Assignment

Statistical testing and bivariate analysis Part 2

Dear student,
Below you find the assignment of Week 4 of Quantitative Research Methods and Statistics.
- Make the assignment using SPSS and put the answers/tables in the file below.
- Upload the assignment as pdf in Canvas.
- Go to the internet quiz https://tue-be-usre.my-survey.host/index.php/676465?lang=en
and fill in the answers there. Copy-and-paste the link to your browser.

Please enter the id-number that is represented on your student ID-card in the
orange section of the card.

Otherwise you will not be recognized!

The computer will process your answers and you will receive personalized feedback on what
you did correct or wrong.

1
Used Tables

2
3
4
5
Theory two-paired T-test and ANOVA

The questions below are based on the lecture

L1. A researcher has a sample of dwelling transactions from the Amsterdam neighbourhood
Jordaan. She wants to test whether the average price of these sold dwellings equals the -
known - average price of dwellings in Amsterdam. Which test should she use? __

A one-sample t-test
B independent samples t-test
C paired-samples t-test
D F-test (Levine)

L2. A researcher has two samples: dwellings sold in Amsterdam and dwellings sold in
Eindhoven. She wants to test whether the average transaction price in these two samples is
the same. Which test should she use? __

A one-sample t-test
B independent samples t-test
C paired-samples t-test
D F-test (Levine)

L3. The researcher expects that the dwellings sold in Amsterdam are more expensive than the
dwellings sold in Eindhoven and wants to test this. What kind of test should she use? __

A one-sided t-test
B two-sided t-test
C paired sample t-test

L4. To test whether the unknown variances of dwellings sold in Amsterdam and Eindhoven are
equal, which test should she use? __

A one-sample t-test
B independent samples t-test
C paired-samples t-test
D F-test (Levine)

L5. The newspaper reports that the fiscal value of the dwellings (WOZ) in Amsterdam falls
short of their transaction value. You want to test whether this is true. __

What is here your null hypothesis (H0)?

What is here your alternative hypothesis (H1)?

A The average price of dwellings sold in Amsterdam equals the average fiscal value (WOZ) of these
same dwellings
B The average price of dwellings sold in Amsterdam does not equal the average fiscal value (WOZ) of
these same dwellings
C The average price of dwellings sold in Amsterdam is higher than the average fiscal value (WOZ) of
these same dwellings

6
L6. Which test should we use to test H1 in question L5? __

A A one-sample t-test
B An independent samples t-test
C A paired-samples t-test
D An F-test (Levine)

L7. The test yields a p-value of 0.0075. What conclusion can we make? __

A The H0 can be rejected at a 5% significance level

B The H0 can be rejected at a 1% significance level
C The H0 cannot be rejected

L8. In an independent samples t-test __

A The number of observations in the two groups needs to be the same

B The number of observations in the two groups needs not to be the same

L9. In a paired t-test __

A The number of observations in the two groups needs to be the same

B The number of observations in the two groups needs not to be the same

L10. You want to test whether dwellings sold in Amsterdam, Eindhoven and Rotterdam have
significantly different average prices. Which statement is true? __

A Anova should be used; using the t-test increases the probability of drawing a wrong conclusion
B The t-test should be used; using Anova increases the probability of drawing a wrong conclusion
C Both, Anova and the Student t-test can be used interchangeably

L11. In Anova, the post-hoc test should be used: __

A Always
B Only if Anova rejects H0 of no differences in average values between groups

L12. In Anova, if H0 is rejected, it means that: __

A Between-Groups Variance ≈ Within-Groups Variance

B Between-Groups Variance < Within-Groups Variance
C Between-Groups Variance > Within-Groups Variance

L13. If the F-test (Levene’s test) rejects the hypothesis of equal variances: __

A Anova can still be used

B Anova cannot be used

7
SPSS CASE 1

Use the file ‘comfort_s.sav’. These data are inspired by research done in the groups BPS and USRE.

Imagine a housing association which is considering investment in the energy efficiency of dwellings.
They can invest in two different technologies: solar panels or insulation. They want to know the effect
of an energy investment on the feeling of comfort of tenants. For this purpose an experiment was
performed. A group of dwellings was divided in three parts.
- the control group did not get any energy investment
- treatment group 1 got insulation
- treatment group 2 got solar panels.

Then a survey was performed measuring the subjective feeling of comfort in each of these dwellings.
The resulting dataset contains the following variables:

• id: identification number

• gender: 0=male, 1=female
• age: age of the person in years
• comfort thermal sensation (how do you feel at the moment?): -3=cold, -2=cool, -
1=slightly cool, 0=neutral, 1=slightly warm, 2=warm and 3=hot
• treatment a dichotomous variable (0/1) indicating whether the dwelling received an
energy efficiency investment; 0 is no, 1 is yes
• treatment2 a categorical variable (0/1/2) indicating whether the dwelling has been made
energy-efficient, and if yes, how; 0 is not treated, 1 is insulation, 2 is solar
panels

8
SPSS Case 1 Comparing the means of two groups

1. Calculate and report (round up at 2 digits) the average comfort for:

- all people in the control group (treatment=0) __
- all people in the treatment groups (treatment=1) __
- all people in the insulation treatment group (treatment2=1) __
- all people in the solar panels treatment group (treatment2=2) __

2 A researcher compares the control group and the whole treatment group (use variable
treatment).
The researcher expects that energy efficiency investment increases the living comfort.

What is the null hypothesis (H0)? __

What is the alternative hypothesis (H1)? __

A. mean comfort of the treatment group is equal to the mean comfort of the control group
B. mean comfort of the treatment group is higher than the mean comfort of the control group
C. mean comfort of the treatment group is not equal to the mean comfort of the control group

3 Should you test one sided or two sided? __

A. One-sided test
B. Two-sided test
C. Does not matter

Explain your choice: __

A. There is a direction in the null hypothesis
B. There is a direction in the alternative hypothesis
C. There is no direction in the alternative hypothesis

4 What type of t-test should you use here? __

A. 1-sample t-test
B. 2-sample t-test

Explain your choice: __

A. We compare averages of two groups, so there are two samples
B. We use the same data in H0 and H1, so there is one sample.

5 Are the conditions met for this test?

Condition 1: there is no overlap between the groups __
A. yes
B. no
Condition 2: Normality of the distribution: both groups have more than 30 cases __
A. yes
B. no

9
6 Conduct the t-test.
Use Analyze/Compare Means/Independent-Samples T Test.
Look at the outcomes of the Levene’s test. Can we assume equal variances? __
A. yes
B. no

Explain: __
A. The Levene’s test is not significant. So we cannot reject the null hypothesis that the variances are
equal.
B. The Levene’s test is not significant. So we can reject the null hypothesis that the variances are
equal.

7 Look at the outcomes of the t-test.

How large is the two-sided (two-tailed) p-value? __
And how large is the one-sided p-value? __
Which value should we use? __
A. Two-sided
B. One-sided

8 Is the test significant at a 5% significance level? __

A. Yes
B. No

Explain: __
A. The p-value is smaller than 0.95
B. The p-value is smaller than 0.05

9 Choose the correct conclusion based on the test (look at H1 in question 2). __
A. The null hypothesis cannot be rejected. The average comfort of the treatment group does not differ
from the average comfort of the residents of the control group.
B. The null hypothesis must be rejected. The average comfort of the treatment group is higher than the
average comfort of the residents of the control group.
C. The null hypothesis must be rejected. The average comfort of the treatment group differs from the
average comfort of the residents of the control group.

10 Now the researcher wants to test whether the treatment and the control group are
comparable in average age and in gender.
Because if they are not, then the measured difference in comfort might be because different people
have different feeling of comfort.

Indicate by characteristic whether you will perform.

A. A t-test
B. A Chi-squared test.
Remember that the Chi-square test is used to compare two nominal variables. The independent
samples t-test is used to compare interval/ratio variables.

Which test
Gender
Age

10
11. Perform the tests.
Report below the p-value and the conclusion. Use a 5% significance level and a 1% significance level.
A. H0 of equal means can be rejected
B. H0 of equal means cannot be rejected

p-value (2 digit) Conclusion 5% Conclusion 1%

significance significance
Gender
Age

11
SPSS Case 2 Comparing the means of three groups

12 Now the researcher wants to examine whether the type of technology used matters for the
feeling of comfort.
For this she uses variable treatment2 (0=non-treated, 1=insulation, 2=solar panels).
Which technique do you need to perform this analysis? __
A. Chi-squared test
B. 1-sample t-test
C. 2-sample t-test
D. ANOVA

Why can the researcher not use a Chi-squared test? __

A. Because she wants to compare the mean comfort for more than two groups.
B. Because the variable comfort is an interval (ratio) variable.
C. Because the variable comfort is a nominal variable.

13 Perform the test. Use Analyze/Compare Means/One-Way ANOVA.

What is the mean comfort level of each group?
Control __
Treatment 1 __
Treatment 2 __

What is the F-value of the test? __

What is the corresponding p-value? __

What is your conclusion? __

A. We cannot reject the null hypothesis that the mean comfort does not differ between the three
groups
B. We can reject the null hypothesis.

14 The researcher wants to understand which energy investments add to the comfort in house.
Which of the two treatment groups differ significantly in comfort from the control group? __
(Use Analyze/Compare Means/One-Way ANOVA/post-hoc test, option Bonferroni)

A. The insulation treatment group differs in average comfort from the control group
B. The solar panel treatment group differs in average comfort from the control group
C. Both, the insulation and the solar panel treatment groups differ in comfort from the control group
D. There is no difference in comfort between the two treatment and the control group.

15 Are the conditions to apply the ANOVA technique met in this case? __
The three conditions are: (i) groups are independent of each other, (ii) sampling distribution is normally
distributed (each group is larger than 30), (iii) variances of the groups are equal

A. No, (i) is not met

B. No, (ii) is not met
C. No, (iii) is not met
D. Yes, all are met

12
16 Perform the Homogeneity of variances test (Levene). Use options in the ANOVA test menu.
What is the p-value? __

What does this mean? Use a 5% significance level. __

A. There is no statistically significant difference between the variances of the two groups.
B. There is a statistically significant difference between the variances of the two groups.

17. Here again the researcher wants to be sure that the two treated and the control groups are
comparable. Indicate by variable which test you will perform.
A. t-test
B. Chi-squared
C. Anova

Which test?
Gender
Age

18. Perform the test.

Report the p-value and the conclusion in the table below. Use a 5% and a 10% significance level.
A. H0 of equal means can be rejected
B. H0 of equal means cannot be rejected

p-value Conclusion 5% significance Conclusion 10% significance

Gender
Age

13