 (Aggression). These correlations are as expected indicating: as a person’s fearfulness increases as well as a person’s aggression level, a person’s positive motivation level decreases. Also that a person’s fearfulness is positively correlated with a person’s aggression level indicating as fearfulness increases so does a person’s aggression level.

 

Step 3:

A) Seven variables (items) were used to measure participants positive motivation on a 5-point Likert scale ranging from 1 (very slightly) to 5 (extremely) (α = .85).

 

Step 4:

 

Descriptive Statistics

 

N

Minimum

Maximum

Mean

Std. Deviation

PositiveMotivation

436

1.00

5.00

3.4014

.75826

Valid N (listwise)

436

 

 

 

 

 

 

DRAFT: This module has unpublished changes.