Soc.3155 Homework 4 Answers

 

11.6  The column percentages are:

Alertness

Partic.

Nonpartic.

Total

High

67.6

45.5

56.7

Low

32.3

54.5

43.3

Total

100

100

100

                                               

A greater percentage of participants in pet therapy reported having high alertness, compared to nonparticipants.

 

The calculations for the Chi Square are:

H0: No association in the population between alertness and participation in pet therapy program.

HA: There is association in the population.

 

Cell

fo

fe

fo - fe

(fo - fe)2

(fo -fe )2/fe

Part., high

23

19.3

  3.7

13.69

0.71

Part., low

11

14.7

- 3.7

13.69

0.93

Non., high

15

18.7

- 3.7

13.69

0.73

Non., low

18

14.3

  3.7

13.69

0.96

 

 

 

 

 

3.33 (Calculated Χ2)

 

Critical Value: Χ2 (1) = 3.84 (.05 level)

Retain H0: There is no evidence of association between participation in pet therapy and alertness in the population.

 

Other measures:

Lambda: Alertness is the dependent variable.

            Method 1: Predict “High” for all 67 cases;   E1 = 29

            Method 2: Predict “High” for participants   errors = 11

                             Predict “Low” for nonparticipants, errors = 15    E2 = 26

            L = (29-26)/29 = .103   Indicates a weak association between pet therapy and alertness.

 

Gamma: PA = (23)(18) = 414 PD =   15(11) = 165    

G = (PA - PD)/(PA + PD) =(414 - 165)/(414 + 165) = .43 Indicates a moderate positive association.

 

Summary: There is somewhat higher alertness among those who participate in pet therapy but the relationship is not generalizable to the population.   Lambda indicates that participation in the program contributes little to the prediction of alertness, and Gamma shows a moderate consistency of rankings among the cases in the cells.  If the goal is to draw conclusions about the effect of this program in the population, I would use the Chi Square.  But if we just want to describe association in the sample, percentages are easy to understand, and Gamma is a good summary measure because it uses all of the information provided by ordinal level data.

 

SPSS Part 1: The relevant parts of the SPSS output are:

 

                                                                                                                 Crosstabs

 

 got into fight

052D06F: R likes to take risks sometimes.

Total

 

DISAGREE:(1)

MOST DIS:(2)

NEITHER:(3)

MOST AGR:(4)

AGREE:(5)

 

 

Did not

140

157

306

398

176

1177

 

 

90.3%

92.9%

90.3%

89.8%

77.2%

88.2%

 

Did

15

12

33

45

52

157

 

 

9.7%

7.1%

9.7%

10.2%

22.8%

11.8%

Total

155

169

339

443

228

1334

 

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

 

                                                                                                                           

 shoplifted

052D06F: R likes to take risks sometimes.

Total

 

DISAGREE:(1)

MOST DIS:(2)

NEITHER:(3)

MOST AGR:(4)

AGREE:(5)

 

 

Did not

135

143

271

322

141

1012

 

 

87.1%

84.6%

79.9%

72.9%

62.1%

76.0%

 

Did

20

26

68

120

86

320

 

 

12.9%

15.4%

20.1%

27.1%

37.9%

24.0%

Total

155

169

339

442

227

1332

 

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

 

                                                                                                                               

 had 5 or more drinks

052D06F: R likes to take risks sometimes.

Total

 

DISAGREE:(1)

MOST DIS:(2)

NEITHER:(3)

MOST AGR:(4)

AGREE:(5)

 

 

Did not

127

137

236

292

109

901

 

 

87.6%

82.5%

74.2%

68.4%

49.8%

70.7%

 

DId

18

29

82

135

110

374

 

 

12.4%

17.5%

25.8%

31.6%

50.2%

29.3%

Total

145

166

318

427

219

1275

 

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

 


                                                                                                                            

 drove after drinking

052D06F: R likes to take risks sometimes.

Total

 

DISAGREE:(1)

MOST DIS:(2)

NEITHER:(3)

MOST AGR:(4)